p 4, 7 DAYS \/ | o3^ Steller s Sea Cow Amchitka, Alaska ' -;f™, > £■ *'!*■ ■ :'-‘’ . J. >^r* " ’ ■ “ ‘•^-‘■-^ f' ■ b r ’ JLv fi ~ • • # &»&££ * iP •*b \ ,V‘ aS?^\ 'iBB  Steller’s Sea Cow (Hydrodamalis gigas) of Late Pleistocene Age from Amchitka, Aleutian Islands, Alaska By FRANK C. WHITMORE JR., and L. M. CARD, JR. GEOLOGICAL SURVEY PROFESSIONAL PAPER 1036 Description of the first specimen of the extinct Arctic sea cow, Hydrodamalis gigas, to be found in place in Pleistocene deposits outside the Commander Islands, U.S.S.R. UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON . 1977UNITED STATES DEPARTMENT OF THE INTERIOR Cecil D. Andrus, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Whitmore, Frank C. Steller’s sea cow (Hydrodamalis gigas) of late Pleistocene age from Amchitka, Aleutian Island, Alaska. (Geological Survey Professional Paper 1036) Bibliography: p. Supt. of Docs, no.: I 19.16:1036 1. Hydrodamalis gigas, Fossil. 2. Paleontology—Pleistocene. 3. Paleontology—Alaska—Amchitka Island. I. Gard, Leonard Meade, 1923- joint author. II. Title: Steller’s sea cow... III. Series: United States Geological Survey Professional Paper 1036 QE882.S6W49 569'.5 77-608068 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 Stock Number 024-001-03001-2CONTENTS Page Abstract................................................... 1 Introduction .............................................. 1 Acknowledgments ........................................... 3 Occurrence and age of the Amchitka specimens .............. 3 Stratigraphic and structural relations.................. 3 Other fossils .......................................... 5 Environment............................................. 6 Age of the deposit ..................................... 7 Composition of the collection ............................. 7 Description of the Amchitka Hydrodamalis................... 8 Measurements ........................................... 9 Page Description of the Amchitka Hydrodamalis — Continued Mandible ............................................. 9 Vertebrae............................................. 10 Scapula............................................... ix Humerus............................................... 12 Radius and ulna....................................... x3 Ribs.................................................. 13 Chevron bone ......................................... 16 Morphologic conclusions............................... 16 Discussion ............................................... is References cited.......................................... 19 ILLUSTRATIONS [Plates follow “References Cited”) Cover art Steller’s sea cow, reproduced by permission from the artist from a painting by Alfred G. Milotte. Plates 1—8. Hydrodamalis gigas. Figure 1. Index map of the Aleutian arc ............................................................................. 2 2. Index map of Amchitka Island.......................................................................... 4 3. Photograph of interglacial beach deposit.............................................................. 5 4. Cross section of South Bight graben .................................................................. 6 5. Photograph of mandibles of Hydrodamalis .............................................................. 6 6. Photograph of rib of Hydrodamalis .................................................................... 8 7-10. Drawings: 7. Key to measurements of mandibles of Hydrodamalis .................................................... 10 8. Key to measurements of scapula and humerus of Hydrodamalis........................................... 13 9. Key to measurements of radius and ulna of Hydrodamalis............................................... 15 10. Rib of Hydrodamalis.................................................................................. 16 TABLES Page Table 1. Foraminifera and other fossils associated with Hydrodamalis from the South Bight interglacial beach deposit, Amchitka Island.......................................................................................... 7 2. Measurements of mandibles of Hydrodamalis gigas .................................................... 11 3. Measurements of vertebrae of Hydrodamalis gigas.................................................... 12 4. Measurements of forelimb of Hydrodamalis gigas..................................................... 14 5. Comparison of Amchitka specimen with immature specimen from Bering Island measured by Von Nordmann (1863) 17 illSTELLER’S SEA COW (HYDRODAMALIS GIGAS) OF LATE PLEISTOCENE AGE FROM AMCHITKA, ALEUTIAN ISLANDS, ALASKA By Frank C. Whitmore, Jr., and L. M. Gard, Jr. ABSTRACT A partial skeleton of Hydrodamalis has been collected from a Pleistocene interglacial beach deposit, 35 meters above present sea level, on Amchitka, Aleutian Islands, Alaska. This is the first discovery of the genus in place in Pleistocene deposits outside the Commander Islands (Komandorskiye Ostrova), U.S.S.R. Parts of the young but apparently nearly full grown individual were found in a position which suggested that the animal, after death, had been washed into shallow water or up onto a beach. The Pleistocene animal was toothless, as was the modern Steller’s sea cow (Hydrodamalis gigas), which was exterminated by Russian hunters about 1768. Like the modern form, the Pleistocene sea cow undoubtedly subsisted on kelp. It is indistinguishable from the modern species. A uranium-series date of 127,000±8,000 years on bone from this beast is consistent with a generally accepted age of the last major interglacial stage. INTRODUCTION The giant Arctic sirenian Hydrodamalis gigas (Zim-mermann), popularly known as Steller’s sea cow, was discovered in 1741 along the coast of what is now called Bering Island (Beringa Ostrov) in the Commander Islands (Komandorskiye Ostrova), U.S.S.R. (fig. 1). Its discoverer, and the only naturalist who observed it alive, was G. W. Steller, a German who sailed on the second voyage of the Danish explorer Vitus Bering, in the service of the Russian government (Golder, 1925). Steller’s description of the animal was published (1751) posthumously and was translated from Latin into English in 1899. A female specimen of Hydrodamalis, measured and dissected by Steller (1751, p. 294), was 7.4 m long from the extremity of the upper lip to the extreme right cornu of the caudal fork. The animals are estimated to have attained a weight of 10 metric tons (Scheffer, 1972, p. 913). Steller’s sea cow was discovered when Bering’s crew was shipwrecked on Bering Island in the course of their return voyage from North America. The starving Russians captured some sea cows with great difficulty, not because they were fierce but simply because their bulk made them hard to haul ashore, and found them to be very good eating. This news was passed by the survivors to subsequent Russian voyagers who, in succeeding years, made the Commander Islands a victualing stop on North Pacific fur-hunting expeditions. Stejneger (1887, p. 1049) stated that from 1743 until 1763 “hardly a winter passed without one or more parties spending eight or nine months in hunting fur-animals there, diming which time the crews lived almost exclusively on the meat of the sea-cow.” The result was that, by 1768, Hydrodamalis was extinct (Sauer, 1802; von Baer, 1840; Brandt, 1846; Stejneger, 1887, Lucas, 1891). Steller, in 10 months on Bering Island, had ample opportunity to observe the daily activities of Hydrodamalis from his hut on the shore. He reported that the animals congregated in herds, feeding incessantly on kelp in shallow water. They were fond of shallow sandy places along the seashore, especially along the mouths of rivers and creeks. “As they feed they move first one foot and then the other, as cattle and sheep do when they graze, and thus with a gentle motion half swim and half walk” (Steller, 1899, p. 198). In his anatomical description, Steller stated that the neck is short but movable: It “has its independent action, a motion observed in the living animal only when it feeds; for it bends its head in the same way as cattle on dry land” (1899, p. 187). The need for vast amounts of seaweed in shallow water was certainly a limiting factor in the distribution of Hydrodamalis, but, even for an animal of such specialized requirements, our knowledge of its range is slight. Only in the Commander Islands is there convincing evidence of the sea cow’s existence in historic time, although there has been much speculation on its existence elsewhere in the North Pacific area (Gard and others, 1972, p. 867). A single rib was found on the island of Attu in the Aleutians (Brandt, 1868, p. 294). During the course of the U.S. Atomic Energy Commission operations on Amchitka, Gard had occasion to talk with Mr. Paul Higdon, then labor foreman for Holmes and Narver, Inc. Higdon had been on Shemya Island (56 km east of Attu) during construction of an airstrip in World War II and remembered a complete skeleton of a large marine mammal being uprooted and pushed aside by construction equipment. Shafer (in Gates and others, 1971, p. 783) reported that the 12 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA Figure 1. — Index map of the Aleutian arc, showing Amchitka and Bering Islands. bedrock surface of Shemya is a wave-cut platform “most likely of pre-Wisconsinian age” covered by marine deposits, and then glaciated. Perhaps an opportunity to recover the entire skeleton of a sea cow was lost at that time. Many sea-cow bones were found along the shore of Bering Island (Chelnokov, 1969; Scheffer, 1973). Stej-neger (1884) took advantage of this situation to make a collection for the U.S. National Museum. Some bones were collected by Stejneger himself; others were purchased from natives. The composite skeleton now exhibited at the U.S. National Museum of Natural History consists of bones from the Stejneger collection. Among composite skeletons mounted in European museums, one of the best is that at the Naturhistoriska Riksmuseet in Stockholm, Sweden (pi. 1). Sea-cow bones on Bering Island were commonly covered by a layer of earth and gravel; Nordenskjold (1882, p. 280) reported it to have been 30 to 50 cm thick: “In order to find them, as it would be too troublesome to dig the whole of the grassy bank, one must examine the ground with a pointed iron rod * * All the bones were apparently found on a low terrace just above the beach “which may be accounted for by the accumulation of storm-wash under the present conditions or very nearly so” (Dawson, 1894, p. 125). Some bones were found at a considerable distance inland. In the Stejneger papers in the Archives of the Smithsonian Institution are notes by Prof. B. W. Evermann “concerning the Rhytina recently acquired by the Museum,” which state that the bones, which were purchased from a “creole” on Bering Island in 1892, lay under about 1 m of sand, three versts (about 3 km) back from the water. Most of the bones in the Stejneger collection at the U.S. National Museum of Natural History are white and hard but not per-mineralized or replaced. An exception to this is USNM 218376, an os occipitis and the first and seventh ribs of the right side, which have a red-ocher color. These specimens are from the anterior part of a skeleton found by Stejneger (1884, p. 61-62) in a sand bank 3.6 m high, 150 m from the sea, and not less than 3-4 m above high tide. The skeleton was on its back. The bones were soft and of a soapy consistency, most of them too soft to be collected. In both softness and red color, these bones resemble some of those discovered on Amchitka, which raises the possibility that there may be sea-cow bones of Pleistocene age on Bering Island in addition to those left by the Russian fur hunters. Considering the frequency and violence of storms in the Commander Islands, it is not surprising that bon is of sea cows butchered in the 18th century should have since been buried under a meter or so of storm-tossed sand. Those buried by the infrequent storms that caused waves to roll far inland could have subsequently been covered by vegetation. The mode of occurrence of the bones led Woodward (1885, p. 457-458) to concludeOCCURRENCE AND AGE OF AMCHITKA SPECIMENS 3 that specimens collected by Robert Damon for the British Museum (Natural History) were of Pleistocene or early Holocene age: “The specimen now in the British Museum was obtained from compact peat, and all the vertebrae and other bones having cavities in them were full of peat-growth when they arrived, as was also the skull.” Without careful field study it is impossible to tell whether sea-cow bones from Bering Island are of Pleistocene or Holocene age, but the historic evidence for extensive killing of sea cows in the 18th century leads to the conclusion that most of the bones date from that period. The 18th-century range of Hydrodamalis, restricted to the Commander Islands, was certainly that of a relict species, perhaps on its way to extinction even without the influence of man. It is logical to assume that the earlier range of Hydrodamalis extended eastward along the Aleutian chain. If this was so during the last 10,000 years or so, Hydrodamalis would have been ideal prey for the sea-hunting Aleuts (Domning, 1972). However, there is no known archeological evidence that the Aleuts hunted the sea cow. Desautels and others (1969-70), in a report on the archeology of Amchitka, recorded many sea-mammal bones from old Aleut middens but included no Hydrodamalis. Six sites (Roger Desautels, written commun., 1971), which revealed occupation over a 2,500-year span, were investigated. Jean S. Aigner (written commun., 1971) reported that there are no known examples of Hydrodamalis from any Aleut middens and that a large number of middens had been sampled, some as much as 4,000 years old. Domning (1972) speculated that some middens ancient enough to contain Hydrodamalis bones may still exist but that most of them have been covered by the rising sea. The only occurrence known to us of Hydrodamalis bone in a midden was reported by Hall (1971) from a 16th-century Eskimo site at the confluence of the Noatak River and Kangiguksuk Creek in northwestern Alaska (67°58' N., 161°50' W.). A rib, which had the distal end sawed off and which was heavily marked with cuts on both sides, was found in floor fill within the house. It is unquestionably associated with the rest of the cultural material at the site (Hall, written commun., 1976). Hall (1971, p. 23) also reported the presence at the site of 133 fragments of fossil mammoth ivory and a bison horn and phalanx, all of Pleistocene age. He pointed out (1971, p. 56) that “mammoth ivory and the bones of extinct animal species are frequently washed out of the permanently frozen banks when the Noatak shifts its course.” The Kangiguksuk site is about 100 km from the ocean. The rib from this site is described in more detail on p. 15. Hydrodamalis was unknown south of Alaska until 1967, when Jones (1967) reported that a badly eroded cranial fragment had been brought up in a trawl from the bottom of Monterey Bay, Calif., and that it had yielded a C14 radiometric age of 18,940±1,100 years B.P. The Monterey specimen is in the lower part of the size range of a series of modern skulls from Bering Island. Shikama and Domning (1970) reported a rib of Hydrodamalis from the late Pliocene of the island of Honshu, Japan. They also mentioned that two specimens of a new species of Hydrodamalis have been discovered in the Pliocene of California and that this species is morphologically intermediate between Metax-ytherium jordani Kellogg of the late Miocene and Hydrodamalis. (See also Domning, 1970.). ACKNOWLEDGMENTS We thank Daryl P. Domning for many helpful discussions and for reviewing this paper before publication. The photographs of the bones were taken by Robert H. McKinney and Haruo E. Mochizuki, and the drawings for figures 7 —10 were made by Richard J. Mjos. Thomas D. Washburn prepared the specimens for study. We are grateful to Dr. Tor Orvig and Mr. Carl Edelstam of the Naturhistoriska Riksmuseet, Stockholm, Sweden, for furnishing photographs of their mounted composite specimen of Hydrodamalis and for permission to reproduce them. The drawing on the cover is from a painting by Alfred G. Milotte; permission to use it has been granted by the artist. We thank the U.S. Atomic Energy Commission for logistical support on Amchitka. The Department of Anthropology, Yale Peabody Museum, kindly loaned us a rib of Hydrodamalis from an archeological site in northwestern Alaska. OCCURRENCE AND AGE OF THE AMCHITKA SPECIMENS STRATIGRAPHIC AND STRUCTURAL RELATIONS In 1969 Gard discovered a partial skeleton of Hydrodamalis gigas 35 m above present sea level in unconsolidated sediments exposed in the sea cliff at the head of South Bight on Amchitka Island (figs. 2, 3). This exposure, first described by Powers, Coats, and Nelson (1960, p. 542) as an emerged Pleistocene interglacial beach deposit, was preserved because it was penecon-temporaneously downdropped in a small (800-m wide) east-northeast-trending graben within the Amchitka Formation of early Tertiary age (fig. 4). These beds were conformably deposited on an erosion surface that truncates earlier Pleistocene lacustrine and marine sediments. The earlier semiconsolidated sediments also4 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA Figure 2. — Index map of Amchitka Island. were trapped in this graben and were tilted and faulted prior to their truncation. The erosion surface, itself tilted and faulted, ranges from 34 m above mean sea level on the north side of the graben to 15 m above mean sea level on the south side of the graben. This surface and the overlying beach deposit are believed to be related to a stillstand of the sea during the South Bight II marine transgression that cut the most extensive terrace on Amchitka Island (Szabo and Gard, 1975). This terrace is as much as 1.5 km wide. The inner edge of the terrace is now 37 to 49 m above sea level.1 The beach deposit ranges in thickness from about 6 m on the north side of the graben to about 30 m on the south side of the graben. The deposit was not found on the Bering