30 ASTACID7E. 



mediate, and if not, finally, Nature never agrees with the strict prin- 

 ciples of a particular scheme, so that apparently capricious aberrations 

 are to be found everywhere the stumbling-blocks of the naturalist who 

 wishes to arrange everything in a regular series. 



A principal character suggesting the division of Cambarus into groups 

 is to be found in the hooked legs of the males. This character divides 

 all the species into two great groups, one with hooks on the third and 

 the fourth pair of legs, the other with hooks only on the third pair of 

 legs. This characteristic seems preferable to the later mentioned one, 

 because it unites forms which are related in all other respects, while the 

 groups divided according to the form of the rostrum unite species which 

 are otherwise quite unlike. 



In number and situation the hooks are always identical and very sure. 

 Among nearly a thousand males I have found only one abnormal male 

 with no hooks at all. I have never observed any aberration in the 

 group with hooks on the third and on the fourth pair of legs, except a 

 few second-form males of C. pellueidus, with the hooks on the fourth Legs 

 very small, even in one case not at all developed. In the other group, 

 with hooks on the third set of legs, sometimes, but very rarely, males 

 are to be found with hooks more or less developed on the second pair 

 of legs, but never on the fourth pair. I may add, that the second-form 

 males always have less developed hooks, and that all show the hooks 

 except the very young and newly hatched males. 



It is worthy of remark, and seems to prove the importance of this 

 character, that the hooks are situated on the same joint and at the 

 same place as the embryonal appendages of the legs in the young lob- 

 sters (Hbmarus), described by Professor Rathke and others. These 

 afterward disappear, and are not to be found at all in the young of 

 Astacus flumatiMs. Indeed, these hooks do not exist in the young Cam- 

 bari; their development is later; but the analogy is too striking to be 

 overlooked, and suggests the great desirableness of an accurate ac- 

 quaintance with the embryological development of Cambarus. 



According to the number of hooked legs, Cambarus is also divided 

 into two groups : — 



I. Third and fourth legs hooked, — group of C. acutus. 



II. Third legs hooked, — all the rest. 



The second important character for the division of Cambarus into 

 groups is the shape of the rostrum. Mr. Girard has employed this 

 character as a principal one, and forms three groups, with the following 

 characters : — 



Rostrum subquadrangularly elongated, tridentated at the tip, — C. 

 affirm and allied species. 



Rostrum short, broad, conical, toothless, — C. Bartoni and allied species. 



Rostrum very much elongated, conical, with a small and acute spine 



