October 23, 1896.] 



SCIENCE. 



597 



ing ways it lias been used in the systems of 

 Bleeker and Haeckel. It is evident, how- 

 ever, that more groups than the old conven- 

 tional ones, which alone Agassiz admitted, 

 would be useful at present. A happy mean 

 seems to be realized in the following list : 



Branch Superfamily 



Subbranch Family 



Superclass Subfamily 



Class Supergenus 



Subclass Genus 



Superorder Subgenus 



Order Species 



Suborder Subspecies 



There are only two (or three for trino- 

 mialists) of these which are ' sonant,' all the 

 others being '■ mute ' (to use the expression 

 of Linnaeus) ; but a question of termination 

 affects several of them. 



All the supergeneric groups, like families, 

 were originally chiefly designated by de- 

 scriptive names, but the trend in all the 

 years has been towards names which are 

 based on the stems of existing genera. 



FAMILY. 



In 1796-7 (' an 5 de la E.'), Latreille, in 

 his ' Precis des Caracteres generiques des 

 Insectes,' for the first time employed the 

 term ' familj?^ ' as a subdivision of an order, 

 but only gave the families numbers (' Fam- 

 ille premiere,' ' Fam. 2,' etc.).* He re- 

 marked that it might be desirable to have 

 the families named, but deferred doing so 

 till he could review the subject with greater 

 care.f 



In 1798 {' an 6 '), Cuvier, in his ' Tableau 

 Elementaire de I'Histoire naturelle des 

 Animaux,' in the introduction, when treat- 

 ing of graded characters (' caracteres 

 gradues'), named only the genus, order, 



* "Les rapports anatomique, ceux de V Habitus, des 

 metamorphoses, ont 6te mes guides dans la formation 

 des families. Elles sont prec6d6es d'un chiffre arabe. ' ' 

 p. ix. 



t "On eut desire que j'eusse donne des noms aux 

 families ; mais prevoyant que je serois contraint d'y 

 f air e plusieur s changemens, j ' eusse ainsi expos6 la nom- 

 enclature a une -vicissitude tres contraire a I'avance- 

 ment de la science. ' ' p. ix. 



class, and the kingdom. In the body of the 

 work, sometimes he used the word family 

 instead of order (as for the Birds), but for 

 two orders of the Insects he formally adopted 

 a division into families which were regu- 

 larly named. The first (unnamed) order 

 ('ordre'), with jaws and without wings 

 (' Des insectes pourvus de machoires, et 

 sans ailes'), was divided into several fam- 

 ilies (' plusieurs families naturelles ') — ' les 

 Crustaces,' ' les Millepieds,' ' les Arac- 

 neides,' and 4es Phtyreides.' The order 

 Nevropteres was disintegrated into three 

 families (' trois families naturelles ') — ' les 

 Libelles,' 'les Perles,' and 'les Agnathes.' 

 The representatives of the other (six) or- 

 ders were distributed directly into genera. 



This, so far as I have been able to dis- 

 cover, was the first time in which an order 

 of the animal kingdom was regularly di- 

 vided into named families, designated as 

 such . 



In 1806 Latreille, in his ' Genera Crus- 

 taceorum et Insectorum,' gave names to 

 families, but on no uniform plan, provid- 

 ing descriptive names for some, as ' Oxy- 

 rliinci ' for the Maioidean crabs — names 

 based on typical genera, with a patronymic 

 termination, as Palinurini and Astacini, and, 

 in other cases, names also based on a typi- 

 cal genus but with a quasi plural form, as 

 Pagurii. (In the same work, it may be 

 well to add, Latreille also admitted more 

 categories than usual, using ten for the ani- 

 mal kingdom — Sectio, Classis, Legio, Cen- 

 turia, Cohors, Ordo, Familia, Tribus, Genus 

 and Species.) 



In 1806 A. M. Constant Dumeril, who 

 had previously contributed tables of classi- 

 fication to Cuvier's 'Lecons d'Anatomie 

 Comparee,' and published his own ' Elemens 

 d'Histoire Naturelle,' brought out his 

 ' Zoologie Analytique.' In this volume he 

 gave analytical tables for the entire animal 

 kingdom and admitted families for all the 

 classes. The families, were generally sub- 