Sea side of the island and does not appear to extend more than 500-600 m back from the sea cliff. The deposit consists of loosely consolidated, poorly bedded sand to coarse boulder gravel, which is richly 1 Altitudes at base of abandoned sea cliffs have been determined from 1968 topographic maps by Holmes and Narver, Inc., scale 1:6,000, contour interval 10 feet. This terrace was erroneously reported as 52 m (Gard and others, 1972, p. 868). fossiliferous. Some beds near the base of the deposit are composed almost entirely of cross-stratified shell fragments; in a few places, discontinuous beds in the lower part of the deposit are composed entirely of unbroken pecten shells, some of which are articulated. Beds in which the partial skeleton of Hydrodamalis was found (fig. 5) are 9 m above the base of the deposit (near the upper middle part) and consist of loose pebbly sand that contains random cobbles and boulders and abundant shell material, including echinoid plates and spines and Foraminifera. Many beds as much as 30 cm thick are composed of well-rounded hornblende-bearing white pumice pebbles, which average 2.5 cm in diameter but which may be as large as 7 cm. The buoyant low-density pumice pebbles probably were derived from eruptions of one of the nearby active volcanoes. Near the base of the deposit, many boulders are coated with calcium carbonate algal secretions, and fragments of these coralline algae (probably Lithothamnion sp.), similar to those found on modern Aleutian Islands beaches, are interspersed between the boulders. An isolated rib of Hy-OCCURRENCE AND AGE OF AMCHITKA SPECIMENS 5 Figure 3. — Interglacial beach deposit overlying tilted unconformity (1) in the cliff face at South Bight, Amchitka Island. Hydrodamalis (USNM 170761) was found at (2). Hydrodamalis rib (USNM 181752) was found at (3), just above unconformity. White bed above (3) is composed of cross-stratified shell fragments. Photograph by R. H. Morris. drodamalis was found in place near the base of the deposit (fig. 6). The difference in thickness between the north and south ends of the deposit is the result of contemporaneous southward tilting of the graben block while these beds were being deposited, which caused a discordance of the bedding in the southern part of the deposit and tilting of the erosion surface at the base. At the southern end, the material that forms the generally coarser grained upper part of the section apparently was eroded from the scarp of the more active south-bounding fault. We assume that these boulders, which are as much as 0.6 m in diameter, were deposited in deeper water, where they escaped abrasion, as many are angular or only slightly subrounded. Although the deposit has been studied in detail, no erosional unconformity is recognized within it. OTHER FOSSILS In addition to bones of Hydrodamalis, the gravel deposit yielded a partial skull of the Steller sea lion, Eumetopias jubata Schreber, the distal half of the radius of a large whale, and a fragment of the tusk of a small walrus. The presence of a rich invertebrate fauna in the deposit has been known for some time. From 1946 to 1951, U.S. Geological Survey personnel collected a molluscan fauna (Powers and others, 1960). A report on Foraminifera collected by R. R. Coats in a gravel pit 200 m north of the Hydrodamalis site was published by Cushman and Todd (1947). In 1969, Allison (1973) collected a large invertebrate fauna from this deposit. The present authors collected samples of sand that contains Foraminifera during excavation of the sea-cow bones in 1971. These were collected from 0.3 m above the base of the deposit, from the sea-cow horizon, and from 1 m above the sea-cow horizon. Foraminifera in these samples, identified by Ruth Todd (table 1) are all living species and provide no evidence of age. In addition, in 1969, a collection of Foraminifera was made by Gard from about 1 m below the sea-cow horizon and was identified by the late R. L. Pierce, U.S. Geological Survey. According to Todd (written commun., 1971), the 1969 collection (table 1, col. 3) is quite similar to the other three. The 1969 collection and the 1971 collections (taken as a whole) have seven species in common. She noted that the 1947 collection (Cushman and Todd, 1947) is much richer in variety of foraminiferal species and that only six species which she identified in these more recent collections are the same as those in the 1947 list. All these collections of Foraminifera are from the interglacial beach deposit, and all but one are from about the same horizon, although Todd suggested that the fauna collected in 1969 and 1971 might be from shallower water than those in the 1947 collection.6 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA 0 200 400 600 FEET 1 -----1-----H---------H 0 100 200 METERS EXPLANATION QUATERNARY TERTIARY -----*------------------V --*-- DUNE SAND MARINE DEPOSITS o o o o o o o o AMCHITKA FORMATION ±-zzz-z INTERGLACIAL BEACH DEPOSIT Triangle shows Hydrodamalis locality LACUSTRINE AND FLUVIAL(?) DEPOSITS Figure 4. — Diagrammatic cross section of South Bight graben, Amchitka Island. Faults mainly generalized. Vertical exaggeration x 2. Figure 5. — Partial mandibles of Hydrodamalis gigas, as found on Amchitka Island, upside down in beach sand, anterior end pointing toward lower left corner of photograph. Photograph by F. C. Whitmore, Jr. ENVIRONMENT Hydrodamalis apparently lived in an environment much like that of Amchitkan waters today. According to Cushman and Todd (1947, p. 60) “while the [foraminiferal] fauna is decidedly one of cold waters, it is not by any means arctic, and most of the species found today are in waters to the south of this area. * * * The great abundance of two species of Globigerina, pelagic forms, would indicate that ocean currents from warmer areas, such as the present Japanese Current, also influenced this area.” A comprehensive study by Allison (1973) on the paleoecology of the invertebrate megafossils from the deposit indicates that the water temperature at that time was almost the same as it is today. Allison reported that when these mollusks were living, the mean February water temperature was probably about 3.9°C, and the mean August water temperature was 10.0°—11.7°C; these temperatures compare with present-day water temperatures of 3.9°C for February and 10.0°C for August. Depth of the water in which these fossiliferous beds were deposited could not have been more than 25 m and probably was much less. At least part of the time, this area was within or even slightly above the intertidal zone. A maximum water depth may be assigned by the presence of coralline algae on and between boulders near the base of the deposit. The maximum depth at which sublittoral algae will grow is primarily dependent on the amount of light that will penetrate the water. This, in turn, depends on latitude and water turbidity (Smith, 1938, p. 296). Coralline algae (Rhodophyta) in the North Atlantic Ocean are rarely found at depths below 30 m. At Amchitka Island today, they are found from the intertidal zone to a depth of at least 23 m. The foraminifers Quinqueloculina and Cibicides are known from nearshore deposits and comparatively shallow water (Cushman and Todd, 1947, p. 60). Allison (1973, p. 42) stated that “the majority of the South Bight fauna lived in shallow (0-23 m(?)) water on the open coast, * * * [and] the environment must have been almost identical to that of South Bight today.” The presence of beds of pumice pebbles indicates that the pumice fragments floated to this area, were rounded by wave action, and stranded on the interglacial beach.COMPOSITION OF THE COLLECTION 7 Table 1.— Foraminifera and other fossils associated with Hydrodamalis from the South Bight interglacial beach deposit, Amchitka Island Collections Fossils 1971 1 m above Hydrodamalis horizon1 1971 Upper part of Hydrodamalis horizon1 1969 1 m below Hydrodamalis horizon2 1971 15 m above beach. 0.3 m above base of Hydrodamalis horizon1 Bolivina decussata Brady X X sp. cf. B. pseudopicata Heron-Alien and Earland X Buccella frigida (Cushman) X Cassidulina californica Cushman and Hughes X X islandica Norvang X X teretis Tappan X X tortuosa Cushman and Hughes X Cibicides lobatulus (Walker and Jacob) X X X Elphidiella arctica (Parker and Jones) X Elphidium alaskense Cushman and Todd .... X clavatum Cushman X frigidum Cushman X X X Globigerina sp. aff. G. bulloides d’Orbigny . . . X X pachyderma (Ehrenberg) X X X X Karreriella baccata alaskensis Cushman and Todd X Genus rel. Neoconorbina tabernacularis (Brady) X Oolina borealis Loeblich and Tappan X Planulina alaskensis Cushman and Todd .... X Polymorphina kincaidi Cushman and Todd . . X Pyrgo sp. cf. P. elongata (d’Orbigny) X Quinqueloculina agglutinata Cushman X akneriana d’Orbigny X seminulum (Linne) X X X Rosalina wrightii (Brady) X Rotalia columbiensis (Cushman) X X X X Trichohyalis ornatissima (Cushman) X Triloculine trigonula (Lamarck) X Microscopic gastropods X Echinoid spines (very abundant) X 1 Identification by Ruth Todd. 2 Identification by R. L. Pierce. AGE OF THE DEPOSIT Uranium-series dating of fossil shells and bone from the interglacial beach deposit at South Bight has yielded an average age of 127,000±8,000 years (Szabo and Gard, 1975). A lengthy stillstand of the sea is indicated by the extensive terrace that was cut on the island. Apparently, when the sea reached its maximum height, it rapidly bevelled the underlying semiconsoli-dated sediments in the graben. Continued subsidence of the graben allowed the thick beach deposits to accumulate and to be preserved while the sea persisted in cutting the terrace on much harder Tertiary bedrock elsewhere on the island. Although Allison (1973) presented arguments that this deposit is of Kotzebuan(?) age (pre-Illinoian, according to Hopkins, 1967, p. 50), the radiometric age date of 127,000±8,000 years seems inconsistent with recent fission-track dating of 0.6 m.y. (million years) for ash beds in late Kansan or early Yarmouth deposits in the Western United States (Izett and others, 1970). The radiometric age of the sea-cow bone from the upper middle part of the deposit seems to be reasonably consistent with age dates of the Barbados III sea-level maximum (Mesolella and others, 1969), which has an average radiometric age of 125,000±6,000 years, and with Terrace C in southern California (Szabo and Rosholt, 1969), which has radiometric ages averaging 131,000+15,000 years. These dates are believed by Richmond and Obradovich (1972) to be of late Pleistocene age. COMPOSITION OF THE COLLECTION Gard’s initial discovery was of bone fragments at the base of a cliff. He traced the bones to their source and collected several vertebral centra and parts of a forelimb. These were identified by G. E. Lewis as belonging to Hydrodamalis. Lewis’ identification was corroborated by Whitmore, who compared the bones with specimens of Holocene age, which were collected on Bering Island by Leonhard Stejneger in 1882 and8 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA Figure 6. — Rib of Hydrodamalis (USNM 181752) at base of South Bight II interglacial beach deposit, Amchitka Island. Photograph by F. C. Whitmore, Jr. 1883 and which are now in the collections of the U.S. National Museum of Natural History. In May 1971 Gard and Whitmore uncovered more bones in place in the cliff (fig. 5). The bones found on the cliff face and those uncovered just beneath the surface were hard and well preserved, although not permineralized; they apparently had case hardened upon contact with the air. As we dug farther into the cliff, the bones found were of increasingly poor quality, at first being crumbly, then, about 1 m in from the cliff face, being soft and of a soapy consistency, easily scratched with the fingernail. Stejneger (1884, p. 61—62) found bones similarly preserved on Bering Island. Altogether, the following bones were collected at the Amchitka locality: The horizontal ramus and a small part of the ascending ramus of the left mandible; the anterior half of the horizontal ramus and the angular region of the right mandible; centra of the last three cervical vertebrae and of the first thoracic vertebra; left scapula; distal half of right humerus and distal end and fragments of proximal end of badly preserved left humerus; left and right radius and ulna; rib fragments; and a chevron bone. About half a meter back in the cliff from the mandibles was a large amorphous mass of soft red material. Perhaps this was all that remained of the skull. All these bones came from a single animal that probably had drifted into shallow water after death and had been quickly buried by sand. The left forelimb, including the scapula but excluding carpals and metacar-pals, was in articulation when found; the position of the limbs relative to that of the vertebrae and jaws indicated that the animal was on its back when it was buried. The epiphyseal cartilage of the long bones was still unossified, and none of the articular ends of the long bones was found. The radius and ulna were not coossified as they are in fully adult specimens of the modern Steller’s sea cow, and the mandibular symphysis was only slightly fused. Although the Amchitka specimen was as large as some fully grown Bering Island Hydrodamalis, it was a young individual. These bones were assigned by the U.S. National Museum of Natural History, Department of Paleobiology, the number USNM 170761. In addition to the bones just described, a single rib of Hydrodamalis was found 0.1 m above the base of the interglacial beach deposit, 400 m south of the other specimens (fig. 6). The specimen consists of the distal two-thirds of about a twelfth rib of a very large animal. It was assigned USNM 181752. A small humerus, doubtless from a very young individual, was collected by Gard in August 1972 at the base of the cliff, about 50 m south of the locality of USNM 170761. It was assigned USNM 186807. Also, four fragments of a large sirenian rib were collected at South Bight by Richard C. Allison and deposited in the National Museum by Daryl P. Domn-ing. DESCRIPTION OF THE AMCHITKA HYDRODAMALIS The bones of the Amchitka specimen will be compared with those of specimens collected on Bering Island by Leonhard Stejneger in 1882 and 1883. Stej-neger’s specimens, preserved by the U.S. NationalDESCRIPTION OF THE AMCHITKA HYDRO DA MA LIS 9 Museum of Natural History, are probably the remains of animals killed for food by Russian fur hunters in the 18th century. MEASUREMENTS The system of measurements illustrated in figures 7-9 was designed by Daryl P. Domning for his studies of Sirenia and is used here, with only slight modifications, in the interest of uniformity. Because the Amchitka bones are incomplete, Domning’s system could not be used throughout. MANDIBLE Plate 2, all figures; plate 3, figures 1-4; table 2 The Amchitka mandible was compared with 15 mandibles from Bering Island. Nine measurements were made on each specimen (fig. 7; table 2). All measurements of the Amchitka mandible are within the range of measurements of the Bering Island group. In all but one measurement (“as,” which may be correlated with youthful age, or may simply be due to individual variation), the Amchitka specimen is at or near the upper end of the range of measurements. This may signify that the Pleistocene Hydrodamalis population was of larger average size than the Bering Island population of historic time, for the Amchitka specimen is immature, whereas the condition of the limb bones of the Bering Island collection indicates that almost all those specimens come from adults. Domning (1970, p. 219) advanced the idea that the 18th-century Bering Island population “was reduced to only about 2,000 animals living in marginal environment in the Bering Sea, their growth stunted so that they never reached the size of their Pliocene ancestors in more favorable habitat to the south.” A notable characteristic of Hydrodamalis, in contrast to all other Sirenia, is its lack of teeth. The anterior occluding surface of the rostrum and of the mandible (pi. 2, all figs.; pi. 3, figs. 3, 4) was, as in all modern sire-nians, deflected downward and flattened; it served as a grasping device (analogous to the toothless rostrum of ruminant Artiodactyla) with which the animal tore off the kelp on which it grazed. Steller (1899, p. 186) stated, on the basis of his dissection of a specimen of H. gigas, that mastication was done by “two strong white bones, or solid tooth masses, one of which is set in the palate and the other is fastened in the inferior maxilla, and corresponds to the first.” These bones, Steller said (p. 186), were not fastened in the maxillae and mandibles, but were “held by many papillae and pores.” The bones were said to be perforated below and to have many little holes (p. 186) “in which the arteries and nerves are inserted in the same way as in the teeth of other animals.” The “masticatory bones” to which Steller referred were undoubtedly horny plates, of cutaneous origin, covering the palate and the anterior part of the mandible. Brandt (1846, pi. Ill) illustrated some of these palatal coverings. These plates were certainly hard, and if they were white when dissected out, Steller may have mistaken them for bones. Owen (1838, p. 41) said of the dugong, “the alveoli in the deflected portion of the lower jaw contained ligamentous processes given off from the internal surface of the thick callous integument covering that part of the jaw: they serve the purpose of fixing more firmly to the bone this dense and horny plate, which is beset externally with short coarse bristles, and is doubtless used in scraping and tearing off the seaweeds and other alimentary substances which may be fixed to the rocks.” The Amchitka mandibles are, in all major respects, identical with the modern ones from Bering Island. The symphysis of the Amchitka mandible is about 25 percent fused, evidence of the immaturity of the specimen, despite its large size. In profile, the anterior border of the mandible is bluntly rounded (pi. 2, figs. 1, 2), in contrast to the pointed anterior profile of most Bering Island mandibles (pi. 2, figs. 3, 4). The fossa for the genioglossus muscle (pi. 3, fig. 