ALBERT R. MANN 
 LIBRARY 
 
 New York State Colleges 
 
 OF 
 
 Agriculture and Home Economics 
 
 AT 
 
 Cornell University 
 
^iSiii™,,™?,* °' '-^•'" ^"1 Greek delf^ 
 
 3 1924 002 803 751 
 
Cornell University 
 Library 
 
 The original of tiiis book is in 
 tine Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924002803751 
 
CONTRIBUTIONS TO BIOLOGY 
 
 FROM 
 
 The Hopkins Seaside Laboratory 
 
 OF THE 
 LELAND STANFORD JR. UNIVERSITY. 
 
 s: 
 
 Scientific Names of Latin and 
 Greek Derivation. 
 
 BY 
 
 Walter Miller, 
 
 Professor of Classical Phiiology, Leland Stanford Jr. University. 
 
 Reprinted from Proceedings of the California Academy of Sciences, 
 3dSer., Zool., Vol. I. 
 
 SAN FRANCISCO, CAL. 
 
 : 1897- 
 
PREFATORY NOTE. 
 
 This memoir is the tenth of a series designed to illustrate 
 investigations and explorations connected with the Hopkins 
 Seaside Laboratory, an adjunct of the biological labora- 
 tories of the Leland Stanford Junior University. These 
 investigations have been carried on by means of the assist- 
 ance given of Timothy Hopkins, Esq., of Menlo Park, 
 California. The memoirs forming this series are published 
 in the Proceedings of the California Academy of Sciences, 
 the present edition being a reprint from these proceedings. 
 
 Charles H. Gilbert, 
 Oliver P. Jenkins, 
 Vernon L. Kellogg, 
 
 Directors Hopkins Laboratory. 
 Date of publication, April lo, 1897. 
 
SCIENTIFIC NAMES OF LATIN AND GREEK 
 DERIVATION. 
 
 BY WALTER MILLER, 
 Professor of Classical Philology, Leland Stanford Jr. University. 
 
 While the first volume of Jordan and Evermann's "Fishes 
 of North and Middle America" was in preparation, Dr. 
 Jordan invited me to review and verify the etymologies of the 
 names of the fishes described. In the course of this review 
 some of the inconsistencies into which nomenclators have 
 fallen in the construction of names of classical derivation 
 came frequently under discussion. Two sets of words 
 seemed to cause more trouble than any others : what should be 
 the gender and inflection of compounds in -pus (genitive, 
 -podis) and -stoma, -soma, etc. (genitive, -stomatis, 
 -somatis, etc.), and what the form of the family names to 
 be derived from them, were the questions that called for an 
 immediate answer. As thorough an investigation as possi- 
 ble followed and a satisfactory conclusion was reached. 
 But the examination of the problems these questions pre- 
 sented naturally suggested others in regard to the forma- 
 tion, gender and inflection of New Latin compound names; 
 for most of the irregularities in scientific nomenclature are 
 of that sort. Here and there we find masculine nouns in -a 
 (genitive, -ae) masquerading as feminine, feminine nouns 
 in -us (genitive, -i) and neuters in-us (genitive, -us, Greek 
 -on?) as masculine, and neuters in -ma (genitive, -matis, 
 Greek -iMaro's) as feminines in -a (genitive, -ae). Again, 
 through a misconception of the inflection of a word it often 
 appears in compounds and in family names in a sadly muti- 
 lated form. 
 
 Various scientific writers have arbitrarily departed from 
 the philologically correct method of nomenclature estab- 
 lished by Linnaeus;' moreover some difference of opinion 
 
 'Even Linnffius, however, is not absolutely free from errors in translitera- 
 tion: e. g. Systema Natures, P^: Oxy-rinchus (pp. 395 and 512), Callo- 
 rynchus (p. 402), Gono-rynchus (p. 528), Di-delphis (p. 71), Syn- 
 gnathus, as masculine (p. 416). 
 
 [llSl Aprils, 1897. 
 
Il6 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 now prevails in regard to the formation, gender and inflec- 
 tion of certain New Latin words derived from the Greek. 
 Definite rules have been wanting, or at least not readily 
 available. Accordingly, at Dr. Jordan's request, and with 
 his kind assistance, I have undertaken to formulate a set of 
 rules based upon philological principles and at the same 
 time agreeing with the practice of fconsistent nomenclators. 
 Ultra-purism, however, as the writing of ai and oi for the 
 Greek ai and ot or of k for Greek k, shall have no more 
 consideration than the philological monstrosities produced 
 by a Rafinesque or a Swainson. 
 
 There is no essential difference in the method of proced- 
 ure whether we are adopting from the classical languages 
 simple words or compound words ; but as not all the desired 
 compounds exist ready made for us in Greek and Latin, and 
 as their formation causes more trouble, the compounds will 
 be assumed as the subject of these rules. 
 
 Compounds in scientific terminology, whether Latin or 
 Greek, are usually made up of two parts, which would nat- 
 urally be separate words standing in some syntactical rela- 
 tion to each other and which combine that relation in one 
 word; either member may be a form that is not a word by 
 itself but appears only in compounds: e. g. Eng. M«-told; 
 un- appears only as inseparable prefix; in •a^-ward -ward 
 appears only as inseparable suffix. So Latin we-sanus (in- 
 separable prefix ve- +sanus) out of one's senses; inn-am- 
 bulus, rope-walker; Greek i v a -aka>TO'i , hard to be caught; 
 TToXv-ir p dry fiwv (ttoXv, much, -\- TrpdrTco, do), meddlesome. 
 Besides these two elements, a third is in every case to be 
 taken into consideration — the inflectional ending, which de- 
 termines the sentence relation of the word; this inflectional 
 ending may or may not be identical with that of the last 
 member of the compound: e. g. in Xev«-o'(^0aX/io? (Xeu/eo?, 
 white,-\- oj>0a\,iJi,6<s, eye), white-eyed, and in YleKoirov-v-qaot; 
 (Tie\oTro'i,Pelofs' , -j-yTjo-o?, island), Peloponnesus, it is; but 
 it is not so in /j,aKpo-Ke(j)a\o^ , long-headed, derived from 
 ftaKpoi, long,-\- Ke<^akri, head, nor in iMaKp-ovpo<i, long-tailed, 
 derived from fiaKp6<;, long, -\-ovpd, tail; but in passing into 
 
ZOOL— Vol. Ij MILLER— GREEK AND LATIN DERIVATIVES. I17 
 
 a compound adjective the stem ice<f>aXa- of Ke^aXri and the 
 stem ovpa- became, after the irresistible analogy of the o- 
 stems, Ke<^aKo- and ovpo-, and to those stems the masculine 
 nominative case-ending -9 was added. Such changes will 
 be more apparent as we proceed. 
 
 Hybrid Words. 
 
 In building a compound name the two classical languages 
 should never be mixed. The new word should be wholly 
 Greek or wholly Latin. Hybrid words are always objec- 
 tionable, and such compounds as Swainson's Felichthys, 
 asLongicephalus, Leptogunnellus, Flavigass- 
 ter; Gymnocorvus, Arboro-phila and the like, are 
 enough to make one's hair stand on end. 
 
 Latin Compounds. 
 
 Latin is, comparatively speaking, poor in compound 
 words. In place of doubling up words, significant sufHxes 
 are added, or the words, retaining their proper syntactical 
 relations, are simply written together as one: e. g. pater- 
 familias, iuris-dictio, etc. Still, the language contains 
 many genuine compounds of every part of speech — substan- 
 tives, verbs and adverbs. 
 
 In entering into composition as first member of a com- 
 pound, a noun or adjective appears not in its familiar, nom- 
 inative singular form, but as its stem or a modification of its 
 stem.^ If the stem ends in a vowel and the second member 
 begins with a vowel, the vowel of the first is elided: e. g. 
 somn-ambulus (somnus, stem somno--|- -ambulus), 
 
 'But less than the stem must not be used in any case; no change in a word 
 may be made except in the variable terminations. "A name made up of the 
 first half of one word and the last half of another is as deformed a monster in 
 nomenclature as a mermaid or a centaur would be in zoology.'' A worse 
 linguistic monstrosity than Cor-corax (from Corvus and Iluppo-xopaS), 
 both hybrid and mutilated at that, it would be difficult to conceive; and yet 
 it stands. Other examples of the same violation of law we have in Cyps- 
 nagra {xO<psXoi-\- Tavd-j-pa^, Merulaxis (Merula -|- -uva/^Aaft?), Bu- 
 corvus (^BouyJpws -\- Corvus), etc. 
 