3; text-fig. 7) always large in Hydrodamalis, is especially large in the Amchitka mandible. This muscle functions in protrusion and retraction of the tongue. The posterior border of the Amchitka mandible is thickened for 80 mm above the mandibular angle; at the angle (“d” in fig. 7 and table 2) it is 14 mm thick. It is thinner for 70 mm above the angle and thicker again (attaining a maximum thickness of 19 mm) for a distance of 80 mm on the uppermost part of the specimen as preserved (pi. 3, fig. 2; table 2). The preserved angular region of the right mandible is broken at the base of the ascending ramus. The posterior thickening of the angular region of the Amchitka specimen is in contrast to this region in adult Bering Island specimens. In the latter, the posterior border of the mandible thins almost to a knife edge ventrally and thickens only slightly at the angle. The posterior thickening of the jaw of the Amchitka specimen resembles more closely that in a young specimen from Bering Island (USNM 218381; table 2) and the Miocene genus Metaxytherium. By contrast another immature mandible from Bering Island (USNM 218401; table 2) has only slight thickening (9 mm) at the angle and thus resembles adult mandibles in this respect. Many of the adult Bering Island specimens have a wrinkled surface on the posteromedial side of the ascending ramus (pi. 2, fig. 4). This contrast in configuration of the posterior edge of the mandible probably reflects differences in the areas of insertion of the medial pterygoid and masseter muscles. These mus-10 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA Figure 7. — Key to measurements of mandibles of Hydrodamalis. (See table 2.) cles raise the jaw and have great power in keeping the jaws clenched. Despite the immaturity of the Amchitka specimen, the maximum thickness of its posterior mandibular border is greater than that of all but one Bering Island specimen (table 2). The mental foramen is a prominent opening 10 mm high (pi. 2, fig. 1), situated 235 mm posterior to the anterior end of the mandible and approximately 50 mm anterior to the base of the ascending ramus. The base of the foramen is 63 mm above the ventral side of the mandible; its top is 31 mm below the dorsal side. The groove anterior to the mental foramen is about 65 mm long, widening gradually in an anterior direction. Because of breakage, the posterior extent of the mandibular canal cannot be determined exactly, but enough of the lateral wall of the mandibular foramen is present to show that its position and shape were approximately the same as that of the Bering Island specimens. VERTEBRAE Plate 3, figures 5 -8; plate 4, figures 1 -8; table 3 Only the centra of the last three cervical vertebrae and the first thoracic vertebra were found. They have the dorsoventrally flattened, subrectangular shape characteristic of the Sirenia. Daryl P. Domning (written commun., 1971) has suggested, on the basis of his studies of North Pacific sirenians, that there is a tendency for the width-to-height ratio of cervical centra to decrease in the course of their evolution. This trend is corroborated by comparison of the Amchitka specimen with Holocene specimens from Bering Island. Perhaps this trend can be correlated with increased lateral flexibility of the neck, associated with browsing in shallow water, as described by Steller. Caution must be exercised, however, in drawing conclusions from measurements of cervical centra in Hydrodamalis, for the boundary between the centrum andDESCRIPTION OF THE AMCHITKA HYDRODAMALIS 11 Table 2. — Measurements, in millimeters, of mandibles of Hydrodamalis gigas [See fig. 7 for key to measurements; leaders (...) indicate no measurement was made; entries followed by “e” indicate measurement is estimated] Amchitka Bering Island specimens (L. Stejneger colln., 1882 -83)' specimen Measurement USNM 2170761 21255 21260 21262 ‘21266 21269 218371 218377 1 00 CO 00 N 218399 218400 | O oo 218402 218403 218408 269064 Mounted composite skeleton, USNM as 142 160 165 148 144 139 145 175 140 143 124 167 152 166 175 191 165 rr' 71 66 62 67 57 61 72e 68 44e 50 55e 66 58 62 60e 72 64 aj 167 149 150 148 122e 125 123 163 110 145 100 166 141 155 167 159 173 ab 436e 418 425 410e 404 362 428 460 352 382 380e 419 428 430e 438 445 428 ad 361 340 349 345 340 345 355 385 314 312 285 334 354 380 363 374 359 ef 166 157 155 167 120 156 163 136 127 127 130 151 165 141 160 160 135e df 278 241 243 250 264 248 263 280 236 222 215 240 280 293 284 260 272 mo 90 92 76 86 82 90 86 91 81 67 82 89 95 92 82 103 81 ap 240 209 219 223 197 191e 221 235 200 206 185 215 223 217 237 222 206 Thickness at“a” 14 16 10 8.3 11 11.4 11 12 13 11 9 12 10 10.5 8 8 Maximum thickness of posterior border .... 19 16 10.7 17.3 12 11.4 14 12.8 16.5 13 14 12 13.5 12 20 17 1 Numbers are from the Division of Mammals, U.S. National Museum of Natural History. 2 Young specimen: symphysis not completely fused Number from U.S National Museum of Natural History. Department of Paleobiology. Note low symphysis. the transverse process is poorly defined and probably varies in position with age. In the sixth cervical vertebra, the lower half of the left transverse foramen is present (pi. 3, fig. 7). It is approximately 14 mm wide. Examination of specimens from Bering Island showed the size of the transverse foramen to be variable, even on two sides of the same individual. On the sixth cervical vertebra of USNM 218808, an adult specimen from Bering Island (pi. 3, fig. 8), the left transverse foramen is 11 mm wide, and the right is 14 mm wide. The posterior width of the centrum in this specimen is 135 mm, compared with 140 for the Amchitka specimen (table 3). The seventh cervical vertebra of the Amchitka specimen has a small facet for rib articulation on its posterior face, on the ventral side of the base of the transverse process (pi. 4, fig. 3). The entire left transverse foramen is present (pi. 4, fig. 2); as in the Bering Island specimens, it is much smaller than the corresponding foramen in the sixth vertebra, being 5.5 mm wide and 2.8 mm high. The first thoracic vertebra has both anterior and posterior demifacets for articulation of the capitulum of the ribs (pi. 4, fig. 6). The anterior demifacet is much smaller than the posterior one. SCAPULA Plate 5, figures 1—3; plate 6, figure 3; plate 7, figure 4; table 4 As in all Sirenia, the scapula of the Amchitka Hydrodamalis is fan-shaped — that is, very wide in its dor- sal (vertebral) part and having a narrow neck just above the glenoid cavity. The result of these dimensions is that the prescapular and postscapular fossae, in which originate the muscles used in rotating the limb, are smaller than in most mammals. By contrast, the vertebral part of the scapula is wide, its posterior part being especially well developed. The anterior border of the scapula in Hydrodamalis is almost straight, with a slight angulation about halfway up its height, in contrast to that of Metaxytherium and the modern Sirenia, which is strongly curved. The gently curved edge dorsal to the angulation in Hydrodamalis may be the anterior edge of the insertion of the serratus cervicis muscle. The straight posterodorsal edge of the scapula (fig. 8, tmo) is the origin of the teres major muscle, which flexes the shoulder joint and abducts the arm. A branch of the serratus muscle was inserted on the costal side of the scapula medial to the teres major. The prescapular fossa is small (fig. 8, ek; table 4), much smaller than in Metaxytherium (cf. Kellogg, 1966, pi. 43, figs. 1, 2) and somewhat smaller than in the modern Dugong dugon and Trichechus manatus. The postscapular fossa is of the same width as the prescapular (fig. 8). The scapular spine in the Amchitka specimen is restricted to approximately the ventral half of the scapula (pi. 5, fig. 2; pi. 6, fig. 3). Spines in modern specimens from Bering Island occupy, on the average, 6012 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA Table 3. — Measurements, in millimeters, of vertebrae of Hydrodamalis gigas [USNM, U.S. National Museum of Natural History; UCMP, University of California, Museum of Paleontology; UCMP measurements by D. P. Domning; leaders (...) indicate that no measurement was made; entries followed by “e" indicate that measurement is estimated; an asterisk * indicates front and back reasurements respectively) Amchitka specimen Bering Island specimens Bone measurement 'USNM 170761 2USNM 22182 2USNM 218807-218810 UCMP 23050 Mounted composite skeleton, USNM Fifth cervical Anteroposterior thickness of centrum 42 31 (USNM 218807) 35 32 Width of centrum . . *140/140 *102/108 *118/118e Height of centrum at middle 78 71 82 77e Width/height of centrum 1.79 *1.43/1.52 1.42 Sixth cervical (USNM 218808) Anteroposterior thickness of centrum 41 36 43 39e Width *130/140 *116/110 *127/135 120e Height Width/height of 80 73 86 64e centrum *1.62/1.75 *1.59/1.50 *1.46/1.56 1.89 Seventh cervicaJ Anteroposterior (USNM 218809) thickness of centrum 45 31 41e 56e Width *141/152 *109/110 *136/. .. *127/123 104e Height Width/height of 78 75 83 66 63e centrum *1.80/1.94 *1.45/1.46 1.63 *1.92/1.86 1.63 First thoracic (USNM 218810) Anteroposterior thickness of centrum 46 59 58 Width *145/157 *147/139 *119/123 104e Height Width/height of 78/72 74/85 74 65e centrum *1.85/2.18 *1.98/1.63 *1.60/1.64 1.60 1 Department of Paleobiology, U.S. National Museum of Natural History. 2 Division of Mammals, U.S. National Museum of Natural History. percent of the scapular height. The lower half of the HUMERUS spine (acromion process) of the Amchitka specimen is expanded and heavy, much more so than in the Bering Island specimens. The tip of the acromion process has been broken off; in Holocene specimens from Bering Island (pi. 7, fig. 4) it is less well developed than in Metax-ytherium (cf. Kellogg, 1966, pi. 43, figs. 1, 2). The glenoid fossa and coracoid process are absent from the Amchitka specimen because the epiphyseal cartilage was not ossified. The scapula of the Amchitka specimen is noticeably thicker and heavier, especially in the dorsal part of the blade (pi. 5, fig. 3), than are those in the Stejneger collection at the U.S. National Museum of Natural History. It is slightly narrower relative to its length than is any Bering Island specimen, but in length and width it is within the size range of that collection. In all Hydrodamalis specimens observed in this study, the scapula is heavier than in modern Sirenia. Plate 4, figures 9, 10; plate 6, figures 4, 5; plate 7, figure 1; plate 8, figure 1; table 4 The distal half of the right humerus of USNM 170761 (pi. 6, figs. 4, 5) was found in a good state of preservation, although it lacks the distal articular surface because the epiphysis had not yet ossified. The left humerus was also present but had deteriorated to such an extent that no measurements or morphologic observations could be made. In addition, the right humerus of a smaller individual (USNM 186807; pi. 4, figs. 9, 10), collected at the foot of the cliff, has open eiphyses and lacks both proximal and distal articular facets. The anterior part of the proximal end of the bone is present, however, allowing measurement of the length of the shaft (245 mm). Although much smaller than USNM 170761, this bone is as heavily constructed. A prominent feature of the humerus of Hydrodamalis is the large shield-shaped deltoid tuberosity (pi. 7, fig.DESCRIPTION OF THE AMCHITKA HYDRODAMALIS 13 a Figure 8. — Key to measurements of scapula and humerus of Hydrodamalis.iSee table 4.) 1). The part of the bone bearing the tuberosity is not present in USNM 170761; in the smaller humerus (USNM 186807), it is rounded and smaller relative to the size of the bone. This is probably due to the immaturity of the individual and possibly also to erosion of the specimen. RADIUS AND ULNA Plate 8, figures 2-6; text-figure 9 The forearm bones of both limbs were collected. Those of the right side are well preserved, whereas those of the left were soft and of soapy consistency. The left radius and ulna were collected in a plaster block and subsequently hardened with beeswax; however, they are so badly cracked that measurement was impractical. In all the bones, the epiphyses were open and the articular surfaces were missing. In adult Holocene specimens from Bering Island (pi. 8, fig. 6), the radius and ulna are strongly fused together at their proximal and distal ends and have scattered small areas of fusion in the middle part of their shafts. By contrast, the radius and ulna of the Amchitka specimen are separate, undoubtedly a function of the youth of the animal (pi. 8, fig. 5). Near the proximal end of the anterior surface of the radius of the Amchitka specimen is a single tuberosity (pi. 8, figs. 2, 5). In Dugong, such a tuberosity serves for insertion of the brachialis muscle (D. P. Domning, written commun., 1973). This tuberosity is double in some observed Holocene specimens from Bering Island (von Nordmann, 1863). RIBS Plate 5, figure 4; plate 6, figures 1, 2; text-figure 10 The rib fragments of USNM 170761 are from the anterior region of the thorax. Four pieces show characteristics worth recording. One is a fragment of the proximal end of an anterior rib (probably the first or second), proximally compressed (thickness about 18 mm) and expanding distally to an oval cross section, which measures 70 mm high by 40 mm wide at a distance of 115 mm from the articular surface. The capitulum of this rib is preserved and consists of a small articular surface, spongy in appearance, as would be expected in a young specimen, and measuring 28 mm high by 15 mm wide. The tuberculum has been broken off. The second fragment is the capitulum of an anterior rib, probably the second or third of the left side. Its articular surface, also spongy (pi. 5, fig. 4), consists of two facets, intersecting at an angle of about 120°, that articulated with the demifacets of two of the thoracic vertebrae. The articular surface has a maximum height of14 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA Table 4. — Measurements, in millimeters, of forelimb of Hydrodamalis gigas (See figs. 8 and 9 for key to measurements. Bering Island specimens having the same number are not necessarily from the same individual. USNM, U.S. National Museum of Natural History; Museum numbers of Bering Island specimens are assigned by USNM Division of Mammals; Museum numbers of Amchitka have been assigned by the USNM Department of Paleobiology; leaders (. ..) indicate that no measurement was made; entries followed by “e” indicate that measurement was estimated; (right) or (left) indicates side of skeletonl Bone measurement Amchitka specimens Bering Island specimens Scapula USNM 170761 (left) USNM 269193 USNM 218828 USNM 35638 USNM 218409 Mounted composite (right)1 (right) (right)2 (left)2 skeleton, USNM (right) ab 513 481 570e 574 620 597 cd 338 350 412 414 430 424 ef 104 120 126 130 145 118 ek 60 110 75 120 117 gh 143 153 214 203 223 195 bi 375 77 125 117 112 120e bj 108e 102 140 139 177 138 kl 110 118 117 133 139 123 tmo 119 146 152 154 197 JP 200e 220 196 221 198 207 ap 250 183 240e 230 263 291 USNM USNM Humerus 170761 186807 USNM 35638 USNM 35638 USNM 21251 Mounted composite skeleton, USNM (right) (right) (right) (left) (left) (left) (right) ef 183 130 225 211 214 207 185 ij 47 56 81 131 3122 113 102 ab 245 534 521 545 489 454 USNM 170761 USNM 21251 USNM 218380 USNM 218380 USNM 218371 USNM 218393 Radius-ulna (right) unfused4 (right) fused (right) ulna2 (left) proximal (right) partial ulna (left) fused half of ulna2 ab 517 294 517 ag 486 cd 372 250 374 Cg 253 324 gh 70 87 62 ij 121 158 99 107 144 159 mn 62 86 65 91 99 op 65 76 66 62 72 93 rt 369 (?) qr 72 88e 92e st 73 79e 91e UV 103 155e 159 WX 63 68 69e USNM 218394 USNM 218414 USNM 218415 USNM 218415 Mounted composite Radius-ulna — (right) ulna2 (left) fused (left) radius (right) radius skeleton, USNM Continued proximal half4 (left) fused ab 506 ag 421 604 cd 603 Cg 345e 316 gh 150 ij i33e 151 149 mn lOOe 80 95 op 65e 73 79e rt 409 344 363 qr 95e 72 73 77e st 63e 81 49 UV 139 145 139 50 WX 57e 73 73 1 Suprascapular cartilage ossified. 3 Eroded. 2 Immature, but epiphyses attached. 