Il8 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 magn-animus (magnus, stem magno- -|- animus).^ 
 
 7 But if the second member of the compound begins with a 
 consonant, the final vowel of the first member will appear 
 as i. For even in the case of a, o-"^ and u-^ stems the final 
 vowel is, in accordance with well known phonetic laws, 
 naturally weakened to -/- [m a n-c e p s, contractor, and m a n- 
 cipium, contract (derived from manus, hand -\- ciapere, 
 take) have lost even the -i-\ : e. g. 
 
 tub-i-cen(tuba, Aorn,-{- c a n-e r e, sing) , horn-flayer; 
 
 curv-i-rostris (curvus, stem curvo-, curved, -\- 
 rostrum, beak) , having a curved beak ; 
 
 spin-i-barbus (spina, s^z«e, + barba, beard), hav- 
 ing a beard of seines ; 
 
 hort-i-cola (hortus, stem horto-, garden,-\- coX- 
 ^m, dwell) , dwelling in a garden ; 
 
 navi-ger(navis, stem navi-,sAi!^,-l-ger-ere, bear), 
 ship-bearing ; 
 
 corn-i-ger (cornu,stem cornu-,^o;'«,+ger-ere, 
 bear), horn-bearing. 
 
 8 In the case of stems in -io-, the o simply falls out and 
 leaves the first member ending in -/-, as in the other cases : 
 e.g. 
 
 ^Forms like multi-angulus (side by side with mult-angulus) are of 
 later manufacture and the result of analogy with forms where no hiatus occurs. 
 
 ^Where -o- appears in the junction of compounds, it is due to the influ- 
 ence of the Greek compounds: e. g. albo-galerus, theflamen's hat; uno- 
 m a m m i u s, single-breasted. Accordingly, such formations asalbo-cauda- 
 tus, white-tailed, as participial form to albi-cauda (adjective), or rubro- 
 punctatus, red-spotted, or rubro-vinctus, might be perfectly legitimate 
 in epic poetry but not in Latin prose nor in strictly correct New Latin con- 
 struction. 
 
 ^A -u- seems sometimes to remain unchanged: e. g. nau-fragus (navis, 
 stem navi-,-t-\/frag, frangere, to break), ship -wrecking, for navi-fra- 
 gus, arose in the same way as Greek vao-ayA'i, and in the same way b u- 
 caeda (cf. bovi-caedium) and bu-mammus were made after the anal- 
 ogy of the Greek compounds of ySoD?; manu-mitto, manu-factus, etc. 
 are not due to composition, properly speaking, at all, but to juxtaposition, 
 manu- being not the stem but the ablative case. 
 
Zoot.— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. II9 
 
 medi-terraneus(medius, stem medio-, middle, -\- 
 
 terra, land), between lands; 
 heredi-peta (heredium, /eg'acjj-f pet-ere, hunt), 
 
 legacy-hunter ; 
 offici-perda (officium, ojice, perd-ere, lose), 
 
 labor-losing. 
 
 The same analogy has forced itself through in the case of 
 consonant stems as well, so that a connecting vowel -/- is 
 inserted between the consonant stem and the second mem- 
 ber:^ e. g. 
 
 part-i-ceps (pars, stem part-, ^ar^, + cap-ere, 
 take), sharing; 
 
 dent-i-f rangibulus (dens, stem dent-, tooth, -\- 
 frang-ere, break), tooth-breaking ; 
 
 soror-i-cida (soror, stem soror-, sister,-j- caed- 
 ere, kill), sister-slayer; 
 
 fratr-i-cida (frater, stem fratr-, brother,-{-caed- 
 ere, kill), brother-slayer. 
 
 The treatment of the first member is the same for both 
 nouns and adjectives. 
 
 The final member of a Latin substantivum compositum 
 may be a substantive (noun, adjective, participle) or a verb, 
 and the whole will appear, according to the will of the 
 coiner, as noun or adjective; its form will be decided by 
 
 'Rarely, instead of adding -i-, the consonant stem loses its consonant: e. g. 
 iu-dex, (ius, stem iur-, law, -\-dic-ere, speak); but it seemed unnatural 
 and so a iur-i-dic-us was made according to rule, with the same mean- 
 ing; in the same way lapi-cida (lapis, stem lapid-, stone, -^caed-ere, 
 cut) was also followed in due time by the normal lapid-i-cida, stone-cutter. 
 The forms with the connecting vowel -i- are by far the most numerous; 
 they are the only ones that we may take as our models, for, analogous forms 
 though they are, they represent the only method of formation in use in class- 
 ical times. The older method, of which only a few specimens are left, was 
 simply to prefix the stem of the first member to the second and leave the con- 
 sonant to take care of itself: e. g. 
 
 sol-stitium (sol), solstice ; 
 
 mus-cipula (mus), mouse-trap; 
 
 nomen-clator (nomen), name-caller. 
 
I20 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 the form of the final member. If the whole is to be a 
 NOUN, its form will be governed by the following rules : 
 
 11 I . If the final member of a Latin nomen compositum is a 
 noun, the compound noun will have the form and gender 
 and inflection of that noun: e. g. 
 
 pro-avus, -i, masc. {3.v\xs), great-grandfather ; 
 decem-vir, -i, masc. (vir), decemvir; 
 de-decus, -oris, neut. (decus), disgrace; 
 meri-dies, -ei, fem. (dies), midday; 
 bi-dens, -dentis, masc. {Aens), fork. 
 
 12 2. If the final member of a Latin nofnen compositum is a 
 verbal root, the compound noun will be («) a masculine 
 
 13 noun of the first declension ending in -a, (6) a noun of the 
 third declension: e. g. 
 
 (a) agri-col-a, -ae (ager, f eld, -^ co\-ere, till), 
 
 farmer ; 
 fratri-cid-a, -ae (f rater, brother, -\- caed-ere, 
 
 slay) , fratricide ; 
 lapi-cid-a, -ae (lapis, stone, -{-C2i&d- ere, cut), 
 
 stone-cutter ; 
 (d) tibi-cen, -cinis (tibia, _;f«/e, + can-ere, sing), 
 
 flute-player ; 
 man-ceps, -cipis (manus, hand, -\- cap ere 
 
 take), contractor ; 
 iu-dex, -dicis (ius, right, -\- die -ere, speak), 
 
 judge. 
 
 14 In forming compound adjectives, the final member, 
 as well as the first, is strongly influenced by the analogy of 
 the i- stems, so that we have compounds with stems in -o- 
 and -a- (nominative singular in -us and -er, masc, -a, fem., 
 -um, neut.), with stems in -i- (nominative singular in -is, 
 masc. and fem., -e, neut.), and with stems ending in a con- 
 sonant (nominative singular in -s, preceded by a consonant, 
 or in -r), according to the following rules : 
 
ZooL— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. 121 
 
 15 I. If the final member is a noun (a) of the first declension, 
 with stem ending in -a-, or (3) of the second declension, 
 with stem ending in -o-, or (c) of the fourth declension, 
 with stem ending in -u-, the compound adjective (A) may 
 
 16 have the endings -us, masc, -a, fem., -um, neut., or (B) 
 after the analogy of the -i- stems, it may appear with the 
 endings -is, masc. and fem., -e, neut.; as there is nothing 
 by which to decide, we often find both: e. g. 
 
 (A) (B) 
 
 (a) im-herh-ns{im-,not,-\-barha, beard), beardless, (a)ira-berb-is; 
 in-fam-us(in-, noi, +fama, fame), infamous, i n - f a m - i s ; 
 
 in-anim-us (in-, jzo^.+anima, life), lifeless, in-anim-is; 
 
 se-cur-us (se-, Tvithout,-\-c.\it3., care), careless, 
 (5) in-anim-us (in-, rao^-j-animus, life), lifeless, (b) in-anim-is; 
 
 raulti-iug-us (iugum of many yokes), manifold, multi-iug-is; 
 
 in-erm-us (pi. arma, arms), unarm,ed, in-erm-is; 
 
 multi-cav-us (cavus, cave), of many caves, and ex-somn-is; 
 
 (c) quadri-man-us (manus, stem manu-, hand), (c) quadri-man-is; 
 four-handed 
 multi-fruct-us(fructus, fruit), abounding in fruit, but bi-corn-is, 
 (bi-too,-l-cornu, horn), two-horned; so angui-corn-is, serpent- 
 horned; etc. 
 
 17 2. If the final member is a noun of the fifth declension, 
 with stem ending in -e- (nominative singular in -es), the 
 compound adjective ends in -us, -a, -um: e. g. 
 
 per-di-us (dies, day^, lasting all day; 
 levi-fid-us (fides, credit), of slight credit; 
 per-fid-us (fides, honor'), treacherous, but 
 ex-spes (spes, ho-pe), homeless. 
 
 18 3- If the final member is a noun of the third declension, 
 with stem ending in -i or in a consonant, the compound 
 adjective will have the form and inflection of its final mem- 
 ber: e. g. 
 
 (2) Aprils, 1897. 
 
122 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Sbr. 
 
 multi-fori-s (fori-s, door), many-do or ed^; 
 
 semi-cani-s (cani-s, dog), half-dog; 
 
 multi-cauli-s (cauli-s, stalk), many-stalked; 
 
 tri-dens (dens, dent-is, tooth), three-toothed; 
 
 bi-frons (frons, front-is, brow), two-browed; 
 
 in-quies (quies, quiet-is, rest), restless; 
 
 per-nox (nox, noct-is, night), lasting all night; 
 
 dis-cors (cors, cord-is, heart), disagreeing ; 
 
 multi-color (color, color-is, color), many- 
 colored; 
 
 bi-pe(d)s (pes, ped-is,ybo^j, biped; 
 
 se-grex (grex, gveg-\s,Jlock), separate; 
 
 multi-radix (radix, radic-is, root), many- 
 rooted. 
 
 19 Note. — Compounds of genus have either -genus, -a, 
 -um, or -gener-us, -a, -um, or -gener-is, -e: e. g. 
 multi-genus, -a, -um, or ,multi-generus, -a, -um, 
 or multi-generis, -e, of many kinds. 
 
 20 4- If the final member is an adjective, it will suffer 
 no change in entering into composition: e. g. per-mag- 
 nus, very great; multi-potens, very -powerful; semi- 
 nudus, half -naked; semper-viv us, ever-living. 
 
 21 5 • If the final member is a verbal root, it will simply add 
 the case-ending of the adjective — in the nominative either 
 
 22 («) -5 (subject to the phonetic laws), or (b) -us, -a, -um — 
 directly to the root: e. g. 
 
 (a) parti-cep-s (cap-ere, take), takingpart; 
 morti-fer (fer-re, bring), death-bringing; 
 sacri-fecs (x) (fac-ere, make), sacrifice-making; 
 
 (b) magni-fic-us (fac-ere, make), great-making; 
 perterri-crep-us (crep-ere, rattle), terribly rat- 
 tling; 
 
 male-vol-us (vol-o, wish), ill-wishing; so bene- 
 vol-us, bene-fic-us, etc. 
 
 ' The analogy of the o-declension has made itself felt here also and we have 
 multi-forus side by side with multi-foris. 
 
ZOOL.— Vol. 1.] MILLER— GREEK AND LATHSf DERIVATIVES. 123 
 
 The formulae for making the various classes of adjectives 
 are as follows : 
 
 a o -o 
 
 « > a 
 
 £ g I I 
 
 £+ I + S+ 5 
 
 23 I. im- berb(a)- i- s, beardless; 
 
 ve- san- u- s, mad; 
 
 de- men(t)- s, mad; 
 
 per- noc(t)- s, i. e. per-nox, lasting all night; 
 
 bi- pe(d)- s, i. e. bi-pes, biped; 
 
 se- greg- s, i. e. se-grex, separate. 
 
 a 
 
 s 
 
 Si a a g 5^ 
 
 5 + 2.+ !?+ £+0 
 
 24 2. magn(o)- anim- u- s, magnanimous; 
 
 miser(o)- i- cor(d) s, compassionate; 
 
 mult(o)- i- mod- u- s, manifold. 
 
 25 (a) The numerals, as first member of compounds, 
 deserve especial mention : 
 
 (i) unus follows the general rule and appears as 
 uni-: e. g. 
 
 uni-color, having one color ; 
 uni-formis, uniform. 
 
 (2) du-o appears as a prefix in the form of either 
 du-or bi-: e. g. 
 
 du-plex, double, but bi-pes, biped; 
 du-plus, double, bi-cornis, two-horned; 
 
 du-bius, doubtful, bi-lobus, two-lobed. 
 
 (3) tres always appears as prefix tri-': tri- 
 color, three-colored; tri-cornis, three-horned; tri- 
 fasciatus, three-striped, etc. 
 
 'Except in tre-decim, tre-centi (and their derivatives) and tressis. 
 
26 
 
 124 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 (4) quattuor always appears in the form of a pre- 
 fix, regularly quadri-cornis, yoKr-Aor«eJ; quadri- 
 punctatus, four-sfotted; quadri-dentatus, four- 
 toothed; but influenced by the form and sound of quattu- 
 or and quartus, that quadri- may also appear as 
 quadru-: e. g. quadru-manus, as well as quadri- 
 radinus, four-handed ; quadru-pes as well as quadri- 
 pes, guadru-ped; etc. Elision will, of course, take 
 place when the second member begins with a vowel: 
 e.g. quadr-angulus, quadrangle. 
 
 (5) From quinque on, the indeclinable numerals 
 are simply prefixes, suffering no change when they enter 
 into composition, subject, of course, to the usual phonetic 
 laws^: e. g. <\\iinc{\i.&-\ihvis ,fve-pound; sex-punc- 
 tatus, six-spotted; septem-flu-us, sevenfold-flow- 
 ing; octo-punctatus, etc. The declinable numerals 
 of the higher orders are treated like other adjectives, 
 while centum also may follow the same analogy 
 making centi-peda, hundred-footed; centi-manus, 
 hundred-handed; but also centum-peda; centu- 
 plex; etc. 
 
 
 1 
 g + 
 
 's 
 1 
 
 > 
 
 I 
 
 Verbal-root 
 
 (Stem-vowel 
 
 + 
 
 ! 
 
 
 
 3- 
 
 magn(o)- 
 
 i- 
 
 fic- u- 
 
 s,*, great-making; 
 
 
 fun(i)- 
 
 
 ambul- u- 
 
 s, rope-walking ; 
 
 
 sacr(o)- 
 
 i- 
 
 f ec- 
 
 s, sacrifice-making; 
 
 
 p art- 
 
 i- 
 
 cep- 
 
 s, sharing; 
 
 
 mart- 
 
 i- 
 
 fer. 
 
 death-dealing. 
 
 '5. quinque- is the rule; but tlie rule is not meant to include the com- 
 pound numerals like quin-decim, quinqua-ginta, quin-genti, etc., 
 nor strange forms like quincu-plex and quincu-pes, which are obviously 
 formed after the analogy of quadru-plex and quadru-pes. 
 
 6. ses-centi, se-mestris and their derivatives, conform to the rule, 
 while se -pes is made after the analogy of se-mestris. 
 
 7. Besides septem- we find sometimes septu- and septi-, which arose 
 from analogy with quadru- and quadri-, octu- and octi-; or possibly 
 
27 
 
 Zooi..— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. 125 
 
 £+ ^+ a+ o 
 
 in- iug- i- s, unyoked; 
 se- cur- u- s, careless. 
 
 28 S- Verbs are rarely used as first member of a compound ; 
 when they are so employed, their stem vowel appears as 
 -i-, either from a natural weakening or from analogy with 
 stems that do suffer such weakening: e. g. 
 
 horri-ficus (horre-re, dread, +fac-ere, make), 
 
 horror-stirring ; 
 posci-nummus (posc-ere, ask, -\- num-mus, 
 
 coin), begging; 
 terri-ficus (terre-re , /r«]g-^^e«, f ac-ere, make), 
 
 terror-causing; 
 perterri-crepus (perterre-re, terrify, crep- 
 
 ere, rattle), terribly-rattling. 
 
 29 6. Adverbs may be prefixed to almost any adjective 
 without affecting the form: semper-vivus, ever-living; 
 bene-volens, well-wishing ; male'dicens, evil-sfeak- 
 ing; etc. 
 
 Greek Compounds. 
 
 The construction of compound words is an art; the con- 
 summate mastery of that art is attainable or attained only in 
 the language of Greece. The Germans pile words together 
 in a convenient way and call them compounds, but they are 
 inartistic and artificial. Sanskrit is incomparably richer in 
 compositions than either Greek or German, or both together, 
 
 septu- was influenced by septu-aginta, and septi- is the natural weak- 
 ening of septu-. 
 
 8. octo- is the rule; but when octo became octo, it naturally followed 
 the course of the o- declension and became first octu- and then octi-. 
 
 9. novem- and noven- are both in use and both regular; noven is 
 the older form of the word: e. g. noven-dialis, of nine days; Novem- 
 ber, the ninth month. 
 
126 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 but those great mountains of words massed one upon the 
 other — a single word often translatable only by a long, in- 
 volved sentence — are no more to be compared artistically 
 with the Greek compounds than the hundred -handed 
 Vishnu is to be compared with the Hermes of Praxiteles. 
 
 30 In forming New Latin compounds of Greek elements 
 there are, just as in the case of Latin compounds, two 
 possibilities: (i) that the word formed will be a noun; 
 (2) that the word formed will be an adjective. Which it 
 will be depends entirely upon the will (and, perhaps, the 
 skill) of the coiner. The rules in either case might be 
 formulated thus : 
 
 L Nouns. 
 