4 Epiphyses missing; measurement less than original length of bone. 55 mm and width of 39 mm. The surviving proximal part of the rib shaft is extremely compressed, being 15 mm thick by approximately 80 mm wide. The other two rib fragments lack the proximal and distal ends. They do not match the broken ends of the two fragments just described. One has a slight angulation in the form of a low ridge on its convex side; the very gentle curvature of this bone indicates that it is the second or third rib, although it is heavier than such ribs in the Stejneger collection from Bering Island. It isDESCRIPTION OF THE AMCHITKA HYDRODAMALIS 15 Figure 9. — Key to measurements of radius and ulna of Hydrodamalis. (See table 4.) from the right side. The fragment is 170 mm long, 76 mm wide at the angulation, its widest (mesolateral) dimension, and 50 mm thick anteroposteriorly at the same point. The smaller (ventral) end of the rib fragment presents a spongy appearance, indicating that it was near the true end of the bone, where there was a cartilaginous connection with the sternum. At this point the rib is tapering rapidly, having a diameter of 50 mm mesolaterally and 39 mm anteroposteriorly. The other fragment is probably from the dorsal part of the third rib of the left side. It is more gently curved and heavier than the first fragments and is 150 mm long. At its proximal end, its cross section has the shape of a triangle with rounded angles; the apex points anteriorly. From base to apex, it measures 60 mm; the length of the base is 65 mm. The distal end of the fragment is oval in cross section. It measures 48 mm anteroposteriorly and approximately 60 mm mesolaterally. A more nearly complete rib fragment (USNM 181752; pi. 6, fig. 1; text-fig. 10) the distal two thirds of approximately the twelfth rib of the left side, was found in place 0.1 m above the base of the deposit 15 m above the beach and about 9 m stratigraphically below the level at which the Hydrodamalis skeleton was found. The bone is heavily oxidized, reddish brown, and was very brittle when collected, in contrast to the white color and hardness of bones found near the surface at the top of the cliff (fig. 6). The rib is evenly curved, having been broken distal to the angulation (fig. 10). It is ovoid in cross section; the greatest thickness is anteroposterior. Cross-section measurements are as follows (fig. 10): No. 1, 98x49 mm; No. 2, 84x56 mm; No. 3, 80x57 mm; No. 4, 72x52 mm; No. 5, 62x50 mm. The distal end of the rib has a relatively smooth area running anteroposteriorly and a coarsely pitted area on either side of it. In cross section, where broken at its proximal end, the rib has cancellous tissue in its outer posterior quarter; the rest of the cross section shows pachyostotic bone. This rib is within the size range of modern ribs from Bering Island. The rib from the Kangiguksuk archeological site (Yale Peabody Mus. No. 233862) is an anterior rib from the left side. The sawed distal end shows it to be pachyostotic except for an area of cancellous bone that runs mediolaterally through the middle of the bone. Such presence of cancellous tissue is typical of young Sirenia; in adult animals the bone is fully pachyostotic, and there is no trace of cancellous tissue. The distal end of the rib has been sawed off; as preserved, the specimen measures 582 mm along the outside curve. At the sawed distal end it is suboval, being slightly narrower at the outer than at the inner side. In cross section it measures 52 mm transversely and 33 mm anteroposteriorly. At the outer extremity of the curve (approximately at the position of cross-section 3 of fig. 10), the rib measures 46.5 mm transversely and 37 mm anteroposteriorly. At the proximal end of the curve, immediately distal to the tuberculum, it measures 43 mm transversely and 36 mm anteroposteriorly, having been reduced slightly in circumference to form the neck of the rib. The proximal end of the rib, in the area of the capitulum and tuberculum, is anteroposteriorly compressed and gently curved, being concave anteriorly and convex posteriorly. The distance from capitulum to tuberculum is 101 mm; midway in16 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA Figure 10. — Rib of Hydrodamalis (USNM 181752) from base of South Bight II interglacial beach deposit, Amchitka Island. (See also fig. 6.) this distance, the rib is 21 mm thick anteroposteriorly and 45 mm dorsoventrally. The capitular articulation is very small, 17 mm dorsoventrally by 18 mm anteroposteriorly. This end of the rib has been worked or gnawed (it is marked by many dorsoventral scratches); probably, the capitulum originally had a somewhat larger articular area. The tuberculum is a low anteroposteriorly compressed ridge. The Kangiguksuk rib appears to come from a more lightly built animal (possibly because of its youth) than those represented on Bering Island. CHEVRON BONE Half a chevron bone, probably from the right side of the animal, was recovered (pi. 7, figs. 2, 3). The chevron bone, named from its shape in cross section, forms the hemal arch below the centra of the caudal vertebrae. Through the V of the hemal arch run the major blood vessels of the tail. There are no chevron bones of Hydrodamalis in the collections of the U.S. National Museum, but hemal arches of Holocene Dugong, a member of the same family as Hydrodamalis (Dugongidae Gray, 1821), are for the most part composed of paired, separate bones in contrast to the condition in the Cetacea, whose hemal arches almost invariably consist of a V-shaped bone, resulting from fusion of the bones of the two sides. The chevron bone of Hydrodamalis from Amchitka measures 66 mm in greatest vertical dimension, 42.5 mm anteroposteriorly on its dorsal side, and 36.5 mm anteroposteriorly at the level where it starts to taper ventrally to a rounded posteriorly slanting edge. The size of the bone indicates that it was associated with one of the posterior caudal vertebrae. MORPHOLOGIC CONCLUSIONS The Amchitka skeleton shows no significant morphologic differences from 18th-century specimens from Bering Island. Its mandible is larger than would be expected in the Bering Island population, especially when one considers that it was immature at death; the other Amchitka bones, however, are farther down in the size range of the Bering Island collection. This disparity in ratio between mandible size and dimensions of postcranial bones may be because of the youth of the Amchitka specimen, or it may be because the Pleistocene Hydrodamalis had a larger head relative to the body than did members of the 18th-century Bering Island population. D. P. Domning (written commun., 1973) has pointed out that in the transition from Metax-ytherium to Hydrodamalis, body size increased perhaps 100 percent, and skull length, only about 40 percent. Ratios between mandible size and that of postcranial bones in the Bering Island population of HydrodamalisDESCRIPTION OF THE AMCHITKA HYDRODAMALIS 17 (U.S. National Museum collection) are of questionable significance because the collection consists of bones picked up at random; it does not include any associations of bones from one individual. The only description known to us of a skeleton of a single individual of H. gigas from Bering Island was published by von Nordmann (1863). This skeleton, in the zoological museum at Helsinki, Finland, is that of an immature animal, which facilitates comparison with the Amchitka specimen (table 5). In von Nordmann’s specimen, the epiphyses of the scapulae, the humerus, the ulna, and the radius had been lost, being separate from the shafts of the bones. The skeleton is 5 m long, as compared with an estimated length of 7.6 m for an adult Bering Island specimen. Von Nordmann’s specimen and that from Amchitka appear to have been near enough to the same ontogenetic age that comparisons between them will have some validity; caution is dictated, however, because the size of the adult Pleistocene Hydrodamalis is unknown. As an index of the relation of head size to body size, ratios were established between the length of the jaw and various measurements of postcranial bones in the von Nordmann and Amchitka specimens (table 5). In many of these ratios no significant difference exists between the two specimens; where a difference does exist, the postcranial measurement is smaller relative to jaw length in the Amchitka specimen than in that described by von Nordmann. This relationship fits Domning’s hypothesis that body size increased relative to head size in the evolution of the Metaxy-therium-Hydrodamalis line. The jaw described by von Nordmann is much shorter than the Amchitka jaw and shorter than the mean of jaws from Bering Island. It is also smaller in other dimensions than are the Bering Island jaws, the exception being the width of the symphysis, in which it measures 65 mm compared with a mean of 61 mm for the Bering Island specimens. With three exceptions, the postcranial bones of the von Nordmann specimen are smaller than the mean for the Bering Island bones. The exceptions are the thickness of the humerus at its lower end and the diameter of the lower ends of the radius and ulna. The small size of these measurements of the Bering Island specimens may result from erosion of the ends of the limb bones, which is evident upon inspection. However, no such cause can account for the fact that, in these measurements, the von Nordmann specimen is larger than the Amchitka specimen, in which the bones concerned are well preserved. The specimen described by von Nordmann has a larger jaw relative to postcranial bones than do the specimens in the mixed collection from Bering Island, on the basis of mean measurements of bones. In this Table 5. — Comparison of measurements of the Amchitka specimen of Hydrodamalis gigas with measurements of an immature specimen from Bering Island, measured by von Nordmann (1863) [Measurements given in millimeters, “e" following value indicates estimated! Amchitka von Nordmann specimen specimen First thoracic vertebra: Length of body at base ..................... 46 41 Height of body............................ 78 71 Greatest width of vertebral foramen . 126e 106 Scapula: Length from end of collum to middle of upper border .......................... 513 455 Greatest width of upper part............. 338 377 Width of neck at border of epiphysis 143 152 Thickness of neck in middle ............. 104 85 Greatest height of spine.................... 67 75 Humerus: Greatest width, lower end ............... 183 177 Greatest width in middle.......... 86 98 Greatest thickness at lower end.... 61 82 Greatest thickness in the middle .... Ill 102 Ulna: Length of bone without epiphysis.... 320 320 Greatest width of upper end...... 121 112 Greatest diameter in middle....... 60 62 Greatest diameter of lower end ........... 71 78 Radius: Length of bone without epiphysis .... 265 280 Greatest width of upper end................ 105e 113 Greatest diameter in middle....... 57 65 Greatest diameter of lower end ........... 66 81 Jaw: Length from most posterior border of angle to point of symphysis...... 395 374 Greatest height of body of jaw at posterior border of foramen max- illare [= mental foramen]....... 92 84 Grestest height of body of jaw in front of foramen maxillare .................. 144 140 Length of symphysis of both jaw halves......................... 143 146 Greatest width of symphysis....... 71 65 Ratios: Jaw length/length, first thoracic vertebra ..................................... 8.58 9.12 Jaw length/height, first thoracic vertebra ..................................... 5.06 5.26 Jaw length/scapula length................... .77 .80 Jaw length/scapula width .................. 1.13 .99 Jaw length/width of humerus, lower end...................................... 2.15 2.11 Jaw length/width of humerus in middle ....................................... 4.59 3.81 Jaw length/humerus: greatest thickness, lower end ........................... 6.47 4.56 Jaw length/humerus: greatest thickness in middle............................. 3.55 3.67 Jaw length/ulna: width of upper end 3.26 3.33 Jaw length/ulna length..................... 1.23 1.16 Jaw length/ulna: diameter of lower end...................................... 5.56 4.79 Jaw length/radius length................... 1.49 1.33 characteristic, the von Nordmann specimen resembles the Amchitka specimen. The allometry shown by these two animals is probably due in part to their youth. The large size of the Amchitka jaw, however, may indicate,18 STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA in addition, that the population from which it came consisted of larger individuals than those constituting the 18th-century population on Bering Island. The larger size of the Amchitka animal indicates that size reduction took place during Pleistocene time within the species H. gigas. One vexing problem of the anatomy of Hydrodamalis was unfortunately not solved by the Amchitka find. This problem is the morphology of the wrist and manus. Steller (1899, p. 188) described the anterior extremity of Hydrodamalis as terminating “bluntly with tarsus and metatarsus [sic]. There are no traces of fingers, nor are there any of nails or hoofs; but the tarsus and metatarsus are covered with solid fat, many tendons and ligaments, cutis and cuticle, as an amputated human limb is covered with skin. But both the cutis, and especially the cuticle, are much thicker, harder, and drier there, and so the ends of the arms are something like claws, or rather like a horse’s hoof; but a horse’s hoof is sharper and more pointed, and so better suited to digging.” No carpal bones of Hydrodamalis have been identified. The skeleton mounted in the Naturhistoriska Riksmuseet in Stockholm (pi. 1) has a bone mounted in the position of a metacarpal, which Domning (in press) has identified as the transverse process of a vertebra. DISCUSSION Hydrodamalis gigas, approximately 127,000 years ago, was established on Amchitka Island and was characterized by the same large size and lack of teeth that distinguished the population that became extinct on Bering Island in the 18th century. Remains of three individuals were found within a few hundred meters in the cliffs at South Bight on Amchitka. The South Bight exposure, preserved in a graben bordered and protected by early Tertiary rocks, is a rare occurrence of late Pleistocene interglacial deposits in the Aleutians; the abundance of Hydrodamalis gigas in this limited exposure suggests that the species may have been widely distributed in the Aleutians at that time. A former wide distribution of Hydrodamalis is supported by other, although scanty, paleontologic evidence — the occurrence of (1) a skull in Monterey Bay, Calif. (Jones, 1967); (2) representatives of the genus in the Pliocene of California (Domning, in press); (3) a rib in the Pliocene of Japan (Shikama and Domning, 1970); and (4) a rib in an archeological site in northwestern Alaska (Hall, 1971). Further evidence pointing to a low-latitude origin of Hydrodamalis is the distribution of fossil as well as modern Sirenia. The presence of Hydrodamalis in the Pliocene and Pleistocene of California emphasizes the question of how the genus, unlike any other representatives of its order, achieved adaptation to life in cold water. Proba- bly this adaptation began in California latitudes; Durham (1950) pointed out that by middle Pliocene time the 20°C marine isotherm, as a result of a cooling trend, had approached its present position near the southern tip of Baja California. Addicott (1969) analyzed the distribution patterns of Tertiary shallow-water molluscan faunas and, although detecting a middle Miocene warming trend, corroborated a decline in the temperature of California coastal water in later Miocene and Pliocene time. Turning to Pleistocene temperatures of the California coast, D. M. Hopkins (written commun., 1972) stated that subarctic mollusk faunas have been collected from the floor of Monterey Bay, “containing, most notably, Astarte benneti, a shallow-water mollusk that now ranges from Puget Sound to Point Barrow.” As Hopkins pointed out, however, we do not know whether the Arctic mollusks are of the same age as the dredged Hydrodamalis skull from Monterey Bay. It seems certain that the Commander Islands must have had a Quaternary history similar to that of the Aleutians and that Hydrodamalis once ranged along the Aleutian chain as well. Whether the sea cow lived in the Aleutians in Holocene time is, however, unknown. If Hydrodamalis lived throughout the Aleutians in Holocene time, its presence may have provided incentive for a rapid westward migration of the Aleut, who, like the Russians, would have found that these animals were easily caught and were excellent food. This speculation is supported by Martin’s (1973) intriguing proposal of a spectacularly rapid advance of hunting man throughout the Western Hemisphere 11,500 to 10,500 years ago. The Aleuts’ westward expansion could account for the disappearance of these animals from the Aleutian Islands (Domning, 1972, p. 188). The Aleuts never reached the Commander Islands (Laughlin, 1967, p. 444), and Hydrodamalis survived there until the 18th century. That bones of Hydrodamalis have not been found in Aleut middens may be because not enough middens in the Aleutians have been excavated, or perhaps because the bones simply have not been recognized; or it may be that the middens that have been explored postdate the extinction of Hydrodamalis in the Aleutians. Aleuts were present on Umnak as much as 8,000 years ago (Laughlin, 1967), and, as Domning speculates, middens where the bones might be found are now mainly below sea level. The presence of Hydrodamalis in the late Pleistocene of Amchitka suggests, despite the presence of icecaps on some, if not all, of the Aleutian Islands, that the water temperature continued to be warm enough during Wisconsinan time to support kelp (which is presently the predominant plant element of Arctic and Antarctic seas, according to Smith, 1938, p. 220) and, in turn, to support the sea cow.STELLER’S SEA COW, AMCHITKA, ALEUTIAN ISLANDS, ALASKA 19 REFERENCES CITED Addicott, W. O., 1969, Tertiary climatic change in the marginal northeastern Pacific Ocean: Science, v. 165, no. 3893, p. 583-586. Allison, R. C., 1973, Marine paleoclimatology and paleoecology of a Pleistocene invertebrate fauna from Amchitka Island, Aleutian Islands, Alaska: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 13, no. 1, p. 15-48. Baer, K. E. von, 1840, Untersuchung iiber die ehemalige Verbreitung und die ga’nzliche Vertilgung der von Steller beobachteten nor-dischen Seekuh (Rytina III.): Acad. Imp. Sci. St. Petersbourg Mem., ser. 6, v. 5, pt. 2, p. 53-80. Brandt, J. F., 1846, Symbolae sirenologicae, quibus praecipue Rhytinae historia naturalis illustratur: Acad. Imp. Sci. St. Petersbourg M£m., s£r. 6, v. 7, pt. 2, p. 1-160. ______1868, Symbolae sirenologicae. Fasc. II et III: Acad. Imp. Sci. St. Petersbourg Mem., s0.0625 mm <0.0625 mm >0.0625 mm 0.0625 0.088 0.125 0.177 0.250 0.350 0.500 5-1-67. 1330 8.55 159 249 118 184 1.0 7.0 29.8 63.8 91.8 100 100 5-2-67. 2000 '10.2 171 254 151 224 1.3 8.0 30.2 61.1 94.9 100 100 5-3-67. 1420 10.6 133 274 122 250 0.3 5.6 23.3 58.2 87.7 100 100 5-6-67. 