 31 I. We may find the Greek compound already borrowed 
 entire or manufactured by the Romans and take it just 
 as it stands in the Latin lexica. [Most words of that kind 
 available for generic names in Zoology or Botany have 
 already been found and used.] 
 
 32 2. We may take the compound as it exists in Greek and 
 simply transliterate it into Latin letters.^ 
 
 33 
 
 'The accepted transliteration of the Greek letters into Latin is as follows: 
 
 a a 
 
 ( 
 
 i 
 
 P 
 
 r (p rh; pp rrh) 
 
 /3 b 
 
 K 
 
 c 
 
 cr 
 
 s 
 
 7 g 
 
 \ 
 
 1 
 
 T 
 
 t 
 
 S d 
 
 M- 
 
 m 
 
 V 
 
 y 
 
 e e 
 
 V 
 
 n 
 
 <f> 
 
 ph 
 
 ? z 
 
 ? 
 
 X 
 
 X 
 
 ch 
 
 ■n e 
 
 
 
 
 
 ^ 
 
 ps 
 
 e th 
 
 TT 
 
 P 
 
 CO 
 
 
 
 The rough breathing ' is represented by h, the smooth breathing ' is taken 
 for granted and not represented in Latin. Final -09 and -ov, of the second 
 declension, are represented by -us and -um. But final -oij is, of course, ren- 
 dered by Latin -os. This distinction is an obvious one, and yet the error of 
 transliterating -<u? by -us is one into which even the best may fall. For in- 
 stance. Dr. GUI, in his address before the American Association for the Ad- 
 vancement of Science ( ' ' Science, ' ' Vol. IV, p. 587), protests against the rule as 
 
ZoOL.— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. 127 
 
 34 3. We may take simple Greek words and manufacture 
 of them the compound as it would have existed in Greek 
 and then transliterate it into Latin. The manufacture will 
 take place in accordance with the remaining formulae. 
 
 35 4. If the final member of a nomen comfositum is a noun, 
 the compound will have the form and gender and inflec- 
 tional stem of that noun: e. g. 
 
 just given and says: " In accordance with such rules, Rhinoceros has been 
 turned into Rhinocerus. But Rhinoceros was admitted into classical 
 Latinity, and with it the corresponding oblique cases, Rhinocerotis, etc.; 
 in fact, the word was current in the language of description, satire, and 
 proverb — as when used by Juvenal for a vessel made of the horn, or by 
 Lucilius for a long-nosed man, or by Martial in the proverbial expression, 
 'Nasum rhinocerotis habere'; i. e., to turn the nose up, as we should 
 say. These authorities are good enough for me." Most naturally; for they 
 are true to the law! Dr. Gill is unfortunate in his choice of example. 
 Rhinoceros is correct, according to rule, and Rhinocerus is wrong; the 
 spelling -us does not occur in any classical author (except in a false reading 
 in Lucilius, where -os is now read). Of course, there can be no question 
 about the existence of forms in -os for Greek -of and -on for -ov in a cer- 
 tain class of writers of the Silver Age, when, according to that same Juvenal, 
 Rome was more Greek than Latin. Apuleius occasionally so transliterates; 
 but it is simply one of his many Greek mannerisms. Pliny the Elder is re- 
 sponsible for the most of the transgressions of that law which are to be found 
 in the dictionaries; his reason for such violation of the Latin law is found in 
 his desire to present the technical or scientific words of his Greek original in 
 an untranslated, unchanged form. The same phenomenon, with the same 
 reason underlying, is to be found in the Church Fathers. And so we have, 
 hammo-chrysos, hady-osmos, ophidion, gingidion, and many 
 more; most of them, as I said, are Pliny's Latin, but Pliny's solecisms or 
 mannerisms should not be good enough for any one when a better than Pliny 
 (in point of Latinity, at least) can be found; and the law both before and 
 after Pliny was: Greek .09 is Latin -us, and Greek -ov is Latin -um. 
 The diphthongs: 
 
 ai 
 
 se 
 
 av 
 
 au 
 
 VI 
 
 yi 
 
 V 
 
 e 
 
 ei 
 
 e or 1 
 
 ev 
 
 eu 
 
 a 
 
 a 
 
 m 
 
 
 
 01 
 
 oe 
 
 ov 
 
 u 
 
 
 
 
 
 ai and oi for at and 01 before another vowel appear only in a few proper 
 names, as Achaia, Troia, etc., and therefore do not concern us. The strict 
 Latin orthography cannot be too rigorously insisted upon; consistency will 
 in no other way be obtainable. 
 
128 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 anthropo-latra,-ae, masc. (-XdTpr}<;,-ov, masc, 
 
 stem Karpa-), man-worshi-per ; 
 cerco-pithecus,-i, fem. (iri0riKOi,-ov, fem.), a 
 
 monkey with a tail; 
 chamaer-ops,-opis, fem. {w'f-, (ott-oi;, fem.), tke 
 
 ■wall-germander ; 
 dia-gramma,-grammatis, neut. {^pa^LfLa, '^pd^ 
 
 ft,aTo<i, neut.), diagram; 
 dasy-pus ,-podis, masc. (ttoiJs, ttoSo'?, masc), a 
 
 kind of rabbit; 
 chamae-cyparissus,-i, fem. (^KVTrdpia-a-o^, stem 
 
 KVirapiara-o-, fem.), cypress; 
 aego-ceras,-ceratis, neut. (Kepav, stem xepar-, 
 
 neut. ) , goat's horn ; so also 
 rhync-ops,-opis, fem.; cyclo-stoma, -sto- 
 ma tis, neut.; hippo-potamus, etc., etc. 
 
 36 The gender and stem of the compound are the same as 
 
 those of its final element; the genitive singular of the com- 
 
 yj pound is formed in the same way upon that stem; and if 
 patronymics were to be formed from such words, they 
 would be made by adding -i d a e or -i n a e to the stem' : e. g. 
 anthropo-latr-idae, chamaer-op-idae, dasy-pod- 
 idae, aego-cerat-idae, rhynch-op-idae, cyclo- 
 stomat-idae, etc. 
 
 There are a few apparent exceptions to this in our class- 
 ical lexica, but they are only apparent: e. g. opistho- 
 domus, although derived — borrowed directly — from 
 oiria-ddSoiJ.o';, masc, is given (correctly?) as feminine. 
 A6fio<;, when transcribed into Latin letters, is identical in 
 form with Latin domus, fem. (which is not derived from 
 
 'If the stem ends inavowel, it will, of course, suffer elision before the suffix 
 -idae. The only other patronymic suffix in Greek is -iwv, but it is of rare 
 occurrence except in proper names: e. g. azpoud-iiov, ostrich; nopyup-iwv, 
 the purple coot; and a few others. 
 
 No regard needs to be paid to the Latin patronymics. The Romans had 
 no patronymic formation of their own, and when they came to borrow from 
 the Greek, their lack of consistency in following their models introduced a 
 chaos, out of which no rules can be deduced for our guidance. 
 
ZOOL.-VOI.. I.] MILLER— GREEK AND LATIN DERIVATIVES. 129 
 
 the Greek but a genuine Latin word), and the gender 
 would naturally, in the mouths of the people, be changed 
 by the analogy of the Latin word. 
 
 The same danger threatened all the Greek derivatives 
 that became thoroughly Latinized and entered into the lan- 
 guage of the people ; many succumbed, and yet not so many 
 as one would naturally expect. Twenty- five or thirty will 
 exhaust the list: e. g. the catapult is a Roman rather than a 
 Greek implement of war, and so we are not surprised to 
 find cata-pulta, -ae, fem., although it owes its origin to 
 KaTa--7re\Tr]^,-ov, masc. Dia-dema, dogma, schema, 
 etc., were looked upon as neuter nouns and inflected 
 
 dia-dema, -demat-is, -demat-i, 
 dogma, dogmat-is, dogmat-i, 
 schema, schemat-is, schemat-i, etc., 
 or were considered feminine and inflected 
 diadema, diademae, 
 dogma, dogmae, 
 schema, schema e, etc., 
 
 according as they were felt as genuine Latin words or as 
 Greek derivatives. Cicero took pains to show that he 
 thought dogma a Greek word by writing it with Greek 
 letters, So'7/io. So also we find oxy-gala, gen. oxy- 
 galact-is, neut., and also oxy-gala, -ae, fem., from 
 Greek 6^v-ya\a,-yd\aKT0';, neut.; cata-rrhactes, -ae, 
 masc, and cata-rrhacta, -ae, fem., from Greek xara- 
 ppdKTr)<;,-ov, masc. ; etc. In almost every instance of change 
 of gender and inflection, we find, side by side with the Latin 
 change, Latin forms preserving the gender and inflection 
 of the Greek. 
 