1115 8.38 1030 208 747 151 2.3 12.0 39.2 70.2 97.9 100 100 5-16-67 1111 6.37 443 166 244 91.6 3.3 24.9 52.0 70.1 95.0 100 100 1 Estimated from stage-discharge relations. tions in the study reach on May 8, June 6, and July 14 for the determination of the distributions of the fluorescent tracer particles. Size distributions of the measured suspended sediment, measured suspended-sediment concentrations, and measured suspended-sediment discharges are presented in table 7. Size distributions were determined by the visual accumulation-tube procedure (U.S. Inter-Agency Committee on Water Resources, 1957b). The suspended-sediment discharge was also measured on May 8, the last day of the injection process. However, the median diameter of the sampled sediment was 0.231 mm, which is more characteristic of bed material than of suspended sediment. Also, the measured suspended-sediment discharge was about three times larger than any of the other measured discharges. Hence, it seems probable that the sampler inadvertently was pushed into the bed during sampling. Data from the May 8 measurement are not included in table 7. The measured suspended-sand discharges (>0.0625 mm) ranged from 91.6 to 250 t/d (101 to 276 tons/d) and averaged 180 t/d (199 tons/d). The median fall diameter of the measured suspended sand ranged from 0.122 to 0.162 mm, averaged 0.146 mm, and had a mean geometric standard deviation of 1.47. The geometric standard deviation, cr, is defined as cr = V2(rf84/rf50 + rf50/rf16), where d is the diameter for which 84, 50, or 16 percent, respectively, of the size distribution, by weight, is finer. The mean discharge of sediment finer than 0.0625 mm was 276 t/d (304 tons/d). The large variations in the discharge are assumed the result of variations in the upstream supply of this sediment. Size distributions and parameters of the distributions of the bed-material samples obtained with the 0.1-m-long (4-in-long) core-sampler are presented in table 8. These distributions were determined by the visual accumulation-tube procedure (U.S. Inter-Agency Committee on Water Resources, 1957b). The mean median fall diameter of the three bed-material samples was 0.242 mm, and the mean geometric standard deviation was 1.34. Mean values of the rflg, rf50, d84, and cr parameters of the size distributions of bed-material samples obtained with the 0.6-m-long (2-ft-long) core sampler are presented in tables 9 and 10 for the May 8 and the June 6, July 14 samples, respectively. Not all segments ofHYDRAULIC AND SEDIMENT MEASUREMENTS 11 Table 8. — Size distributions and parameters of the distributions of bed material as determined by the visual accumulation-tube procedure; 0.1- m core samples Date Time (24-hour clock time) Percentage finer than size (in millimeters) indicated q (mm) dR4 (mm) cr 0.0625 0.125 0.177 0.250 0.350 0.500 1.00 c/|g (mm) 5-1-67 1313 0.2 1.0 9.9 57.3 88.0 95.8 99.4 0.188 0.240 0.329 1.32 5-8-67 1110 .5 1.2 9.8 52.4 83.9 93.7 98.7 .189 .247 .351 1.36 5-16-67 .... 1140 .4 1.9 12.0 57.1 88.7 97.9 100 .183 .239 .326 1.34 Table 9. — Mean parameters of the size distributions of bed material from various cross sections as determined by sieve analysis; 0.6-m core samples, May 8 Cross section Cores in which all segments were sieved Cores in which all segments were not sieved Number of segments (n) (mm) clgQ (mm) cfg4 (mm) cr Number of segments (/i) rfjg (mm) rfgQ (mm ) dg4 (mm) 7.5 40 0.213 0.310 0.492 1.52 51 0.212 0.307 0.488 1.52 15 45 .211 .310 .508 1.56 73 .206 .301 .492 1.55 30 30 .210 .310 .512 1.56 45 .206 .298 .492 1.55 60 10 .207 .299 .471 1.51 45 .197 .289 .460 1.53 90 15 .207 .292 .449 1.48 45 .197 .291 .460 1.53 180 10 .201 .288 .462 1.52 42 .202 .291 .453 1.50 270 20 .202 .289 .453 1.50 47 .199 .289 .458 1.52 360 10 .203 .290 .473 1.53 40 .189 .277 .433 1.51 450 35 .196 .293 .528 1.65 47 .194 .289 .492 1.60 540 5 .203 .304 .549 1.65 38 .198 .290 .483 1.56 Composite 220 .207 .301 .492 1.56 473 .200 .295 .473 1.53 the May 8 samples were sieved because some of the bottom segments contained no fluorescent particles. Hence, the samples are divided into two categories, one in which all five segments of the core were sieved and one in which one or more of the segments were ,ot sieved. The number of segments, n, sieved for each cross section or lateral position is also given in tables 9 and 10. A comparison of the values presented in tables 9 and 10 shows that there is essentially no variation in the mean values of c?16 and d50 with either location in the channel or with time. There appears to be slightly more variation in the rf84 and cr values, mostly with respect to longitudinal location for the May 8 values. This variation could be the result of pumice which was found in several of the samples. Pumice has a very low specific gravity and will bias slightly distributions obtained by sieve analysis. The mean values of the parameters in tables 9 and 10, however, are consistently larger than those for the 0.1-m-long (4-in-long) core samples given in table 8. There are several possible explanations for this difference. The 0.1-m-long (4-in-long) core samples were analyzed by the visual accumulation-tube procedure, whereas the 0.6-m-long (2-ft-long) core samples were sieved. Figure 7 of the U.S. Inter-Agency Committee on Water Resources (1957b) shows the difference between fall diameter and sieve diameter for naturally worn quartz particles of various shape factors. However, for particles to have a sieve diameter of 0.30 mm and a fall diameter of 0.24 mm, Table 10. — Mean parameters of the size distributions of bed material from lateral positions as determined by sieve analysis; 0.6-m core samples, June 6 and July 14 Lateral position Number of segments (n) ,dl6 (mm) d50 (mm) d84 (mm) cr Left quarter .. 145 June 6, 1967 0.202 0.300 0.490 1.56 Centerline .. .. 145 .211 .306 .469 1.54 Right quarter . 145 .202 .292 .468 1.52 Composite .. .. 435 .204 .299 .470 1.52 Left quarter . . 75 July 14,1967 .201 .300 .478 1.54 Centerline . ... 75 .211 .302 .461 1.48 Right quarter . 75 .199 .300 .478 1.55 Composite . . . . 225 .202 .300 .472 1.53 the shape factor would have to be less than 0.3; for the particles involved, such a shape factor would seem to be too small to be reasonable. Estimates of shape factors for the 0.125- to 0.177-mm, 0.177- to 0.250-mm, and 0.250- to 0.350-mm sieve classes of bed material from the geometric mean sieve size, the fall diameters presented in table 2, and figure 7 of the U.S. Inter-Agency Committee on Water Resources (1957b) are 0.8, 0.6, and 0.65, respectively. Probably a better explanation for the difference is that the 0.6-m-long (2-ft-long) cores contained particles from throughout the dune whereas, the 0.1-m-long (4-in.-long) cores consisted primarily of the finer topset material on the crests and backs of the dunes. The bed of the canal consisted of dunes that moved slowly downstream. The dimensionless Chezy discharge coefficient, C/Jg, where g is the acceleration of gravity and C is the Chezy coefficient, ranged from 1012 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO Run T able 11. — Hydraulic and bed-form data, June 1966 I From Nordin (1971 )| Date Mean depth im) Mean velocity tm/s) Water- surface slope (m/m) Water temperature (°C) Dune length Length Cy . v , v a netite moved in the upper part of the zone of movement for the dime-bed form. The Student’s t test (Bennett and Franklin, 1954) was used to determine whether significant differences existed among mean values of the depth of mixing for May 8, June 6, and July 14. Mean values for groups of sieve classes were compared in pairs and results are presented in table 40. The May 8—June 6 and May 8-July 14 comparisons, with the exception of the May 8-July 14 comparison for the lead tracer particles, show that the mean depths of mixing for June 6 and July 14 were significantly larger than for May 8 at the 0.005 level of significance. The June 6 -July 14 comparisons show that the mean depths for these dates were not significantly different at the 0.05 level of significance. Thus, the depth of mixing increased significantly between May 8 and June 6CORE SAMPLES 35 Table 40. — Results of the Student’s t test of the mean values of the depth of mixing for the fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Dates, minerals, and sieve classes8 Confidence limits for difference of means*3 Significance level May 8, Q1-Q5: June 6, Q1-Q5 . . . .0.26 ±0.086 0.005 May 8, Q1-Q5: July 14, Q1-Q5 . . .0.26 ±0.10 .005 June 6, Ql — Q5: July 14, Q1-Q5 . . 0.0083±0.066 .05 May 8, G1-G5: June 6, G1-G5 ... 0.23 ±0.10 .005 May 8, G1-G5: July 14, G1-G5 . . 0.26 ±0.12 .005 June 6, Gl—G5: July 14, Gl— G5 .. 0.030 ±0.078 .05 May 8, M1-M5: June 6, M1-M4 . . 0.21 ±0.11 .005 May 8, Ml—M5: July 14, Ml—M3 . 0.22 ±0.15 .005 June6, Ml—M4: July 14, Ml—M3 0.010 ±0.090 .05 May 8, LI—L4: June 6, LI—L4 .... 0.20 ±0.21 .005 May 8, L1-L4: July 14, L1-L4 . . . 0.11 ±0.17 .05 June 6, L1-L4: July 14, L1-L4 .. . 0.089 ±0.21 .05 May 8, Ql— Q5: June 6, Q1-Q5, G1-G5: G1-G2, M1-M5: M1-M2 .. 0.30 ±0.063 .005 June 6, Q1-Q5: July 14, Q1-Q5, G1-G2: G1-G3, M1-M2: Ml-M2 0.0051 ±0.048 .05 a Q, G, M, and L denote quartz, garnet, monazite, and lead, respectively; 1, 2, 3, 4, 5, and 6 denote the 0.125- to 0.177-mm, 0.177- to 0.250-mm, 0.250- to 0.350-mm, 0.350- to 0.500-mm, 0.500- to 0.707-mm, and 0.707- to 1.00-mm sieve classes, respectively. ^ Mean values are not significantly different at the indicated significance level if the confidence limits for the difference of the means include zero (Bennett and Franklin, 1954). but did not change significantly between June 6 and July 14. The second method for determining the depth of mixing, dm, is based on an analysis of a sonic depth-sounding record of the longitudinal profile of the bed surface. The record is divided into sections beginning at the upstream end. The first section lv extends from the first dune trough on the record to the next trough downstream that is at a lower bed elevation. The second section extends from this trough downstream to the next trough that is at a lower bed elevation and so on until the complete record is sectioned. The mean bed-surface elevation is determined for the total record, and the difference between the mean bed-surface elevation and the trough elevation gives the depth of mixing for that section. The mean value of the depth of mixing for the reach is the weighted average of the values for the sections given by where L is the length of the reach, li is the length of the ith section, dm is the depth of mixing for the ith section, and ns is the number of sections. The number of bed profiles obtained during the fluorescent tracer study was limited and the quality was poor because of equipment difficulties. Only one record obtained on May 7 along the centerline between the injection point and cross section 360 was analyzed, and a depth of mixing of 0.31 m (1.02 ft) was obtained. The third method for determining the depth of mixing, dm, is based on the conservation of the tracer particles. The equation is (23) where W is the weight of tracer particles injected, B is the channel width, X is the fraction of the volume of the bed not occupied by the bed-material particles, ys is the specific weight of the bed material, and J0 Cdx is the area under the longitudinal distribution curve for the tracer particles. This method differs for fluorescent and radioactive tracers. For radioactive tracers the integral term can be determined from in situ activity measurements with radiation detectors. For fluorescent tracers core samples are necessary to determine the integral term; hence, the mean value of the depth of mixing in effect must be known to permit determination of the concentrations. The centroid depth, which is the distance from the mean bed-surface elevation to the centroid of the vertical distribution of the tracer material, was also determined from the vertical distributions of the tracers. The mean value of the centroid depth was calculated from d = 1_ N D + % dknk k=\ (24) where dk is the distance from the top of the core to the center of the £th segment, that is 0.06, 0.18, 0.30, 0.42, and 0.54 m (0.2, 0.6,1.0,1.4, and 1.8 ft), respectively, for the five 0.12-m-thick (0.4-ft-thick) segments, and nk is the number of fluorescent particles in the £th segment. The mean value of the centroid depth was calculated from equation 24 for each sieve class of each of the four minerals for May 8, June 6, and July 14. The basic raw data were used in these calculations — that is, small numbers of fluorescent particles in the bottom segments were not neglected as was done in the determination of the depth of mixing. This procedure was possible because the centroid depth is weighted with respect to the number of particles and hence small numbers do not contribute substantially to the summation in equation 24; on the other hand, the depth of mixing is based strictly on the presence or absence of particles in a specific segment of the sample. Mean values of the centroid depth, d, for the various sieve classes and specific gravities of tracers and the36 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO three sample dates are presented in table 41. To determine if significant differences existed among these mean values, the technique of analysis of variance, as was used for the depth of mixing, was used. Results of the analysis of variance of the centroid depths are presented in table 42. Few differences were found among the centroid depths for each of the three sampling dates. For May 8 it was necessary to exclude the depths for the 0.500-to 0.707-mm sieve class of garnet tracer and the 0.250- to 0.350-mm and 0.350- to 0.500-mm sieve classes of lead tracers to have no significant differences at the 0.05 level. For June 6 it was necessary to exclude the depths for the 0.125- to 0.177-mm sieve class of quartz tracer and the 0.250- to 0.350-mm and 0.350- to 0.500-mm sieve classes of lead tracers to have no significant differences at the 0.05 level. For July 14 it was necessary to exclude only the depth for the 0.125- to 0.177-mm sieve class of quartz tracer to have no significant differences at the 0.05 level. Table 41. — Mean values of the centroid depths of the vertical distributions of fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Mineral Sieve class d (m) (mm) May 8 June 6 July 14 Quartz 0.125-0.177 0.18 0.27 0.31 0.177-0.250 .16 .24 .29 0.250-0.350 .16 .21 .25 0.350-0.500 .17 .23 .24 0.500-0.707 .18 .22 .23 0.707-1.00 .19 .23 .25 Garnet 0.125-0.177 .19 .24 .27 0.177-0.250 .18 .22 .26 0.250-0.350 .19 .22 .26 0.350-0.500 .20 .25 .26 0.500-0.707 .23 .25 .29 Monazite 0.125-0.177 .16 .20 .27 0.177-0.250 .16 .20 .26 0.250-0.350 .18 .23 .25 0.350-0.500 .18 .23 .26 0.500-0.707 .21 .24 .26 Lead 0.125-0.177 .18 .26 .25 0.177-0.250 .18 .25 .28 0.250-0.350 .25 .29 .28 0.350-0.500 .26 .27 .27 The Student’s t test (Bennett and Franklin, 1954) was used to determine if significant differences existed among mean values of the centroid depth for May 8, June 6, and July 14. Mean values for groups of sieve classes were compared in pairs and results are presented in table 43. All the comparisons for the three sampling dates, with the exception of three comparisons for the lead tracers, showed significant differences at the 0.005 level. Thus, the results of the statistical tests show that the centroid depths for the quartz, garnet, and monazite tracers increased with time throughout the study. This is in contrast with the depths of mixing which increased between May 8 and June 6 but did not Table 42. — Results of the analysis of variance of the centroid depths of the vertical distributions of fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Date Minerals and Variance ratio, sieve classes F included0 F b vj, v2, ot May 8 .Q1-Q6 0.90 F5,508,0.05= 2'21 G1-G5 1.64 F4,407,0.05 = 237 M1-M5 1.93 F4,389,0.05 = 2-37 L1-L4 5.67 F3,190,0.005 =4.28 LI, L2 . . .04 Fl,111,0.05= 3'27 L3, L4 . . .09 Fl,79,0.05 = 3 96 Q1-Q6, G1-G5, M1-M5, LI, L2 1.67 F17,1410,0.025 = F77 Q1-Q6, G1-G4, M1-M5, LI, L2 1.07 F16,1357,0.05= l-64 June 6 .Q1-Q6 3.13 F5,415,0.05= 2-2l Q2-Q6 . . .91 F4,329,0.05 = 237 G1-G5 1.23 F4,354,0.05 = 237 M1-M5 1.78 F4,342,0.05 = 2-37 L1-L4 .. .35 F3,87,0.05= 2'71 Q2-Q6, G1-G5, Ml -M5, L1-L4 1.65 F18,1112,0.025= 1-60 Q2-Q6, G1-G5, M1-M5, LI, L2 1.29 F16,1075,0.05= l-64 July 14 Q1-Q6 2.81 F5,247,0.01= 392 Q2-Q6 1.38 F4,205,0.05= 237 G1-G5 . . .43 F4,202,0.05= 2 37 M1-M5 .. .20 F4,197,0.05= 237 L1-L4 .. .13 F3,60,0.05= 2 76 Q2-Q6, G1-G5, Ml -M5, L1-L4 .. .55 F18,662,0.05= l-69 a Q, G, M. and L denote quartz, garnet, monazite. and lead, respectively; 1. 2, 3. 4. 5, and 6 denote the 0.125- to 0.177-mm, 0.177- to 0.250-mm, 0.250- to 0.350-mm, 0.350- to 0.500-mm. 0.500- to 0.707-mm. and 0.707- to 1.00-mm sieve classes, respectively. ^Values for F from table IV of Bennett and Franklin 11954 >. v . v ,a change significantly between June 6 and July 14. These observations suggest that the lower limit of particle movement essentially had been established prior to June 6, whereas the centroids of the vertical distributions of the tracer particles apparently were still moving downward at the end of the study on July 14. Another explanation for the increase of the centroid depths between June 6 and July 14 is possible, however. The water discharge, and consequently the sediment transport rate, decreased considerably toward the end of the study. Thus, lower regions of the bed that previously had been involved in the transport process may no longer have been in motion. Because tracers were deposited in these regions of the bed by the higher flow and transport conditions, a layer of bed material con-CORE SAMPLES 37 Table 43. — Results of the Student’s t test of the mean values of the centroid depths of the vertical distributions of fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Dates, minerals, and sieve classes41 Confidence limits for difference of means0 Significance level May 8, Q1--Q6: June 6, Q2—Q6 0.17±0.067 0.005 May 8, Q1--Q6: July 14, Q2-Q6 0.26±0.081 .005 June 6, Q2--Q6: July 14, Q2-Q6 0.085±0.091 .005 May 8, G1-G5: June 6, G1--G5 0.12±0.076 .005 May 8, G1 -G5: July 14, G1-G5 0.24 ±0.090 .005 June 6, G1-G5: July 14, G1-G5 0.11+0.090 .005 May 8, M1-M5: June 6, M1--M5 0.14±0.073 .005 May 8, Ml—M5: July 14, Ml—M5 0.28±0.088 .005 June 6, Ml —M5: July 14, Ml—M5 0.13±0.089 .005 May 8, LI - L2; June 6, LI - L4 0.28±0.17 .005 May 8, L1-L2: July 14, L1-L4 0.28±0.21 .005 June 6, L1-L4: July 14, L1-L4 0.01 ±0.17 .05 May 8, L3--L4: June 6, L1-L4 0.04 ±0.14 .05 May 8, L3-L4: July 14, LI-L4 0.04 ±0.16 .05 May 8, Ql—Q6: June 6, Q2—Q6, G1-G4: G1-G5, M1-M5: M1--M5, L1-L2: L1-L2 0.16±0.04 .005 May 8, Ql -Q6: July 14, Q2-Q6, G1-G4: ‘ G1-G5, M1-M5: M1-M5, L1-L2: L1-L4 0.27+0.05 .005 June 6, Q2--Q6: July 14, Q2-Q6, G1--G5: G1-G5. M1--M5, Ml —M5: L1-L2; L1-L4 0.11+0.05 .005 a Q, G. M. and L denote quartz, garnet, monazite. and lead, respectively; 1. 2. 3, 4, 5. and B denote the 0.125- to 0.177-mm. 0.177- to 0.250-mm, 0.250- to 0.350-mm, 0.350- to 0.500-mm, 0.500- to 0.707-mm. and 0.707- to 1.00-mm sieve classes, respectively. ^ Mean values are not significantly different at the indicated significance level if the confidence limits for the difference of the means include zero (Bennett and Franklin, 1954) taining tracer particles, but which was no longer in motion, may have been formed. Under these circumstances, the depth of mixing as indicated by the core samples would not change, as was found (table 38); but the centroid depth would appear to increase because the tracer particles still in motion moved on down the channel and out of the study reach, whereas the layer of particles not in motion remained, lowering the centroid of the vertical distribution. This effect should be greatest for the small particles because they move the fastest (figure 44). A consideration of the increase of the d values between June 6 and July 14 shows that, with the exception of the lead particles, the increases in general are largest for the small particles, thus supporting the hypothesis. The ratio, R, of the centroid depth to the depth of mixing was used as an index of the vertical concentration gradient by Hubbell and Sayre (1965). For a concentration distribution that is uniform, R = 0.5; for a concentration distribution that has the larger part of the tracers in the upper part of the sample, R <0.5; and for a concentration distribution that has the larger part of the tracers in the bottom part of the sample, R> 0.5. Courtois and Sauzay (1966) computed values of the Table 44. — Values of the ratio R of the centroid depth to the depth of mixing for the vertical distributions of fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Mineral Sieve class R (mm) May 8 June 6 July 14 Quartz . . . . . 0.125-0.177 0.35 0.47 0.53 0.177-0.250 .31 .41 .50 0.250-0.350 .33 .36 .43 0.350-0.500 .35 .40 .41 0.500-0.707 .37 .40 .42 0.707-1.00 .44 .53 .58 Garnet . . . . . 0.125-0.177 .44 .44 .49 0.177-0.250 .39 .40 .47 0.250-0.350 .41 .42 .47 0.350-0.500 .41 .48 .53 0.500-0.707 .50 .54 .56 Monazite . . . 0.125-0.177 .35 .36 .49 0.177-0.250 .35 .38 .47 0.250-0.350 .39 .47 .51 0.350-0.500 .42 .47 .53 0.500-0.707 .46 .52 .60 Lead . 0.125-0.177 .49 .60 .68 0.177-0.250 .53 .58 .61 0.250-0.350 .58 .63 .65 0.350-0.500 .65 .59 .68 ratio for five types of distributions and found that R was between 0.33 and 0.50 for all five. These five included a uniform distribution, a linear distribution decreasing from the maximum concentration at the surface of the bed, two types of parabolic distributions with the maximum concentration at the surface of the bed, and a parabolic distribution with the maximum concentration at some fraction of the depth of mixing below the bed surface. They suggested, therefore, that the depth of mixing could be replaced in the sediment transport equation by 2.5 d, with a maximum error on the order of 25 percent. This procedure thereby circumvents the previously discussed difficulties that are involved in estimating the true lower limit of the zone of movement for determining the depth of mixing. Values of the ratio for the various sieve classes of tracer and the three sample dates are presented in table 44. The ratio values in general increased with time, probably because the centroid depths increased with time. The quartz, garnet, and monazite ratios for May 8 ranged from 0.31 to 0.50, indicating that most of the distributions were weighted toward the bed surface. The ratios for June 6 and July 14 were larger and, in several instances, exceeded the 0.50 of a uniform distribution. Most of the ratios larger than 0.50 were for the lead particles and the largest of the quartz, garnet, and monazite particles. Thus, the vertical distributions for these particles tended to be weighted toward the bottom of the zone of movement— that is, concentrated in the lower regions, rather than uniformly distributed over the depth of movement. The ratio values for July 14 for the 0.125- to 0.177-mm and 0.177- to 0.250-mm sieve classes of quartz tracer were larger than might be expected in comparison with the values38 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO for the other sieve classes. However, these values were probably large because the d values for these sieve classes were larger than for the other sieve classes, for reasons discussed previously. CORE-SAMPLE CONCENTRATIONS In previous sections, the vertical distributions of the tracers within the different core samples were discussed. From this point on, the variations with lateral and longitudinal positions and with time of the concentrations of the fluorescent tracers as determined from the core samples will be discussed. Before computing concentrations of fluorescent tracers from the results of the analysis of the core samples, it is necessary first to consider the problem of how to treat the bottom limit of the cores. Because of the irregular nature of the dune bed and the constant 0.6-m (2-ft) length of the sampler (fig. 5), samples to varying depths below the mean bed elevation were obtained. The depth of flow at the sampling points for the 115 core samples collected on May 8 ranged from 0.6 to 1.1 m (2.0 to 3.6 ft) and averaged 0.79 m (2.6 ft). On June 6 the depth of flow at the sampling points for the 87 samples ranged from 0.52 to 1.0 m (1.7 to 3.3 ft) and averaged 0.73 m (2.4 ft). On July 14 the depth for the 45 samples collected ranged from 0.34 to 0.82 m (1.1 to 2.7 ft) and averaged 0.52 m (1.7 ft). The variation of the bed-surface elevation with longitudinal position may also be seen in the vertical concentration distributions presented on plate 1. In theory, sampling depth should be just to the lower limit of the zone of particle movement. This lower limit, however, undoubtedly varies with time and position in the reach. However, in the calculations in this report, the lower limit of the zone of particle movement was assumed on the average to be a plane parallel to the water surface. This assumption should be valid so long as equilibrium flow conditions (as defined, for example, by Simons and Richardson, 1966, p. J3) exist in the channel. To correct the core samples to the lower limit of the zone of movement, the following procedure was used. First, the depth of mixing was determined, as described previously, for each sieve class of each of the four minerals for each core sample analyzed. Second, the mean value of the depth of mixing was determined from the values obtained in step 1 for each of the three sampling dates (May 8, June 6, July 14). Third, the concentration distribution for each core sample was extrapolated or truncated to a distance below the mean bed-surface elevation equal to the mean value of the depth of mixing for the specific sieve class of mineral under consideration. The extrapolations and truncations were to the nearest 0.03 m (0.1 ft). Estimates of the weight of material in the truncated or extrapolated parts were on a proportional basis — that is, if a segment were truncated at the mid-point, the retained part was assumed to contain half the material; if the segment were extrapolated half the thickness of a segment, the extrapolated part was assumed to contain half as much material as was contained in the segment just above the extrapolated portion. Estimates of the number of tracer particles were made on the basis of the particle distributions in all the segments of the core sample above the truncated or extrapolated part. The concentration of fluorescent tracers in each core sample was calculated from K—O 2 k=\ k=5 2 k=i nk wk (25) where ra^is the number of fluorescent particles of a particular sieve class for a particular mineral in the k th segment of the core sample, wAis the weight in grams of all material in that sieve class in the k th segment, and Np is the particles per gram factor for that sieve class and mineral (table 5). The summation in equation 25 was generally not over precisely five segments because of the extrapolation or truncation of the concentration distributions of the core samples. LATERAL DISTRIBUTIONS Lateral distributions of the fluorescent tracers were determined from the May 8 core samples for cross sections 7.5,15, 30, 60, 90,180, 270, 360, 450, and 540. The distributions for the 0.125- to 0.177-mm sieve classes of quartz, garnet, and monazite are presented in figures 20, 21, and 22, respectively, as examples. The concentrations plotted in the figures are relative concentrations—that is, each concentration has been divided by the area under the lateral distribution curve. The defining equation is CM) = -,Cte)------ (20 /„ c<21 * Several trends are evident in figures 20—22. First, the distributions for the cross sections near the injection point were approximately Gaussian, whereas the distributions for the cross sections farther downstream tended to be more rounded and also more irregular. The increased irregularity was probably because the distributions at the downstream cross sections were not fully developed, whereas the increased roundness was a result of the increased lateral spread or dispersion of the particles with distance downstream. Second, the peak relative concentration decreased with distanceRELATIVE FLUORESCENT TRACER CONCENTRATION CORE SAMPLES 39 Figure 20. — Variation with distance downstream in the lateral distributions of relative concentrations of the 0.125-to 0.177-mm sieve class of quartz tracers as defined by core samples collected on May 8. downstream because of the increased lateral spread. Third, the lateral distributions for the cross sections be- tween the source and cross section 90 tended to have mean lateral positions 0.9 to 1.5 m (3 to 5 ft) to the leftRELATIVE FLUORESCENT TRACER CONCENTRATION 40 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO LEFT BANK, IN METERS Figure 21. — Variation with distance downstream in the lateral distributions of relative concentrations of the 0.125- to 0.177-mm sieve class of garnet tracers as defined by core samples collected on May 8.RELATIVE FLUORESCENT TRACER CONCENTRATION CORE SAMPLES 41 DISTANCE FROM LEFT BANK, IN METERS Figure 22. — Variation with distance downstream in the lateral distributions of relative concentrations of the 0.125- to 0.177-mm sieve class of monazite tracers as defined by core samples collected on May 8.TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO 42 of centerline, whereas the distributions for the cross sections farther downstream tended to have mean lateral positions to the right of centerline. MEAN LATERAL POSITIONS AND VARIANCES The mean lateral positions, z, of the lateral distributions were calculated from equation 12. The z values tended to shift toward the left bank for cross sections between the injection point and cross section 90. This was most evident for the large sizes of tracer particles. For cross sections farther downstream, the z values for the small particles were generally to the right of centerline, and the z values for the large particles varied considerably, presumably because these lateral distributions were not completely developed. Comparison of the z values for the “dustpan” samples with the 2 values for the core samples showed that differences were small; when large differences in the z values did occur, they were almost always for the large particles and (or) downstream cross sections where the distributions probably were not completely developed at the time of sampling. The variances of the lateral distributions were calculated from equation 11 and are plotted as a function of distance downstream in figures 23 and 24 for the quartz and garnet, and monazite and lead tracers, respectively. The variation of the variances was essentially the same as that shown in figure 19 for the variances determined from the “dustpan” samples, that is, the variances increased approximately linearly with distance for several points, then tended to fall away from the linear relation. Similarly, the variances at a particular cross section tended to increase with the size of the tracer particles as for the “dustpan” samples. Several variances exceeded the theoretical maximum for a uniform distribution of about 23 m2 (250 ft2); but, as for the “dustpan” samples, these were for large particles at downstream cross sections where the lateral distributions were not fully developed but were bimodal or multimodal. LATERAL DISPERSION COEFFICIENTS Least-squares lines were calculated for the approximately linear range of the variance versus distance data, and these lines are shown in figures 23 and 24. The slopes of these lines, dcrydx, and the number of cross sections considered in each of the least-squares calculations are presented in table 45. The dcr\ldx values increased with the size of the tracer particles for each mineral, suggesting that the rate of lateral dispersion, as indicated by the rate of change of the variance with distance, increased with particle size. Similar results were obtained for the “dustpan” samples (table 37). Table 45. — Rate of change of the variance with distance downstream, dcrVdx, and lateral dispersion coefficient for fluorescent tracer particles as determined from concentration distributions defined by core samples collected on May 8 Mineral Sieve class (mm) Number of cross sections da-ydx imVm) ^Z (m2/h) Quartz .. . 0.125-0.177 6 0.0573 0.0460 0.177-0.250 6 .0664 .0382 0.250-0.350 6 .0646 .0157 0.350-0.500 5 .109 .0151 0.500-0.707 5 .123 .0140 0.707-1.00 3 .174 .0170 Garnet .. 0.125-0.177 6 .0728 .0292 0.177-0.250 6 .0710 .0142 0.250-0.350 5 .0823 .0124 0.350-0.500 5 .141 .0174 0.500-0.707 4 .183 .0240 Monazite 0.125-0.177 6 .0716 .0218 0.177-0.250 5 .0978 .0174 0.250-0.500 4 .142 .0245 0.350-0.500 4 .193 .0248 0.500-0.707 4 .234 .0307 Lead .... 0.125-0.177 5 .150 .0082 0.177-0.250 4 .145 .0092 0.250-0.350 4 .106 .0097 0.350-0.500 4 .436 .0223 Comparison of the dcrydx values for the core and “dustpan” samples showed that 11 of the core sample values were larger than the corresponding “dustpan” sample results and that 5 were smaller. The largest difference was for the 0.707- to 1.00-mm sieve class of quartz; however, only three cross sections were used in determining dcr\ldx for this size, and the small number of points undoubtedly contributed to the large difference observed. The mean dcrydx value of the 16 sieve classes (excluding the lead) was 0.118 m2/m (0.386 ft2/ft) compared with a mean of 0.110 m2/m (0.361 ft2/ft) for the “dustpan” samples. The d(r\ldx values are plotted in figure 25 as a function of median fall diameter of the sieve class. The trend line through the data shows the increase of dcrydx with increase in the median fall diameter of the tracer particles. There did not appear to be any consistent differences between the dcr ydx values determined from the core samples and those determined from the “dustpan” samples. The scatter of the data precluded any definite conclusions regarding hydraulic equivalence of the particles. It appeared, however, that the scatter increased with diameter, suggesting that the smaller particles of differing specific gravity, which move predominantly in suspension, were more nearly equivalent than the larger particles. The velocities of the centroids of the tracer masses, to be discussed in the next section (table 49), were used to convert the dcrydx values to lateral dispersion coefficients with equation 16. The resulting kz values areVARIANCE, IN SQUARE METERS CORE SAMPLES 43 Figure 23. — Variation with distance downstream in the variances of the lateral distributions of the quartz and garnet tracers as defined by core samples collected on May 8. presented in table 45. Because the d (33) ys(l-k)B dm.pi where W( is the weight of tracer of sieve class i injected for the experiment. The Wi values are given in table 4 and the dm values in table 38. A value of 1600 kg/m3 (100 lb/ft3) was used for ys (1-X), and a value of 17 m (55 ft) was used for the mean channel width, B. Means of the pi values determined from the analyses of the many core samples were used. The Am. values defined by equation 32 are the denominators of equation 4, and the procedure used in the determination of these integrals was described pre-CORE SAMPLES 67 Table 52. — Areas under the measured longitudinal distribution curves (Am) and recovery ratios ('RRj for the fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Sieve class Am, in gram-meter/gram RR Mineral (mm) May 8 June 6 July 14 May 8 June 6 July 14 Quartz ... 0.125-0.177 0.101 0.019 3 0.006 64 0.422 0.0846 0.0304 0.177-0.250 .089 3 .035 4 .016 2 .691 .288 .132 0.250-0.350 .079 6 .072 8 .065 5 .845 .972 .853 0.350-0.500 .231 .251 .283 .785 1.14 1.32 0.500-0.707 .500 .533 .637 1.12 1.37 1.67 0.707-1.00 .115 .112 .142 1.90 1.82 2.27 Garnet .. 0.125-0.177 .019 6 .016 7 .012 5 1.17 1.13 .888 0.177-0.250 .029 0 .029 2 .026 3 1.00 1.09 1.02 0.250-0.350 .043 9 .044 8 .045 1 .800 .954 1.00 0.350-0.500 .051 5 .065 2 .076 8 .547 .833 .966 0.500-0.707 .012 2 .019 0 .022 3 .427 .722 .956 Monazite 0.125-0.177 .091 7 .101 .082 3 .738 .900 .754 0.177-0.250 .035 1 .043 9 .033 2 .858 1.16 .901 0.250-0.350 .024 1 .027 8 .025 0 1.05 1.37 1.14 0.350-0.500 .044 5 .061 3 .057 6 .716 1.26 1.16 0.500-0.707 .046 6 .068 3 .065 2 .754 1.13 1.04 Lead .... 