 38 (fl) The gender of the genus name, when it is made 
 
 a noun, depends, therefore, not on the termination, but 
 upon the gender of the noun forming the final element of 
 the compound. 
 
 This rule is the only one that can have any philological 
 basis. It is deduced from the practice of the classical 
 
130 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 30 Ser. 
 
 word-makers and from the practice of scientists from Lin- 
 naeus down: e. g. (the instances cited on p. 129); from 
 Linnaeus, Systema NaturcB, V^: Didelphis virginiana, 
 Rhynchops nigra, Ophidion barbatum, etc. The 
 variations on the part of naturalists from this rule are due 
 either to arbitrariness or misunderstanding: e. g. Gnathy- 
 pops rhomaleus, Hypsipops rubicundus, and oth- 
 ers of the sort were made in the supposition that tB-i/r was a 
 masculine noun; Hypognathus serenus, in the idea 
 that '■/vddo'i was masculine; so Hybopsis hyostomus, 
 and many more. 
 
 39 Note. — The theory has been advanced and emphasized 
 
 that all these compound names are adjectives, with a noun 
 understood.' That theory, if rigorously insisted upon, will 
 not leave a single compound noun in the whole Greek lan- 
 guage. It will hold good in the case of some but not of all, 
 by any means. Alvo-Xecov, jJUKpo-TroXir-qti, rpioyXo-SvTrii, 
 Kvvoa-ovpa, Trpo-Kvoov, XeovTO-ireraXov, epir-dicavQa, Atoa-icovpoi, 
 (7Ta<j)vX6-SevSpov, x<^fiaL-TnTV<;, ■jrp6-vao<;, ^i(^o-id,d')(aipa, ttitvo- 
 Kafiirrj, ;]^7;j'-a\tB7r77|, and any number more are just as 
 good nouns as were ever produced, and the same is 
 true of the names of genera, as both scientists and philolo- 
 
 ^This theory seems to originate with one Herrmannsen, who is quoted as 
 saying: Vocabula composita Graecse originis, quorum terminalis 
 radix generis neutralis, in a exiens, genitivo -aTO(; flectitur, 
 nonnisi adjectivo sensu adhibentur. Hinc si neutralia numero 
 plurali obveniunt, ut Adeloderma, Cricostoma [sing, -um, vel 
 -on], terrainentur geniiivo-orum; sin autem singular! numero, ut 
 Amphidesma, Cyclostoma, feminini erunt generis, et decli- 
 nabimus "Cyclostomse." Documento sint veterum nomina 
 darpaxddepfia [which is simply neuter plural of the adjective dtrrpaxodepfio^, 
 see §§44 and 55], Tplffrofio?, adeaiio's. — Quoted by Dr. Gill, National Acad- 
 emy of Sciences, Vol. VI, Sixth Memoir, p. 129. 
 
 Herrmannsen' s rule is all very well, with the exception of the two words 
 nonnisi adjectivo; the falsity of this assumption is abundantly proved by 
 a host of examples, such as dia-dema, genitive diadematis; dia- 
 phragma, genitive, diaphragmatis; epigramma, genitive, epi- 
 grammatis; emblema, genitive, emblematis, etc., etc. See § J 35-44. 
 The possibility of an adjective formation of the sort assumed by Herrmannsen 
 is recognized in § 44 and more fully explained in $ 55. 
 
ZOOL-VOL. I.] MILLER— GREEK AND LATIN DERIVATIVES. I31 
 
 gists look at them. If we could assume that they were ad- 
 jectives, it would puzzle one to discover why one piscis is 
 feminine, another masculine and a third neuter. It is the 
 same with aves, reptilia, plantae, etc.; and the word 
 chosen for the assumed adjectives to agree with would be 
 arbitrarily chosen and would have little chance of corre- 
 sponding with the idea in the mind of the coiner when he 
 named his genus. 
 
 40 To say that epi-gramma, melo-drama, acro-polis, 
 macro -cosmus, leonto-cephale, chaen- alopex, 
 Helles-pontus, pyx-acantha, hippo-dromus, and 
 hundreds of others are not nouns, but adjectives used sub- 
 stantively with some imaginable word understood, is no less 
 absurd than to deny that cyclo-stoma, di-delphys, etc., 
 may be nouns of the same sort and subject to the same laws. 
 And yet that theory which would make adjectives of all 
 these words tells us that even mega-therium is not a 
 noun but a neuter adjective modifying animal or mon- 
 strum understood, and that it is subject to the inflection 
 megatherius, megatheria, megatherium, etc.! It 
 is, therefore, properly derived, we are informed, not from 
 lj,e'^a-\-6r)piov directly, but from a fieya-Bi^pio'; sc. O'^p, ' a 
 great-beastly animal.^ That is, to be sure, a perfectly legit- 
 imate course to take in order to get megatherium, but it 
 is no more likely that it was obtained in that way than it is 
 that the discoverer of cyclo-stoma did not simply put 
 together KX)KKo<i -\-aT6p,a, but first built an adjective kvkXo- 
 a-TOfio'ij-ov, then turned that into Latin cyclo-stomus, 
 - u m , and then arbitrarily manufactured a feminine cyclo- 
 stoma to agree with piscis, masculine, understood! The 
 classical precedent for forms in -stomus, masc. and fem., 
 -stomum, neut., is poly-stomus, masc. and fem., 
 poly-stomum, neut. The only rational way to form either 
 would be ii,ir^a-\-6r)pCQv and kh/cXo — \-aT6iJ.a. And as cyclo- 
 stoma goes, so must go the rest of the words in -stoma, 
 
 41 -soma, etc. In general, it is arbitrary and inexact to make 
 adjectives of three endings in Latin out of Greek adjectives 
 of two endings. The old Latinists used the form in -us for 
 
132 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 both masculine and feminine, and -um for the neuter; e. g. 
 curo-troph-us (-os), masc. and fem., curo-troph-um 
 (-on), neut.; dory-phorus, masc. and fem., dory-pho- 
 rum, neut.; disco-phorus, masc. and fem.; eu-pe ta- 
 lus, masc. and fem.; poly-stomus, masc. and fem., 
 -um, neut. The only exception I find given in the Latin 
 dictionary is di-somus, -a, -um; but that is erroneous 
 and should read disomus,- um. 
 
 42 5. Family names formed from neuter names of genera 
 in -stoma, -soma, etc., should end in -atidae not in 
 -idae; for the stem of those words ends in -at, Greek -aT. 
 It is so in Latin (see p. 129) : e. g. dia-phragma , gen. 
 dia-phragmat-is , epi-gramma, gen. epi-gram- 
 matis, em-blema, gen. em-blemat-is , etc. 
 
 43 Note. — The family name from forms in -stomatus, -a, 
 -um, would appear in the same form as that from -stoma, 
 -stomat-is. 
 
 44 6. Family names formed from adjectives (used sub- 
 stantively) in -stomus or -s to mum, -somus or -sonum, 
 etc., end in -idae, not in -atidae. Thus Catostomus, 
 Catostomidae . 
 
 IL Adjectives. 
 
 45 As a matter of course, this division will comprise by far 
 the largest part of all Greek compounds, but not by any 
 means all of them. Many of the compounds are genuine 
 nouns (see §§ 39—41), and all the adjectives may be used 
 as nouns. 
 
 46 In forming compound adjectives from substantives (nouns, 
 adjectives and participles) , it is not the familiar, nominative 
 singular form of the noun, adjective or participle that is 
 used, but the stem.^ If the first element is a substantive 
 with a stem ending in a vowel and the second begins with a 
 vowel, the stem-vowel of the first element is, of course, 
 
 ^See % 5, with foot-note i. 
 