0.125-0.177 .036 9 .065 2 .022 9 .469 .926 .287 0.177-0.250 .010 2 .107 .008 81 .292 3.41 .297 0.250-0.350 .007 56 .035 7 .003 29 .313 1.62 .138 0.350-0.500 .007 01 .025 6 .002 04 .213 .978 .0715 viously in the section on the centroid velocities of the tracers. The Am values for May 8, June 6, and July 14 are presented in table 52. The areas under the measured longitudinal distribution curves for each sieve class of tracer should be identical for the three sample times if the sample collection and analysis procedures are perfect. Consideration of the A„ values in table 52 shows definite trends with both time and particle size. The areas for the 0.125- to 0.177-mm sieve classes of quartz and garnet and the 0.177- to 0.250-mm sieve class of quartz decreased with time. This decrease with time was probably the result of failure to extrapolate the leading edges of the concentration distributions far enough downstream. If the May 8 area is used as a basis, then only about 19 percent and 6.6 percent of the 0.125- to 0.177-mm quartz tracer remained in the study reach on June 6 and July 14, respectively; similarly for the 0.177- to 0.250-mm quartz, about 40 percent and 18 percent remained on June 6 and July 14, respectively. Stated another way, if the mean longitudinal concentrations observed in the study reach on June 6 and July 14 for the 0.125- to 0.177-mm sieve class of quartz were continued downstream, then the lengths of reach required to give an area equal to the area under the longitudinal distribution curve observed on May 8 would be 5420 and 15 200 m (17 800 and 49 800 ft), respectively; similarly, for the 0.177- to 0.250-mm sieve class of quartz, the reach lengths would be 2450 and 5890 m (8040 and 19 320 ft), respectively. These large reach lengths indicate that the centroid velocities for these sieve classes may have been considerably underestimated using the limited extrapolation procedures described previously. The change in area with time for the 0.125- to 0.177- mm sieve class of garnet was not so large as for the quartz tracers. Using the May 8 area as a basis, the percentages remaining on June 6 and July 14 were about 85 and 64, respectively. The area values for the four sieve classes of lead tracers varied considerably with time and showed no consistent trends except that the areas were largest for all four sieve classes on June 6. The large variations in the areas for the lead tracers were probably the result of a problem associated with counting the lead particles. A blue dye was used on the lead and the color of this dye’s fluorescence resembled that of lint and fiber under the ultraviolet light. Hence, it was difficult to distinguish between actual lead tracer particles and extraneous materials. The areas for some of the large particles increased with time. This is believed to be a result of the fact that on May 8 these tracers had, for the most part, moved only a short distance downstream. Hence, sample coverage was poor, and linear interpolation between the limited number of sample points probably resulted in large errors in the estimation of the areas. On the later dates, these particles had moved farther downstream, and sample coverage was improved. The recovery ratios defined by equation 30 are presented in table 52 also. The ratios, in comparison with the theoretical value of 1.00, were satisfactory, with the exception of the June 6 and July 14 values for the 0.125- to 0.177-mm and 0.177- to 0.250-mm sieve classes of quartz tracer, the three values for the 0.707-to 1.00-mm sieve class of quartz tracer, and the values for the four sieve classes of lead tracer. The lead tracers have both the largest and the smallest of all the ratio values and by far the largest variability with time. This variability was not unexpected, in view of the large68 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO variability of the Am values presented in table 52 and the previously discussed problem in counting the lead particles. The decrease with time of the ratios for the two sieve classes of quartz tracer also was expected because of the aforementioned movement of these sizes of tracer out of the study reach and the associated decrease of the Amvalues. The May 8 ratio suggests that only slightly more than 40 percent of the 0.125- to 0.177-mm sieve class of quartz tracer was recovered on this sample date; this means that the estimated extrapolations based on the May 8 areas, discussed in previous paragraphs, would be even longer than those presented. The explanation for the large values for the 0.707- to 1.00-mm sieve class of quartz tracers is unknown. Ratios larger than 1.00 indicate that more tracer material was recovered than was injected. A possible explanation is that the Am values were too large as a result of the poor sample coverage discussed previously. To analyze the calculation of the recovery ratios, consider the various factors contained in equations 32 and 33. The -ys(l —A) and B factors were discussed in the section on the spatial-integration procedure and were considered to be known accurately. The pi factors were determined from the analysis of a large member of 0.6-m (2-ft) core samples and thus were accurately known. The W- factors were determined from the size distributions of the pure tracer materials (table 1) and the total amounts of tracer injected (table 3). The total amount injected was determined for each tracer from the injection rate and the length of the injection period. This amount was checked against the amount determined from the weights of tracers in each bag at the beginning of the experiment and the empty bags at the conclusion of the experiment; the differences for the various tracers were all less than 5 percent. The mean value of the depth of mixing appears both in the denominator of equation 33 and inherently in equation 32 as the upper limit of integration over the vertical direction. To estimate the effect that an incorrect dm value would have on the recovery ratio, recovery ratios were computed using a value of 0.6 m (2.0 ft). It was found that a relatively large error in the depth of mixing had only a very small effect on the calculated values of the recovery ratios. The explanation for this may be seen at least qualitatively by considering again equations 30, 32, and 33. If dm. is overestimated, then the effect in equation 33 is to give an At. value that is too small; the effect in equation 32 will be to include additional bed material containing zero or very small concentrations of tracer at the bottom of the core, thus producing concentrations that are too small and areas under the distribution curves that are too small; the net effect on the recovery ratio in equation 30 tends to cancel. It was concluded, therefore, that the computation of the recovery ratio is relatively insensitive to the value of the depth of mixing, at least for the conditions of the present study. The final factor entering into the calculation of the recovery ratios is the integral term defined by equation 32. Numerous considerations enter into the accurate determination of these integrals, among which are obtaining sufficient samples across the channel and sampling to a sufficient depth to define accurately the cross sectional mean concentration; sieve analysis of the samples and counting of the number of fluorescent particles of each color in each sieve class; determining the particles per gram factors (table 5) for conversion of the number of particles to concentrations; determining the depth of mixing from the vertical distributions of the tracers; conversion of the cross sectional mean concentrations into longitudinal distributions; and extrapolating or truncating the leading edges of the longitudinal distributions to give a finite upper limit of integration for equation 32. The same procedures were used for all 20 sieve classes of tracer materials; thus, if there had been some consistent error in these procedures, one might expect that the recovery ratios would have all been similarly affected. However, the various sieve classes of tracers all moved with different velocities and dispersion rates, and, hence, the same procedure could result in different errors for the different sieve classes and specific gravities of tracers. Mean values of the recovery ratios were calculated, neglecting the 0.125- to 0.177-mm and 0.177- to 0.250-mm sieve classes of quartz tracer for the June 6 and July 14 sample dates and the 0.707- to 1.00-mm sieve class for all three dates. Results are presented in table 53. The mean ratios increased from May 8 to June 6 but changed little between June 6 and July 14. In view of the large volume of bed material in which the tracers were mixed, the recovery ratios are considered satisfactory. Table 53. — Mean values of the recovery ratios for the fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Mean recovery ratio Mineral May 8 June 6 July 14 Mean Quartz ............ 0.772 1.16 1.28 1.02 Garnet............. .789 .945 .966 .900 Monazite...............824 1.16 .999 .996 Mean........ 0795 L08 L05 T77~ Overall mean . ... ... ... 0.966 LONGITUDINAL DISPERSION COEFFICIENTS The recommended procedure (Sayre and Chang, 1968) for calculating the longitudinal dispersion coefficient, Kx, of a dissolved dispersant is U:i do? K*= ~2~ik (34)CORE SAMPLES 69 where U is the mean flow velocity, cr^ is the variance of a concentration versus time curve, and x is the coordinate in the longitudinal direction. Fischer (1966) showed that an equivalent representation for Kx is K. 1 da2 2~ ~dt~’ (35) where a2 is the variance of a concentration versus distance curve, and t is time. Furthermore, Fischer (1966) showed that these relations are independent of the initial distribution of tracers, after some initial period during which the one-dimensional diffusion equation does not apply. In equations 34 and 35 the variances are determined by the method of moments. By analogy, then, the longitudinal dispersion coefficient for sediment tracer particles moving by a combination of transport along the bed surface, saltation, and suspension may be defined as 1 dcr2 2 dt~’ (36) Variances of the longitudinal distribution curves of the various sieve classes of tracers for May 8, June 6, and July 14 were calculated from °V2 (37) where x is the position of the centroid, the calculation of which was described in the section on the velocities of the centroids of the tracer masses. The upper limit of integration is the longitudinal position at which the concentration decreased to 1.0 percent of the maximum concentration. The denominator is the area under the longitudinal distribution curve; these areas have been presented previously in table 52. The integral in the numerator was evaluated in the same manner as the integrals were evaluated in determining the velocities, that is, by plotting x2C versus x and assuming that x2C varied linearly with * between sample points. The area was determined by summing the areas of the trapezoids thus formed. The variances of the longitudinal distribution curves are presented in table 54 and plotted as a function of the median fall diameter of the sieve class in figures 45, 46, and 47 for the quartz, garnet, and monazite tracers, respectively. The variances for the lead distributions were not plotted because they were considered to be less reliable than the quartz, garnet, and monazite values. Figures 45, 46, and 47 show similar trends —that is, a very large dependence of cr2 on fall diameter for the May 8 samples, less dependence on diameter for the June 6 samples, and least dependence on diameter for Figi kk 45. — Variation with fall diameter in the variances of the longitudinal distributions of the quartz tracers as defined by core samples collected on May 8, June 6, and July 14. the July 14 samples. The vertical dashed lines in these figures represent the size limits of the sieve classes used in the analysis of the samples and the numbers next to each point are the sieve class number given in table 54. With the exception of the 0.500- to 0.707-mm sieve class of garnet on May 8 and July 14, the variances in general decrease with increasing particle size. The variance, a2, represents the rate of spreading70 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO Figure 46. — Variation with fall diameter in the variances of the longitudinal distributions of the garnet tracers as defined by core samples collected on May 8, June 6, and July 14. downstream. These variances and the longitudinal distribution curves from which they were derived (figs. 26 40) clearly show that the small particles moved and spread far downstream, whereas the large particles tended to move only a short distance downstream from the source. The variance values presented in table 54 for the lead tracers in general show exactly the opposite effect — that is, 0-2 tended to increase with particle size. The reason for this is uncertain, but it probably reflects the generally less reliable nature of the lead results. Personnel counting fluorescent particles tended to ignore a Table 54. — Variances of the longitudinal distributions of fluorescent tracers as defined by core samples collected on May 8, June 6, and July 14 Mineral Sieve class (mm) 2 times 10 3 (m2 ) May 8 June 6 July 14 Quartz ... . in 0.125-0.177 29.4 80.7 73.5 (2) 0.177-0.250 22.0 64.5 76.0 (3) 0.250-0.350 4.04 33.8 60.8 (4) 0.350-0.500 1.11 16.2 46.4 (5) 0.500-0.707 .920 10.7 33.9 (6) 0.707-1.00 .845 8.45 34.8 Garnet . .. . (1) 0.125-0.177 18.0 55.2 82.5 (2) 0.177-0.250 4.25 36.1 62.8 (3) 0.250-0.350 1.31 24.9 53.2 (4) 0.350-0.500 1.15 13.7 42.4 (5) 0.500-0.707 2.56 10.2 55.3 Monazite . . (1) 0.125-0.177 12.0 43.0 67.6 (2) 0.177-0.250 2.40 28.3 61.3 (3) 0.250-0.350 1.17 19.4 49.2 (4) 0.350-0.500 1.11 11.2 33.4 (5) 0.500-0.707 1.02 9.66 28.9 Lead . (1) 0.125-0.177 .186 .604 .910 (2) 0.177-0.250 .316 .130 .948 (3) 0.250-0.350 .929 .929 3.87 (4) 0.350-0.500 .223 1.03 20.2 few small blue particles because small pieces of lint and fiber looked much the same under the ultraviolet light. To test the concept of hydraulic equivalence, the (m) fi ILO il i Left edge of water (ml Concentration times 10s, at indicated lateral position (in m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 95.1 l.i 2.30 2.77 2.16 3.58 8.44 9.85 6.32 1.56 1.43 .0692 0 108 i.i 9.45 8.49 2.82 5.17 12.9 36.9 1.44 5.10 .359 .133 0 118 i.i 9.50 4.74 3.49 5.36 123 34.7 11.3 5.69 0 2.05 .637 124 i.i 15.4 42.9 14.9 24.9 9.15 49.6 0 3.57 0 1.81 .458 132 i.i 47.1 15.5 6.83 3.31 30.0 10.8 2.92 4.72 1.95 (*) .177 138 i.i 40.5 13.1 4.00 37.4 43.5 12.6 18.8 4.83 1.19 1.68 .858 143 i.i (*) (*) 20.3 10.1 35.6 8.66 1.27 2.95 .467 .694 3.11 149 i.i 2.52 12.1 19.4 38.6 121 11.0 23.0 (*) 3.55 .880 2.85 162 i.i (*) 20.2 21.8 27.8 125 14.6 1.25 7.47 16.7 1.95 3.71 193 1.5 33.8 18.2 45.2 7.67 64.8 14.7 29.4 11.9 18.0 1.05 7.11 238 .5 1.31 4.77 19.4 15.1 57.5 14.4 30.8 9.09 6.03 2.13 285 .5 10.0 3.45 17.6 5.16 30.6 10.6 12.1 3.84 13.4 2.00 339 .5 3.40 27.5 1.70 11.1 20.1 8.80 50.0 14.5 19.5 .386 356 .3 6.35 4.97 7.41 34.1 14.5 31.7 1.40 2.02 3.40 0 429 .3 .152 1.34 .425 10.9 9.10 59.1 33.9 .989 2.99 1.28 525 .3 5.49 7.70 10.7 3.16 3.32 2.92 2.00 5.07 .255 .0801 597 .3 14.3 13.2 2.80 4.90 7.11 (*) 0 .170 0 .368 Table 26. — Concentration of monazite tracers in the 0.350- to 0.500-mm sieve class in “dustpan" samples collected at cross section 90 ILeaders (...) indicate no sample collected at that time and lateral position; (*) indicates sample was misplaced) Hours from injection (h) Left edge of water (m) Concentration times 10s, at indicated lateral position (in m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 0.067 i.i 0.0 (*) (*) 0.0 0.0 0.0 .270 i.i 0.0 0 0.0 0.0 0 0 0 .500 i.i o.o 0 0 0 0 0 0 0 o.o 0.726 1.13 i.i 0 0 0 0 0 0 0 .221 0 0 2.00 i.i 0 2.69 0 0 0 0 0 .325 0 (*) 3.09 i.i 0 0 0 0 (*) 0 0 1.73 0 0 4.05 i.i 0 0 0 0 0 0 0 .597 0 .440 5.06 i.i 0 0 0 0 0 .328 0 0 0 0 6.00 i.i 0 0 0 0 0 0 0 0 0 0 8.06 i.i 0 1.32 0 0 (*) 0 0 .535 0 0 10.0 i.i 0 0 0 0 0 0 0 0 0 0 12.0 i.i 0 0 0 0 0 0 0 0 0 0 22.0 i.i 0 0 0 0 .609 0 0 0 0 0 34.0 i.i 0 0 0 0 0 0 0 0 (*) 0 47.0 i.i o.o .509 0 0 .264 0 0 0 0 0 0 59.8 i.i 4.23 0 0 0 0 .304 0 0 0 1.03 0 70.7 i.i 0 0 0 1.32 0 0 0 0 0 0 0 83.5 i.i 0 0 .460 1.08 0 0 0 0 0 0 0 95.1 i.i 2.93 1.81 0 5.44 11.3 3.35 1.97 0 2.33 0 0 108 i.i 6.80 22.9 5.27 8.64 9.04 49.2 5.05 3.03 0 0 0 118 i.i 15.5 28.3 10.6 20.1 66.5 42.3 15.9 5.66 0 3.78 3.82 124 i.i 61.6 74.6 16.9 34.0 9.54 128 0 12.4 0 3.20 0 132 i.i 209 66.0 5.49 5.08 27.3 26.7 10.9 6.07 7.77 (*) 1.02 138 i.i 155 3.77 5.00 51.2 48.4 46.2 24.8 13.3 2.06 7.44 4.40 143 i.i (*) (*) 51.3 24.5 20.0 20.0 2.04 9.96 2.62 4.36 7.54 149 i.i 12.2 32.6 82.5 42.7 204 36.9 12.2 (*) 16.2 8.69 2.85 162 i.i (*) 131. 166. 79.6 224 27.5 3.36 18.3 38.7 .430 3.12 193 1.5 182 92.3 105. 290. 14.5 33.2 57.2 16.4 51.1 2.77 19.9 238 .5 13.4 26.0 71.0 81.2 20.4 18.2 52.4 15.7 13.4 16.5 285 .5 17.5 21.7 45.2 11.9 22.4 14.3 29.1 20.4 40.1 19.7 339 .5 14.5 15.4 10.4 37.4 54.8 57.1 104 105 86.2 1.05 356 .3 38.8 36.8 50.2 48.7 55.2 120. 5.71 10.5 19.7 4.11 429 .3 0 6.11 1.74 44.8 60.6 28.2 92.5 1.85 14.8 6.71 525 .3 20.2 19.0 30.4 12.8 24.2 (*) 28.6 25.2 1.32 3.01 597 .3 35.1 25.9 53.6 19.9 23.5 (*) .664 0 0 0 90 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO Table 27. — Concentration of monazite tracers in the 0.500- to 0.707-mm sieve class in “dustpan”samples collected at cross section 90 [Leaders (...) indicate no sample collected at that time and lateral position; (*) indicates sample was misplaced] Hours from Left edge injection of water Concentration times 105, at indicated lateral position (in m) 0») (m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 0.067 i.i 0.0 o 0.0 0.0 0.0 0.0 .270 i.i 0.0 0 0.0 0 0 0 0 .500 i.i 6.6 0 0 0 0 0 0 0 6.6 6.6 1.13 i.i 0 0 0 0 0 0 0 0 0 0 2.00 i.i 0 0 0 0 0 0 0 0 0 (*) 3.09 i.i 0 0 0 0 0 0 0 .677 0 0 4.05 i.i 0 0 0 0 0 0 0 0 0 0 5.06 i.i 0 0 0 0 0 2.52 0 0 0 0 6.00 i.i 0 0 0 0 0 0 0 0 0 0 8.06 i.i 0 0 0 0 (*) 0 0 0 0 0 10.0 i.i 0 0 0 0 0 0 0 0 0 0 12.0 i.i 0 0 0 0 0 0 0 0 0 0 22.0 i.i 0 0 0 0 0 0 0 0 0 0 34.0 i.i 0 0 0 0 0 0 0 0 (*) 0 47.0 i.i 6.6 0 0 0 0 0 0 0 0 0 0 59.8 i.i 0 0 0 0 0 0 0 0 0 0 0 70.7 i.i 0 0 0 0 0 0 0 0 0 0 0 83.5 i.i 0 0 0 0 0 1.73 0 0 0 0 0 95.1 i.i 0 0 2.37 5.71 4.89 2.88 0 0 0 0 0 108 i.i 0 64.1 3.41 3.03 3.86 5.30 0 9.14 0 0 0 118 i.i 44.4 19.6 6.41 9.77 15.2 3.40 0 6.55 0 0 0 124 i.i 48.0 112 0 58.6 3.54 160 0 5.49 0 0 19.9 132 i.i 230 81.1 0 1.39 21.8 32.0 8.91 12.5 20.4 (*) 13.1 138 i.i 230 69.3 12.6 62.5 31.0 18.6 26.4 13.0 0 0 0 143 i.i (*) (*) 50.6 19.1 33.2 41.5 2.77 0 2.73 0 0 149 i.i 0 0 114 25.7 185 16.2 3.74 (*) 0 0 41.2 162 i.i (*) 189 120 73.4 305 56.0 5.39 0. 6.42 0 34.7 193 1.5 576 148 122 550 64.0 (*) 50.8 24.0 77.8 4.71 31.5 238 .5 51.3 77.4 216 272 27.1 10.9 46.3 41.6 34.8 94.5 285 .5 51.3 37.6 81.9 9.68 44.6 16.1 77.3 93.8 134 57.2 339 .5 39.0 52.2 7.14 118. 56.0 89.5 330 440 206 23.6 356 .3 49.4 101 260 58.5 20.4 146 23.2 57.6 212 9.30 429 .3 0 19.2 0 60.0 35.6 666 110. 8.42 54.9 53.6 525 .3 26.4 10.7 0 7.20 31.0 12.1 149. 40.5 16.0 0 597 .3 35.4 32.2 74.6 36.1 49.4 (*) 0 0 0 0 Table 28. — Concentration of lead tracers in the 0.125- to 0.177-mm sieve class in “dustpan" samples collected at cross section 90 [Leaders (...) indicate no sample collected at that time and lateral position; (*) indicates sample was misplaced] Hours from injection (h) Left edge of water Concentration times 10s, at indicated lateral position (in m) (m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 0.067 i.i 0.424 o 0.0 0.0 0.0 0.0 .270 i.i o.o 0 0.0 0 0 0 0 .500 i.i 6.6 0 0 0 (*) .652 0 0 6.6 0.0 1.13 i.i 0 0 0 0 0 0 0 1.17 0 0 2.00 i.i 0 0 0 0 0 0 0 0 0 (*) 3.09 i.i 0 0 0 0 0 0 0 0 0 0 4.05 i.i 0 0.227 0 0 0 0 0 0 0 0 5.06 i.i 0 0 0 0 0 .385 0 2.10 0 0 6.00 i.i 0 0 0 0 0 0 0 0 0 0 8.06 i.i 0 0 0 0 (*) .377 0 .403 0 0 10.0 i.i 0 0 0 0 0 0 0 0 0 1.64 12.0 i.i 0 0 0 0 0 0 0 0 0 0 22.0 i.i 0 0 0 0 0 0 0 0 0 0 34.0 i.i 0 0 0 0 0 0 0 0 (*) 0 47.0 i.i 6.199 0 0 0 0 0 0 0 .266 0 0 59.8 i.i 0 0 0 0 .846 0 0 0 0 0 0 70.7 i.i 0 0 0 0 0 0 0 0 0 0 .711 83.5 i.i .465 0 0 0 0 0 0 0 0 0 0SUPPLEMENTAL DATA 91 Table 28. — Concentration of lead tracers in the 0.125■ to 0.177-mm sieve class in “dustpan” samples collected at cross section 90 — Continued Hours from injection (h) Left edge of water Concentration times 105, at indicated lateral position (in m) (m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 95.1 i.i 0 0 0 0 0 0 0 0 0 0 0 108 i.i 0 0 0 0 7.06 0 0 0 0 0 0 118 i.i 0 0 0 0 0 0 0 0 0 0 0 124 i.i .476 6.10 0 0 0 0 0 0 0 0 0 132 i.i 0 0 0 0 0 0 3.70 0 0 0 0 138 i.i 0 0 0 0 0 .618 0 0 .177 0 0 143 i.i (*) (*) 0 .984 0 0 .254 0 0 0 0 149 i.i 0 0 .363 .846 0 1.33 0 (*) .425 0 0 162 i.i (*) 0 0 0 26.6 1.67 0 0 0 0 0 193 1.5 .372 2.47 1.69 0 0 .501 0 32.0 148. 11.9 67.1 238 .5 0 0 0 .597 0 1.27 0 .544 0 0 285 .5 0 0 0 0 0 .459 0 0 0 0 339 .5 0 2.22 .317 0 0 0 0 0 0 .189 356 .3 0 0 .200 0 0 0 0 0 0 0 429 .3 0 0 0 0 0 2.05 10.2 0 0 0 525 .3 0 0 0 .169 0 0 0 0 0 0 597 .3 0 0 0 .258 0 <*) 0 0 .312 0 Table 29. — Concentration of lead tracers in the 0.177■ to 0.250-mm sieve class in “dustpan” samples collected at cross section 90 ILeaders (...) indicate no sample collected at that time and lateral position; (*) indicates sample was misplaced) Hours from Left edge of water Concentration times 105, at indicated lateral position (in m) (h) (m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 0.067 i.i 0.0 (*) (*) 0.0 0.0 0.0 .270 i.i o.o 0 0.0 0.0 0 0 0 .500 i.i 6.6 0 0 0 0 0 0 0 o.o 0.0 1.13 i.i 0 0 0 0 0 0 0 .358 0 0 2.00 i.i 0 0 0 0 0 0 0 0 0 (*) 3.09 i.i 0 0 0 0 0 0 0 0 0 0 4.05 i.i 0 0 0 0 0 0 0 0 0 0 5.06 i.i 0 0 0 0 0 0 0 .944 0 0 6.00 i.i 0 0 0 0 0 0 0 0 0 0 8.06 i.i 0 0 0 0 (*) 0 0 .395 0 0 10.0 i.i 0 0 0 0 0 0 0 0 0 .493 12.0 i.i 0 0 0 0 0 0 0 0 0 0 22.0 i.i 0 0 0 0 0 0 0 0 0 0 34.0 i.i 0 0 0 0 0 0 0 0 <*) 0 47.0 i.i 6.6 0 0 0 0 0 0 0 0 0 0 59.8 i.i 0 0 0 0 0 0 0 0 0 0 0 70.7 i.i 0 0 0 0 0 0 0 0 0 0 0 83.5 i.i 0 0 0 0 0 0 .109 0 0 0 0 95.1 i.i .141 0 0 0 0 0 0 0 0 0 0 108 i.i 0 .110 .536 0 0 0 0 0 0 0 0 118 i.i <*) 0 0 0 0 0 0 0 0 0 0 124 i.i 0 9.09 0 0 0 0 0 0 0 0 0 132 i.i 0 0 0 .967 0 0 0 0 0 0 0 138 i.i .249 0 0 0 0 0 0 0 0 0 0 143 i.i (*) (*) 0 0 0 0 0 0 0 0 0 149 i.i 0 0 0 0 0 .489 0 (*) .119 0 0 162 i.i (*) 0 0 0 16.0 0 0 0 0 0 0 193 1.5 0 33.2 4.73 0 0 0 0 434. 12.0 87.5 4.85 238 .5 0 0 0 0 3.83 0 0 0 0 0 285 .5 0 0 0 0 7.49 6.00 0 0 0 0 339 .5 0 2.00 0 .167 .584 .136 0 0 0 0 356 .3 0 0 0 0 0 0 0 0 0 0 429 .3 0 0 0 0 0 0 1.68 0 0 0 525 .3 0 .105 0 0 .498 .260 0 0 0 0 597 .3 0 0 0 0 0 (») 0 0 0 0 92 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO Table 30. — Concentration of lead tracers in the 0.250- to 0.350-mm sieve class in “dustpan” samples collected at cross section 90 [Leaders (...) indicate no sample collected at that time and lateral position; (*) indicates sample was misplaced] Hours from Left edge injection of water Concentration times 105, at indicated lateral position (in m) * (h) (m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 0.067 i.i 0.0 (*) (*) 0.0 0.0 0.0 .270 i.i 6.6 0 0.0 0.0 0 0 0 6!o .500 i.i 6.6 0 0 0 0 0 0 0 6.6 1.13 i.i 0 0 0 0 0 0 0 .334 0 0 2.00 i.i 0 0 0 0 0 0 0 0 0 (*) 3.09 i.i 0 0 0 0 0 0 0 0 0 0 4.05 i.i 0 0 0 0 0 0 0 0 0 0 5.06 i.i 0 0 0 0 0 0 0 .785 0 0 6.00 i.i 0 0 0 0 0 0 0 0 0 0 8.06 i.i 0 0 0 0 0 0 .134 0 0 10.0 i.i 0 0 0 0 0 0 0 0 0 .276 12.0 i.i 0 0 0 0 0 0 0 0 0 0 22.0 i.i 0 0 0 0 0 0 0 0 0 0 34.0 i.i .140 0 0 0 0 0 0 0 (*) 0 47.0 i.i 6.6 0 0 0 0 0 .143 0 0 0 0 59.8 i.i 0 0 0 0 0 0 0 0 0 0 0 70.7 i.i 0 0 0 0 0 0 0 0 0 0 0 83.5 i.i 0 0 0 0 0 0 0 0 0 0 0 95.1 i.i 0 0 0 0 0 .196 0 0 0 0 0 108 i.i 0 0 0 0 0 .391 0 0 0 0 0 118 i.i 0 0 0 0 0 .391 0 0 0 0 0 124 i.i 0 0 0 0 0 1.56 0 0 0 0 0 132 i.i 0 0 0 0 0 0 0 0 0 0 0 138 i.i 0 0 0 0 0 0 0 0 0 0 0 143 i.i O (*) 0 0 0 0 0 0 0 0 0 149 i.i 0 0 0 0 0 0 0 (*) 0 0 0 162 i.i (*) 0 0 0 37.0 0 0 0 0 0 0 193 1.5 0 2.05 6.19 0 0 .586 0 2.62 218 .510 16.6 238 .5 0 0 0 0 0 1.37 .352 0 0 0 285 .5 0 0 0 0. 4.37 0 0 0 0 0 339 .5 0 0 0 0 0 0 1.90 0 .243 0 356 .3 0 0 0 0 .440 0 0 .156 .339 0 429 .3 0 0 .137 0 0 7.24 7.74 .480 0 0 525 .3 0 0 0 0 3.70 0 0 0 0 0 597 .3 0 0 0 0 0 (*) 0 0 0 0 Table 31. — Concentration of lead tracers in the 0.350- to 0.500-mm sieve class in “dustpan" samples collected at cross section 90 I Leaders (...) indicate no sample collected at that time and lateral position; (*) indicates sample was misplaced] Hours from injection (h) Left edge of water Concentration times 10s, at indicated lateral position (in m) (m) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12 13.5 15 16.5 0.067 i.i 0.0 (*) (*) 0.0 0.0 0.0 .270 i.i 6.6 0 0.0 0.0 0 0 0 .500 i.i o.o 0 0 0 0 0 0 0 0.0 6.6 1.13 i.i 0 0 0 0 0 0 0 .749 0 0 2.00 i.i 0 0 0 0 0 0 0 0 0 :*) 3.09 i.i 0 0 0 0 (*) 0 0 0 0 0 4.05 i.i 0 0 0 0 0 0 0 0 0 0 5.06 i.i 0 0 0 0 0 0 0 4.18 0 0 6.00 i.i 0 0 0 0 0 0 0 0 0 0 8.06 i.i 0 0 0 0 (*) 0 0 0 0 0 10.0 i.i 0 0 0 0 0 0 0 0 0 0 12.0 i.i 0 0 0 0 0 0 0 0 0 0 22.0 i.i 0 0 0 0 0 0 0 0 0 0 34.0 i.i 0 0 0 0 0 0 0 0 (*) 0 47.0 i.i 6.6 0 0 0 0 0 0 0 0 0 0 59.8 i.i 0 0 0 0 0 0 0 0 0 0 0 70.7 i.i 0 0 0 0 0 0 0 0 0 0 0 83.5 i.i 0 0 0 0 0 0 0 0 0 0 0SUPPLEMENTAL DATA 93 Table 31.__Concentration of lead tracers in the 0.350- to 0.500-mm sieve class in “dustpan”samples collected at cross section 90 — Continued Hours from Left edge injection of water Concentration times 105, at indicated lateral position (in m)___________ (h) (m) L5 31) O (TO 7J5 9l) HX5 12 13A 95.1 1.1 0 0 0 0 0 0 0 0 0 0 0 108 1.1 0 0 0 0. 0 1.03 0 0 0 0 0 118 1.1 0 0 0 0 0 0 0 0 0 0 0 124 1.1 0 0 0. 0 0 1.37 0 0 0 0 0 132 1.1 0 0 0 0 0 1.02 0 0 0 0 0 138 1.1 0 0 0 0 0 0 0 0 0 0 0 143 1.1 (’) (*) 0 0 0 0 0 0 0 0 0 149 1.1 0 0 0 0 0 0 0 (*) 0 0 0 162 1.1 (*) 0 1.33 0 0 0 0 0 0 0 0 193 1.5 0 4.53 15.1 0 0 0 0 5.28 260 0 13.7 238 .5 0 0 0 0 6.56 0 0 0 0 0 285 .5 0 0 0 0 C .304 0 0 0 0 339 .5 0 1.58 0 0 0 0 0 1.17 0 0 356 .3 0 0 0 0 .750 0 0 0 0 0 429 .3 0 0 0 0 0 17.8 4.17 0 0 0 525 .3 0 0 0 0 2.00 (*) 0 5.85 0 0 597 .3 0 0 0 0 0 (*) 0 0 0 0 Table 46. — Concentrations of fluorescent tracers for different minerals and sieve classes in core samples collected on May 8 I Leaders (...) indicate no sample collected at that time and lateral position. '*) indicates sample was misplaced. | Concentration times 10r> Quartz Lead section 0.125 - 0.177 - 0.250 - 0.350 - 0.500 - 0.707 - 0.125 - 0.177 - 0.250 - 0.350 - 0.177 0.250 0.350 0.500 0.707 1.00 0.177 0.250 0.350 0.500 7.5 17.5 26.7 70.1 439 1140 411 165 38.4 28.7 41.0 15 27.0 33.9 68.7 320 783 200 131 33.6 19.1 20.7 30 17.8 28.3 60.6 249 719 124 32.6 10.9 5.94 4.75 60 20.6 23.8 45.5 155 294 74.0 5.88 2.57 3.76 1.20 90 22.1 33.4 39.0 92.6 117 16.9 1.15 .478 1.15 .411 180 18.4 20.0 14.3 1.62 1.22 0 .489 .142 .107 .0602 270 16.8 12.4 1.64 .340 .140 .592 .172 .240 .0666 .0282 360 15.1 9.82 .676 .621 .650 .927 .959 .314 .171 .333 450 15.4 8.52 .390 .372 .964 .369 .210 .220 .0973 .130 540 14.0 3.17 .605 .713 .385 .878 .140 .0832 .0378 .0383 Concentration times 10 ' Garnet Monazite .section (m) 0.125 - 0.177 - 0.250 - 0.350 - 0.500 - 0.125- 0.177- 0.250- 0.350- 0.500- 0.177 0.250 0.350 0.500 0.707 0.177 0.250 0.350 0.500 0.707 7.5 11.7 33.3 67.2 134 25.9 91.2 41.3 27.6 102 88.9 15 15.7 36.8 74.9 93.1 28.5 82.3 40.7 31.3 76.0 68.4 30 10.2 28.4 34.2 46.2 13.2 68.4 38.0 16.8 40.0 52.1 60 12.4 17.0 29.6 31.2 5.71 55.0 23.3 14.5 32.0 35.7 90 6.91 12.1 21.0 17.1 2.64 33.0 14.0 16.6 15.1 15.0 180 2.92 2.62 .912 .139 .343 14.4 3.01 .191 .0902 .114 270 1.19 .519 .349 .223 .319 4.25 .399 .109 .0956 0 360 2.12 .402 .255 .294 .243 3.47 .495 .653 .294 1.17 450 .418 .342 .284 .233 .996 1.41 .258 .790 .922 .158 540 .448 .344 .515 .744 .670 1.70 1.36 .187 .859 .50294 TRANSPORT AND DISPERSION OF PARTICLES, ATRISCO FEEDER CANAL, NEW MEXICO T able 47. — Concentrations of fluorescent tracers for different minerals and sieve classes in core samples collected on June 6 Concentration times 105 Cross Quartz Lead section (m) 0.125- 0.177- 0.250- 0.350- 0.500- 0.707- 0.125- 0.177- 0.250- 0.350- 0.177 0.250 0.350 0.500 0.707 1.00 0.177 0.250 0.350 0.500 30 4.95 4.12 3.64 12.0 52.2 10.2 139 309 101 72.7 60 1.59 2.06 3.84 24.4 107 36.8 58.3 38.2 10.1 6.05 90 6.37 4.82 5.52 30.8 99.2 26.5 10.4 4.99 2.62 1.64 120 3.18 2.46 4.66 29.5 102 19.1 3.20 .260 .293 .446 150 2.76 4.60 9.92 72.0 267 58.8 1.67 1.00 .338 1.03 180 4.03 3.47 7.84 60.5 185 59.9 1.79 1.05 1.36 .618 210 .759 2.22 9.44 74.0 176 34.2 .128 1.20 2.01 1.58 240 .629 1.64 8.68 51.5 156 24.4 .415 .0629 .0258 .0228 270 1.08 2.32 13.3 82.4 183 41.6 .305 .113 .0365 .319 300 1.16 2.19 11.1 72.1 130 25.7 .309 .154 .143 .0803 330 1.90 3.45 13.9 72.7 99.2 12.1 .277 .130 .0573 0 360 2.78 3.67 14.7 48.2 63.1 5.10 .394 .0576 .0295 0 390 .874 2.80 15.5 51.8 58.3 8.40 .284 .0116 .0285 .195 420 2.19 4.20 15.5 42.7 24.3 3.02 .0348 .0102 0 0 450 1.18 4.40 18.7 39.9 45.5 1.22 .339 .106 .0572 .0360 480 1.42 4.19 13.1 14.5 4.83 .870 .254 .0431 0 .0446 510 1.56 6.19 12.8 9.60 2.52 0 .303 .0536 .117 0 540 1.37 3.96 12.5 18.4 .877 0 .103 .156 0 0 570 2.21 6.14 10.1 10.7 .325 0 .0570 0 .0609 0 600 2.02 3.32 6.56 2.41 .660 0 .269 .0357 .0742 0 630 2.57 4.51 4.68 1.82 .960 0 .0888 .0335 0 .0919 660 2.70 5.79 5.64 .457 .559 0 .0728 .0104 .0145 .0525 690 1.25 4.57 3.88 .300 .181 0 .124 .0208 0 .0370 720 2.49 3.39 2.57 .223 1.66 1.59 .0815 .0332 .0158 0 750 1.40 4.56 2.80 .535 .385 0 .0463 .111 0 0 780 1.72 8.07 3.24 .268 .295 0 .0337 .646 0 0 810 1.44 3.54 2.25 .478 .440 0 .0590 .0431 .0958 .0368 840 1.35 3.49 1.50 .370 .447 0 .0408 0 0 0 870 1.41 3.02 .923 .194 .183 0 .0874 .0118 0 0 Concentration times 105 Cross Garnet Monazite section (m) 0.125- 0.177- 0.250- 0.350- 0.500- 0.125- 0.177- 0.250- 0.350- 0.500- 0.177 0.250 0.350 0.500 0.707 0.177 0.250 0.350 0.500 0.707 30 T75 2l5 6.47 9.18 2.47 8.13 6.37 5.48 14.7 13.1 60 1.69 3.91 5.86 15.1 4.95 17.0 8.08 5.14 21.3 31.1 90 4.13 7.98 10.1 18.0 3.63 26.6 11.6 7.01 16.6 20.6 120 1.70 2.75 5.91 13.6 3.43 12.9 5.57 3.33 12.8 11.1 150 2.54 6.03 11.7 30.7 12.7 22.4 10.8 8.55 27.4 38.2 180 2.95 7.78 17.2 27.5 8.99 18.9 12.9 10.9 25.8 210 2.86 8.00 16.7 30.2 4.94 23.5 11.9 10.3 24.3 24.4 240 1.43 4.16 4.62 10.7 11.1 11.4 7.49 5.20 10.7 19.2 270 1.98 4.25 8.92 13.0 1.94 15.5 7.26 4.30 8.93 14.1 300 1.17 3.29 7.17 9.94 2.90 11.2 5.60 4.18 11.0 6.39 330 2.12 5.86 8.49 8.69 .406 15.1 8.51 6.08 9.08 4.46 360 1.95 5.61 9.04 8.51 .224 18.5 9.22 4.80 8.58 7.68 390 2.04 4.96 6.88 5.88 .476 12.2 6.52 3.33 4.62 1.53 420 3.76 7.82 8.22 4.31 .458 23.1 8.74 2.93 2.92 2.66 450 2.44 4.50 4.90 2.50 3.44 14.9 5.36 2.32 1.41 2.51 480 1.71 2.25 2.08 1.85 .214 9.72 3.62 .866 .398 .627 510 2.88 2.18 4.28 .669 .166 9.20 2.68 4.80 .543 .322 540 1.64 2.26 2.09 .965 0 11.3 3.25 .626 .150 .275 570 2.34 1.39 1.28 1.68 0 9.56 2.16 .606 .298 0 600 .958 1.06 .630 .436 0 4.38 .834 .144 .129 .902 630 1.83 1.49 .877 .666 .134 7.30 1.74 .0974 .142 .260 660 1.76 .981 .890 .304 .109 5.79 .981 .142 .0746 0 690 .786 .660 .352 .0716 .0889 4.36 .654 .0369 .0794 0 720 .942 .451 .379 .106 .437 3.03 .309 .122 0 .203 750 .983 .890 .593 .364 0 2.49 .416 .0879 .177 0SUPPLEMENTAL DATA 95 Table 47. — Concentrations of fluorescent tracers for different minerals and sieve classes in core samples collected on June 6 — Continued Cross section (m) 0.125- Garnet 0.177- 0.250- Concentration times 105 0.350- 0.500- 0.125- 0.177 0.250 780 810 840 870 1.20 1.49 1.41 .386 .753 .515 .685 .723 0.350 .443 .407 .274 .112 0.500 .279 .0679 .0363 0.707 .228 .438 0 0 0.177 4.20 5.18 2.14 1.31 Monazite 0.177- 0.250 0.250- 0.350 0.350- 0.500 0.500- 0.707 .646 .0580 0 .442 .266 .108 .0398 .851 .216 .0373 .0183 0 .172 .0177 0 0 Table 48. — Concentrations of fluorescent tracers for different minerals and sieve classes in core samples collected on July 14 Concentration times 10 Cross Quart/ Lead section (m) 0.125- 0.177- 0.250- 0.350- 0.500- 0.707- 0.125- 0.177- 0.250- 0.350- 0.177 0.250 0.350 0.500 0.707 1.00 0.177 0.250 0.350 0.500 30 1.03 1.59 2.54 17.0 70.0 18.8 30.3 10.7 1.94 0.562 90 .867 1.42 4.28 24.2 67.8 19.1 14.8 6.37 2.93 1.45 150 .502 .805 2.17 12.4 48.6 20.0 .178 .105 .0758 .144 210 1.60 3.40 8.05 19.2 61.6 13.9 .545 .392 .853 .679 270 .369 1.14 3.94 34.3 113 28.6 .218 .0618 .0815 .0653 330 .359 .597 4.64 40.7 122 24.3 .0578 .115 0 .0484 390 1.37 1.00 5.84 49.8 154 40.8 .493 .115 .0665 .0398 450 .415 .796 4.50 43.8 133 17.7 .207 .160 .155 .354 510 .688 2.17 8.26 48.0 95.0 18.6 .390 .130 .102 .128 570 .559 2.05 9.72 51.0 96.5 17.6 .245 .255 .478 0 630 .559 1.57 7.56 34.8 45.8 6.49 .416 .118 .0660 .132 690 .285 1.59 7.84 30.9 24.8 6.82 .112 .0071 .0684 .197 750 1.30 3.06 14.8 30.8 19.1 1.97 .856 .347 .534 .305 810 .356 2.27 10.7 17.7 10.3 2.72 .0523 .0545 .0255 .0313 870 .615 1.71 6.63 9.92 3.51 .585 .0333 .0303 .0853 .0363 Concentration times 10 Cross Garnet Monazite section (m) 0.125- 0.177- 0.250- 0.350- 0.500- 0.125- 0.177- 0.250- 0.350- 0.500- 0.177 0.250 0.350 0.500 0.707 0.177 0.250 0.350 0.500 0.707 30 4.23 3.39 4.98 7.92 8.24 6.25 3.13 2.23 5.05 9.95 90 1.66 3.89 9.64 14.6 2.86 15.4 5.91 5.73 14.1 19.0 150 .450 1.07 2.52 4.99 .942 3.36 1.83 1.54 5.48 7.97 210 1.13 2.74 5.56 14.8 4.17 6.37 4.11 3.74 12.0 8.22 270 1.36 4.40 7.81 15.2 4.94 13.4 5.17 4.36 11.8 18.4 330 .848 2.03 5.58 11.5 2.59 9.33 3.43 3.63 10.7 9.98 390 1.43 3.34 7.81 13.2 1.94 12.1 4.57 4.25 8.69 12.6 450 .915 3.14 6.64 14.4 3.62 9.19 5.01 4.77 12.3 11.9 510 1.44 3.48 4.48 7.47 2.98 8.46 3.18 2.02 4.93 4.47 570 1.40 3.72 6.35 8.12 1.39 8.38 3.92 2.56 4.20 4.31 630 .832 1.53 2.10 3.56 1.46 4.87 2.17 1.40 2.47 .782 690 .884 2.74 3.47 3.50 .416 7.81 3.07 1.39 1.38 .875 750 2.68 4.94 4.83 6.24 1.74 14.8 5.57 2.22 1.69 .252 810 1.16 2.29 2.04 1.14 .0895 7.36 2.40 .992 .746 .899 870 .791 1.24 1.23 1.17 1.30 6.40 1.10 .668 .281 0CD ac LU 0.3 O 0.6 < LL DC D CO CO rr LL) LU z 0.9 H 3 < “5 LU DQ LU O z < H CO Q 1.2 1.5 0.3 0.6 SEGMENT NUMBER 5 4 3 2 1 J_____I___I 0 4 8 12 J_____L 0 8 16 h J_____L 0 20 40 I I I 0 1 2 J__I__I_L 0 0.5 1.0 0 10 20 0 6 12 J___L 0 1 2 0 2 J__l 4 _____I____I____I___I____I 0 4 8 0 2 4 0 2 J__I 4 J___I_I__L X 0 0.4 0.8 0 8 16 0 0.4 0.8 J___I__I__I J____I____L I I I I I 0 4 8 0 0.6 1.2 0 2 4 I I I I J___I I I J__L o 1 0 1.6 3.2 0 1 2 0 12 3 0 2 4 > 0.9 D 1.2 I I I 0 2 4 6 J_____I_____I 0 4 8 12 J___I___I__I 0 2 4 I I I I 0 2 4 0 2 4 J___I_I__I__L I I___________________________________________________________________I__________________________________I 0 6 12 0 2 4 15o 0 0.4 0.8 _L ----------------L 0 12 3 FLUORESCENT TRACER CONCENTRATION TIMES 105, IN GRAMS PER GRAM J__________I__________I_________L i i i 0 0.30 0.60 J___I_L I I I I I I I I 0 0.4 0.8 J___I_I__I__L 0 2 4 0 4 8 0 0.4 0.8 1 2 0 12 3 30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 DISTANCE DOWNSTREAM FROM SOURCE, IN METERS 540 570 600 630 660 690 720 750 780 810 840 870 900 0.6 0.9 T T T T < 12 1.5 - 1.8 I I I I I 0 160 320 I I I I_____________________________________________________________________________________L 0 40 80 J___L 0 2 4 J___I__I__I_i J__I__I_I__L 0 20 40 J___I___I__I___I 0 0.4 0.8 X 0 100 200 0 200 400 J___LJ____l_J___I 0 6 12 18 B, 0.250- to 0.350-mm sieve class of quartz tracer J___I__L 0 0.4 0.8 co DC LU I— LU 0.3 LLI o 0.6 < LL DC D CO CD cc LU lu Z 0.9 I- 3 < "> 5 1.2 O LU m LU o z < I— 1 5 I— CO o 0.3 0.6 ■4- > 0.9 3) —) 1.2 SEGMENT NUMBER 5 4 3 2 1 J___I_I__L 0 4 8 I I I I I 0 8 16 J___I_I__I__L 0 10 20 J___l_J___I__I 0 20 40 I I I I I I I I I I 0816 0 8 16 0 10 20 0 6 12 J___L l l l i i J___I_I__I__I 0 10 20 0 10 20 0 6 12 0 10 20 J___I__I_I__I 0 6 12 J___I_I__I__I I I I_____________________________________________L 0 4 8 J___I_L 0 4 8 I I I 0 4 8 0 2 4 J___L 0 2 _J___I 4 0 1 2 0 1.6 3.2 0 1 2 J____I__I___L 0 0.8 1.6 0 10 20 1.5 0.6 0.9 I I I I I 0 10 20 _L 0 30 60 l.l I I J___I__I l I 16 32 J___I__L 0 4 8 0 2 4 J___I_L J___I_L 0 6 12 0 4 8 J I I _L. I 5 10 30 60 90 120 150 180 210 240 270 300 330 360 _l____I___I 0 10 20 30 FLUORESCENT TRACER CONCENTRATION TIMES 10s, IN GRAMS PER GRAM J_____________I___________I____________I____________I____________L_ 0 8 16 J___I_I__L J___I_I__L I I I I I 0 10 20 J___I_I I I 0 8 16 0 8 16 0 6 12 0 6 12 390 420 450 480 510 DISTANCE DOWNSTREAM FROM INJECTION POINT, IN METERS 540 570 600 630 660 690 720 750 780 810 840 870 900 CD > 1.2 - 1.5 1.8 L- T T T T J 0 2000 4000 X 0 100 200 I I I I_____________________________________________________________I 0 800 1600 C, 0.500- to 0.707-mm sieve class of quartz tracer 0 3000 6000 CO cc W 0.3 O 06 < LL CC V) <° £ Z 0.9 6? O _l LU DQ LU O z < I— CD Q 1.2 1.5 0.3 0.6 ■sf >- 0.9 => —3 1.2 1.5 SEGMENT NUMBER 5 4 3 2 1 XJ__L 0 60 120 0 60 120 0 80 160 X 111,0 200 400 o 120 240 I I I_________________________________________________________________I_____________________L I I I 120 240 0 80 160 J___I__L 0 8 16 0 20 40 0 15 30 0 40 80 I I I_____________________________________________L 0 300 600 _L 0 240 480 J___I_I__I__L 0 24 48 X_J__LJ___L 0 60 120 0 20 40 _L 0 100 200 J___L I I I________________________________________________________________L 0 60 120 30 60 90 120 150 180 210 240 270 300 330 360 0 100 200 0 100 200 FLUORESCENT TRACER CONCENTRATION TIMES 10s, IN GRAMS PER GRAM J_____________I____________I___________I_____________I_________ J___I_I__L 0 20 40 60 J__I__I_I III I I I J___I__L 0 8 16 I I 0 60 120 0 30 60 0 8 16 0 60 120 390 420 450 480 510 DISTANCE DOWNSTREAM FROM SOURCE, IN METERS 540 570 600 630 660 690 720 750 780 810 840 870 900 VERTICAL CONCENTRATION DISTRIBUTIONS OF THE 0.125- TO 0.177-mm, 0.250- TO 0.350-mm, AND 0.500- TO 0.707-mm SIEVE CLASSES OF QUARTZ TRACER AS DEFINED BY CORE SAMPLES COLLECTED ALONG THE CENTERLINE OF THE CHANNEL ON MAY 8, JUNE 6, AND JULY 14, ATRISCO FEEDER CANAL NEAR BERNALILLO, NEW MEXICO 7 DAYS U.S. Ol->' r