ZOOL— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. 133 
 
 47 elided'; and, at the same time, if such elision brings a sim- 
 ple surd («, TT, t) up against a vowel affected by a rough 
 breathing, that simple surd will become the correspond- 
 ing aspirate (%, ^, ^). If, on the other hand, the second 
 element begins with a consonant, the vowel stands, and the 
 rule is that all a-stems follow the analogy of the o-stems and 
 
 48 change their vowel to -o-. When, however, the stem of the 
 first member ends in -t or -v'', the same analogy has forced 
 itself in and -0- is added to the stem of the first as a con- 
 necting-vowel between it and the second member of the 
 compound^: e. g. epvBp {0^-6^6 aKji,o<i {ipvOpoi;, red,-\-6<j)0a\- 
 IJ.6<;,eye), red-eyed; rnj.epo-hpojjLO'; {riiJ.ipa,-\-hp6iio<;, running), 
 one who runs all day; 6vp(a)-avXo^ (^dvpa, door,-\-avKea), 
 lodge), living out of doors; but Bvp-o-Koiro'; {dvpa, door,-\- 
 KOTTTO), knock), knocking at the door; Se'x^ij/j.epo'i (heKa, ten, 
 -^■fffiepa, day), ten-day; vv)(0-riiJi,ep-o'i (w|, stem vvkt-, night, 
 -{■•^Hepa, day), night and day; fieXav-o-Ke(}>aXoi (yiieXay, black, 
 stem iJi.eKav-,-\-Ke(^a,\ri, head), black-headed; XeovT-o-Ke^aXo<; 
 (Xecov, stem XeovT-, lion, -{-KecfjaXij, head), lion-headed; ifiav- 
 To'-TTOu? (t/iia?,stem if^avT-, strap, -\-'irov'i, foot), leather-footed; 
 6^t-o-Ke<^aXo^ (o<f>i-<;, serpent, -\-Ke<^aXrj, head), snake-headed; 
 TToXi-ovxai, i. e. ■7roXi-d-exo<: (ttoXj-?, city, stem 7ro\t-,-j-e';)^(», 
 hold) , city-holding; lx6v-o-irSX7]'s (lxBv-^,fsh,-\-7ra>X6co, sell) , 
 fsh-monger. But when the final v of the stem is part of a 
 
 'Sometimes contraction takes place; e. g. xaxoup^oi {xax6i-\-*ep^w), 
 working ill; emouxo(;'(sdvTJ,hed,-\-s^w^ tend), eunuch. This would be the nat- 
 ural phonetic course for such combinations to follow, but it is the exception, 
 rather than the rule, for them to do so. 
 
 4g " This rule applies to nouns only; all adjectives with stems in -u, when figur- 
 
 ing as first member of a compound, appear as bare stem without any con- 
 necting-vowel: e. g. ppaxo-Tcows, short-footed; (ipa-^u-oupo's, short-tailed; 
 ^padu-voui, slow of thought; ^SO-oafio^, sweet-smelling; (vxu-ttou^, swift- 
 footed; and all the hosts of words beginning with ttoXu-. 
 
 ,Q ^ a, e, I, and j; occur so rarely in the middle of compound words, even 
 
 before a consonant (and then, too, for the most part, in poetry), that for our 
 purposes we may disregard them entirely. Still, good classical precedents 
 maybefoundfortheuseof anyoneof them; l^du^, e. g., appears in the Greek 
 lexicon as the first member of compounds thirty-one times with an -0- over 
 against fourteen times without the -0-, while the compounds of oyc^ all have 
 the —0— vowel. 
 
134 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 diphthong, the connecting-vowel is never used: e. g. vav- 
 wyo'; (pav-<}, ship,-\- Vay-, dy-vv-fii, break), shi-p-wrecking; 
 ^ov-Ke<f)aKo^ (/Sou-?, ox,-\-Ke(pa\'^, head), ox-head; ^ov- 
 Kpdviov, ox-skull. 
 
 51 The second word also is so powerfully influenced by the 
 same analogy, that a large majority of all compound adjec- 
 tives take the inflection of the -o- declension, with two end- 
 ings: -09, masc. and fem., -ov, neuter. These usually go 
 into Latin as -us, masc. and fem., and -um, neut. ; rarely 
 the feminine has a separate form in -a, and the neuter a form 
 in -on. Some have -5j?,-e?, and some take the original form 
 of the final member. The rules for the three respective 
 cases might be drawn as follows: 
 
 52 I. When the final member of the compound is (a) a 
 substantive of the first declension with stem ending in -a 
 (nominative in -??? or -as masc, -a or -r) fem.), or {b) of 
 the second declension with stem ending in -0 (nominative 
 in -09, masc. or fem., -ov, neut.), or (c) if the final member 
 is made directly from a verbal root, the compound adjective 
 ends in -o?, masc. and fem., -ov, neut. 
 
 («) inrep-^ope- <s (ySo/oe-a?, nortK), beyond the north; 
 ^pax^-ovp- o If (^ovpd, tail), short-tailed; 
 olSi-Kvr)ft-o<; (KvqfiTj, shin), swollen-shinned; 
 vvx6-'niJ'ep-o<; (vvkt-, stem of wf, night, -{-rmepa, 
 
 day), night and day ; 
 Kvvo-/ce<f)aX- 9 (kvv-, stem of kvwv, dog,-\-Ke<f>a\ri, 
 
 head), dog-headed; 
 
 (b) Tpi-SdKTvX-o'i {ZdKTvKo-^ , finger ,) three-fingered ; 
 \evK-6(f>daXiJL-o<; (6(f)ffa\/jLd-<; , eye), while-eyed; 
 'Xpva-o-Trrep- <! (irrepo-v, wing), gold-winged; 
 lj,aKp6-KevTp-oi; (xevrpo-v, sting), long-stinged ; 
 
 (c) vir-rjKO-o'i , (ctKofO), hear), obedient; 
 6r]p-o-Tp6<f>-o<; , (Tpecfxo, breed), animal-breeding; 
 XiOo-cfydy- o <s {y^ay, (f>ay€iv, eat) , stone-eating; 
 avOo-Xoy-o'i (v/x%7, Xe'7eB, gather) , flower-gather- 
 ing. 
 
53 
 
 ZOOI..-VOL. I.] MILLER— GREEK AND LATIN DERIVATIVES. I35 
 
 2. When the final member of the compound is a sub- 
 stantive of the third declension with the stem ending in any- 
 consonant except -V, -p, -8, or -e? (and occasionally also 
 even then), the compound adjective (i) usually follows the 
 same analogy and ends in -0?, masc. and fem., and -ov, 
 54 neut., or (2) retains the third declensional form of the final 
 member : 
 
 (i) ev-crapK-o<; (^crdp^, stem aapit-, jlesh^, incarnate; 
 
 a-')(p6nt,aT- 9 (^'x^pa/j.a, stem ^^jom/tar-, color) , colorless; 
 
 TToXv-'^eip- o ^ or iroXv-yeLp (j^etp, hand), many- 
 handed; 
 
 ■Kaix-jjirjv- ? (f-ijv, month, stem /f t?!"-) , lasting through 
 all the month ; 
 
 (2) fiLKp-wyjr (coijr, stem wir-, eye), sm.all-eyed ; 
 
 epi-^Qs\a^ (^S)\a^, stem ^coXaK-, clod), with big 
 
 clods ; 
 lj,aKp6-')(ei,p (xeCp, stem X^'-P't hand), long-handed ; 
 Bv(7-£pa}<! (epo)?, stem ipeor-, love), cherishing an un- 
 happy love; 
 o\ecrC-67]p {Orip, stem 6rip-, beast), beast-destroying. 
 
 (fl) When the final member is a neuter noun ending in 
 -IJia (stem ending in -/xar-) , the compound adjective properly 
 ends in -p-aT- o-?, as in the rule just given, or (as those words 
 not infrequently in composition shorten the stem from -p-aro- 
 to -p,o-^) in - p,o<s , or we may have both the longer and 
 shorter forms side by side; e. g. 
 
 Tpi-croip.aT-o-'i and Tpi-a-cop.-o-<;, three-bodied ; 
 
 a-<7Q)p,aT-o-<; and d-crQ)p,-o->;, bodiless; 
 
 d-Bepjjiar-o-^ and d-Sepp,-o-?, skinless; 
 
 a-'x^pap.aT-o'i and a-^pm/x-o-?, colorless. 
 
 Linngeus has forms to which the prototypes would be avai- 
 a-Top,-o-<;, •iro\v-vr}p,-o-'i, ')(^pva-6-a-TL'^p,-o-'s, 'x^pvcrd-crTO/j.-o-'; , 
 alp,6-cnop,-o-'i, yovo-a-Tiy/JL-o-'i, apiyvpo-aTO/jL-o-i. 
 
 ' Perhaps following the analogy of words like ypd/j.ij.a,-aTo?, with its parallel 
 ypa^fj.TJ,-rji, from which we have ypa/ifio-ecdTJ?, etc.; cf aT:spno-^oXi<u, 
 aTzspp.o-kuyo'i, Ssp/io-izzepo';, auip-aaxiui, awp.-epa(TT-ijs, etc. 
 
 55 
 
136 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 The shortened forms may occur also when these words 
 are used as first members of compounds: e. g. 
 
 tj-Trep/iiaT-o-Xoy-o-i; and enrepfi-o-Xoy-o-^ , seed-collecting; 
 
 airep/iar-o-cfxiy-o-'! and <T'7repfi-o-(j>dy-o-<i, seed-eating ; 
 
 acoid.d.T-'qy-o-'; but o'wpj-epaaTri'i ; 
 
 alfiaT-o-TTOv; (L.) but alfi-o-irrep-o-^ (L.); 
 
 aijjLar-aihrj'i ■ (-L.) but aip.-6-<TTop,-o-<i (L.). 
 
 56 {b) When dealing with this class of neuter nouns, there- 
 fore, we may be entirely justified in using whichever form 
 may happen to suit our fancy, and say dermato-, sper- 
 mato-, haemato-, somato-, etc., or dermo-, sper- 
 mo-, haemo-, etc., as we will. Numerically, the Greek 
 lexicon reveals a preference for the longer form as first 
 member of a compound, and for the shorter form in the 
 final member. 
 
 (c) For the inflection of these adjectives and the forma- 
 tion of family names derived from them and from the cor- 
 responding nouns, see pp. 127-132. 
 
 57 3- When the final member is a substantive of the third 
 declension, with a stem ending in -v, -p, or -S (which is 
 always dropped by law before the -? of the nominative sin- 
 gular), the compound adjective usually retains the form 
 and inflection of the last member: e. g. 
 
 KaKO-Saifiaiv, ^possessed of a had s-pirit; 
 
 avTo-'xjdmv, sprung from the earth itself; 
 
 <jii\-eWrjv, Greek-loving; 
 
 iroXv-'^eip, many-handed; 
 
 d-7rai<; (a- privative -f 7rat'(S)?, child), childless; 
 
 pi-^-aaiTi'i (piTT-ra, throw, -{-aaTrk, stem a<77rtS-, shield), 
 throwing away one's shield; 
 
 a-TTovi (a- privative +tou9, for -l-*7ro(8)?, foot, foot- 
 less ; 
 
 l/j,avT-d-7rov<; (Ijid';, stem If^avr-, leather strap, 4-*7ro(S)9, 
 foot), leather-footed; so also Xaym-Ttov;, hare-footed; 
 ^paSv-irov;, slow-footed; ipvdpd-irov;, red-footed; 
 yXapo-irovi, green-footed; etc. 
 
ZOOL— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. I37 
 
 58 (a) All words in -pus (Greek -.ttov?) have the stem in 
 -pod; the genitive ends as in Latin in - p o d - i s , the 
 plural in -pod-es. The family name will, therefore, 
 end in -pod-idae. The Latin lexicon reveals forms 
 like 
 
 echino-pus, gen. -pod-is; chytro-pus, gen. -pod-is; 
 haemato-pus, " -pod-is; clino-pus, " -pod-is; 
 himanto-pus, " -pod-is; O edi-pus, " -pod-is; 
 lago-pus, " -pod-is; tri-pus, " -pod-is. 
 
 There is no exception; corono-pus has genitive 
 -podis, but accusative -pum. Poly-pus, -i, has also 
 Greek ttoXutto?, ttoXvttou, as archetype. 
 
 59 (3) In the same way the words in -n and -r (Greek -v 
 and -p) have the genitive in -n-is and -r-is, the plural in 
 - n - e s and - r - e s , and the family names in -n-idae and 
 -r-idae. 
 
 60 4. When the final member is a substantive of the third 
 declension, with a stem ending in -es (nominative in -779, 
 masc. or fem., -09, neut.), the compound adjective will 
 end in -■»??, masc. and fem., and -es , neut. : e. g. 
 
 Si-eTTji (eVos, stem irea--, year), biennial; so rpi-erTj^, 
 
 triennial; 
 \va-L-iJ,eKiq<i (/j.eXo'i, stem fj,e\ecr-, limb), limb-relaxing ; 
 vr]-Keph'ffi (KepSo'i, stem Kephea--, gain), bootless. 
 
 Linnaeus seems to use as adjectives the noun forms oxy- 
 rincus {sic) (pp. 512, 395), gono-ryn-chus (p. 528), 
 callo-rynchus (p. 402), syn-gnathus (p. 416). 
 
 ^i {a) When the last member of such a compound is elSo?, 
 
 preceded by an -o-, the -o--\-eihe<; may contract to -oSe?: 
 e. g. 
 
 rpix-o- e I S ^ s or T/3(%- d)Sr)<; , hairlike; 
 Kr]T-o- e t S 77 ? or KTjT- CO S 7j 9 , cetacean ; 
 kvkX-o- eihri<i or kvkX- Shrft , cycloid ; 
 ryoyyvX-o- e ( S ^ 9 or yoyyvX- to 8 •?; 9 , roundish. 
 
 ( 3 ) AprU 8, 1897. 
 
138 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Sbr. 
 
 62 5. When the final member is a neuter noun of the third 
 declension with stem ending in -ar- (nominative singular 
 in -as), the compound adjective will end in -to? (con- 
 tracted from -a{,T)o^') for the masc. and fem., and -o>v for 
 the neut. ; but («) sometimes it follows the analogy of the 
 adjectives in -09, -ov. e. g. 
 
 'fKvKv-Kpe(o<s ,-aiv or yXvicv-Kpe-o'; , -ov, sweet-meated (^Kpeaf ; 
 
 stem, Kpear-) ; 
 fiovo-Kepoii; or /jLovo-Kepo'; , one-horned {icepa'i ; stem,«:e/oaT-) ; 
 a-77;/3ffl?(7^pa?, age'), ageless; but 
 6p06-Kep-o<!, uprt'gAi-korned ; pivo-Kepw; , horn-nosed ; etc. 
 
 63 6. When the final member is a substantive of the third 
 declension with a stem ending in the vowel -t- or -v- (nomi- 
 native singular -t?,-v?, masc. or fem., and -i or -v, neut.), 
 the ending of the compound will appear identical with the 
 form of the final member: e. g. 
 
 (TWO- 1- TT o X t - ? ( TTo'Xt-s , ctty) , city-saving; 
 7re/3o-e'-7ro\t-9 (ttoX*-?, city), city-destroying; 
 6<f)i-d-\jri-^ (^o-\jri,-i; , look), looking like a snake; 
 \t,'ir6-vav-<; (vav-<;, ship), ship-abandoning; 
 Si-trrjxv-'! ('T'Pjxv^, cubit), two cubits long; 
 *a-Se\(j)v-'; (SeX(j)v<s, womb), wombless; etc. 
 
 64 7. When the final member is a verbal root, it may have 
 either an active or a passive force; {a) if active, the com- 
 pound may have the simple verbal root with the endings 
 
 65 -o<; ,-ov; {b) if passive, it will usually have the suffix 
 
 66 -roii; (c) it may have the ending -i?s, which is either 
 active or passive: e. g. 
 
 (a) 67jp-o-Tpd(f)-o<; (VTp^tf), rpefjuo, breed), animal- 
 breeding; 
 vX(a)-o-TdiJ,-oq (Vt%ij,, rejjL-v-m, cut) , wood-cutting; 
 ovpav-o-aKOTT-o^ (^cncoTr-eaj, see), sky'rgazing; 
 
ZOOL— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. 139 
 
 (3) KaTd-<j)pa/c- T o <; (V<f>pay, (f>pdcrcra), cover), covered 
 
 over ; 
 Sopv-aka}-TO<i (^ydXco-, dXCuKOiMai,, be taken), taken 
 
 with the spear ; 
 a-XeTTtSo)- T o 9 (*X67rtSo'(i), cover with scales), not 
 
 covered with scales ; so 
 /j.aKpo-X.eTTiSco- roi , d-yvo)- t s , ' Apr]i-(j)a- to<; , etc. 
 
 (c) ev-fia6-ij<} , easily learning, or easily learned; 
 a-iiaO- ■^ <; , not learning, or not learned; 
 deo-cre^- r\ 9 , reverencing the gods. 
 
 The principal classes of Greek compound adjectives, 
 the formation of which is concerned in this discussion, are 
 not different from the classes of Latin compounds dis- 
 cussed in §§ 23-29 and will be treated in the same way. 
 They are as follows : 
 
 67 
 
 (+5 ) 
 
 + o 
 
 XSOVT- 
 
 0- 
 
 K€<f>aX(^a)o- 
 
 ?, lion-headed; 
 
 "irir^o)- 
 
 
 ovp(a)o- 
 
 9, horse-tailed ; 
 
 •wajj,- 
 
 
 IJbTjv- 0- 
 
 9, all the month; 
 
 KVV- 
 
 6- 
 
 'yX(»TT(a)o- 
 
 9, dog-tongued; 
 
 KVV- 
 
 0- 
 
 Ke(j)aX(^a)o- 
 
 9, dog-headed; 
 
 poBo- 
 
 
 SaKTvXo- 
 
 9, rosy-fingered. 
 
 6^ The process would be exactly the same for two adjec- 
 
 tives combined into one; or for an adjective with an insepa- 
 rable prefix: 
 
 epv0pd-XevKo-<;, reddish white; 
 XeuK0-/xe'\a(i')-9, black and white; 
 TroXv-j(X(upo-^ , very green ; 
 av-airio-'i , not to blame; d-a-o^o-<;, unwise; etc. 
 
140 
 
 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 69 
 
 a/j,<f)(- 
 
 r^jj,!,- 
 
 Noun-stem 
 
 + 
 
 > 
 
 a 
 
 V 
 
 2). 
 
 1 
 
 
 /3io- 
 
 
 ?, 
 
 amfhibious; 
 
 awfiaT- 
 
 0- 
 
 ?, 
 
 bodiless ; 
 
 XevKO- 
 
 
 ?, 
 
 whitish ; 
 
 Seo-yiio- 
 
 
 ?, 
 
 unbound; 
 
 •X^pm/jiaT 
 
 - 0- 
 
 ?, 
 
 colorless ; 
 
 TTTepo- 
 
 
 ?, 
 
 half-finned. 
 
 70 
 
 71 
 
 < + 1 (+s) + 
 
 3 
 
 
 ^idKp{o-')ovp(^a)o- 
 
 ?, 
 
 long-tailed ; 
 
 rpi- (Tco/j-ar- 0- 
 
 ?, 
 
 three-bodied; 
 
 TrepKVO- irrepo- 
 
 ?. 
 
 dusky-winged ; 
 
 \evK(o-) o^daXfio- 
 
 ?, 
 
 white-eyed. 
 
 (a) Numeral adjectives follow an abnormal course : 
 
 (i) Instead of the numeral el?, /Ji,(a, ev, in compo- 
 sition, im6voi;= single, is used: e. g. /jl v o-yafio'; , married 
 but once; /mov6-^v^, single-yoked; ^i.ovo-Ke.pai'i, having 
 but one horn. 
 
 (2) For %vo- there is an inseparable prefix St- which 
 is always used in compounds: e. g. St-Sa/eTwXo?, two- 
 fingered; Si-oSeov, two-toothed; 8 i -irrepo'; , two-winged; 
 Si-errj's, biennial. 
 
 (3^ The stem of rpei?, -rpia, is tpi-, which is simply 
 prefixed in entering into composition: e. g. t pi-am\i,o<i, 
 trifle-bodied; t p i -hdKTv\o<i , three-fingered; t p i -'^mva , 
 three-cornered. 
 
 (4) rerrape'; has also an inseparable prefix form 
 rerpa-: e. g. t e t p a -SdKTv\o<i , four-fingered; rerpd- 
 ycovoi;, four-cornered; tet pa-err]';, quadriennial. 
 
ZooL— Vol. I.] MILLER— GREEK AND LATIN DERIVATIVES. 14I 
 
 (5) and (6) wevre and ef follow the analogy of 
 Terpa- and kirTa, and make prefixes irevra- and k^a- 
 respectively : •n-evra-haKTvKo'i, five-fingered; e^a- 
 Sa.KTvXo'}, six-fingered. 
 
 (7) eTTTa, oKTO), etc., are simply prefixed without 
 change. 
 
 72 Note. — TLav in composition is usually treated as a neuter 
 
 substantive, with stem irav- instead of 7rai'T-,in an adverbial 
 relation to the other member of the compound: e. g. nrav- 
 a-o<j>oi;, all-wise, ir av-(TeXi)VO<i, full-mooned; ir cut. -yi,r)vo<i , 
 through all the month; Tlav-eXKrjve'i, the Grecian world; 
 rarely TravT-o- : ir avT-o-ha-iro';, all sorts of ; ir avr-o-Tropoi, 
 all-traversed. 
 
 74 
 
 +^ +5+5 
 
 73 4. ai/d,- 6- ppo- o- <;, flowing with blood ; 
 
 TOTTo- ypdtj)- o- 9, -place-describing ; 
 
 \i0o- (l>dy- o- 9, stone-eating; 
 
 ovpavo- tTKoir- 0- 9, sky-gazing; 
 
 Ppayxi-o- CTey- 0- 9, gill-covered. 
 
 
 s 
 
 S V 
 
 £+ 
 
 
 (t5)+5 
 
 d- 
 
 rpOTT- 
 
 0- 9, not turning; 
 
 dvd- 
 
 ySXeTT- 
 
 9, i. e. /3\e'\]r, looking up; 
 
 V 
 
 ev- 
 
 TpO<f)- 
 
 0- 9, living with; 
 
 d- 
 
 TOfl- 
 
 0- 9, uncut; 
 
 Sl'X^d- 
 
 TO/J,- 
 
 0- 9, CM^ 2« ^ZfO. 
 
 Exceptions to these rules will be found, of course, not 
 only in scientific nomenclature, but even in well approved 
 classics. All that has been attempted here has been to give 
 
142 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 
 
 the rule, leaving the abnormal to take care of itself. As I 
 have given the rule and ignored the exceptions for the most 
 part, even where exceptions are numerous, the laws laid 
 down may be called " Draconian," and censured as being 
 "tinctured with more than Roman severity." But the 
 " frequent laments over the instability of our systematic 
 nomenclature" and the "bitter complaints against those 
 who change names " are called forth in nearly every case 
 by some one's having rebelled against some " Draconian " 
 law. Consistency and correctness can be secured only by 
 following out what is recognized as rule, and will surely be 
 thwarted by adopting strange forms even when sanctioned 
 by sporadic use on the pages of a Pliny. Nothing is good 
 enough but the best, whether in science or language or 
 Latin form for scientific nomenclature ; and Pliny, for the 
 last, is not good enough so long as we have something bet- 
 ter. I cannot refrain from concluding with the rest of Dr. 
 Gill's paragraph, alluded to above, in which, after objecting 
 so vigorously to what he calls "Draconian" laws and 
 "laws tinctured with more than Roman severity" and 
 urging that " the language of nomenclature should not be 
 bound by rules of strict philology," he declares so unequi- 
 vocally for humble obedience to the higher law of priority 
 as the only way out of utter confusion: " Frequent are the 
 laments over the instability of our systematic nomenclature; 
 bitter the complaints against those who change names. But 
 surely such complaints are unjust when urged against 
 those who range themselves under laws. We are forci- 
 bly reminded by such complaints of the ancient apologue 
 of the wolf and the lamb. The stream of nomencla- 
 ture has indeed been muddied, but it is due to the acts of 
 those who refuse to be bound by laws or reason. The only 
 way to purify the stream is to clear out all the disturbing 
 elements. In doing so, mud that has settled for a time may 
 be disturbed, but that is at worst anticipating what would 
 have inevitably happened sooner or later. We are suffer- 
 ing from the ignorance or misdeeds of the past. In opposing 
 the necessary rectifications and the enforcement of the 
 
ZOOL— Vol. I.] MILLER— GREEK AMD LATIN DERIVATIVES. 143 
 
 laws, extremes may meet ; conservatives and anarchists 
 agree. But the majority may be depended upon in time to 
 subscribe to the laws, and the perturbed condition will then 
 cease to be." 
 
 We may recognize the law of priority as absolute, and 
 retain the many monstrous and misspelled names to be 
 found on the records of natural history, just as their makers 
 left them. They are historic facts and serve to mark the 
 group of animals or plants to which they apply, but these 
 misshapen forms of words are not ornamental and they are 
 unworthy of scholars. It is to be hoped that, in future, 
 greater care maybe taken to make words that give correctly 
 the idea the author may have intended. Such words as 
 Felichthys, Lefomis, Semotilus demand the constant apology 
 of those who use them, while words like Zalofhus, Eri- 
 cymba, Hylocichla are a pleasure in themselves to those who 
 understand their meaning. It costs no more to frame a 
 name properly than to leave it a monstrosity. 
 
 If this paper shall serve as a stepping-stone toward the 
 attainment of correctness and uniformity in the framing of 
 names of classical origin, its purpose will be fully met. 
 
University Publications. 
 
 {Continued from Second Page of Covers) 
 
 9. Marine Fossils from the Coal Measures of Arkansas. 
 
 By James Pkrrin Smith, pp. 72. ^plates. 
 
 10. Scientific Names of Greek and Latin Derivation. 
 
 By Walter Miller, pp. 31. 
 
 UNIVERSITY ADDRESSES. 
 
 The American University and the American Man. 
 
 Second Commencement Address. By George Elliott Howard, A. M. 
 pp. 22. 
 
 ' THE LIBRARY. 
 
 Hopkins Railway Library Catalogue, By Frederick J. 
 f Teggart, A. B. pp.231. 
 ! Address, 
 
 ; THE REGISTRAR, 
 
 Stanford University, California.