The nets used by the Atka and Attore Aleuts “are from sixteen to eighteen feet long, and six to ten feet wide, with coarse meshes made nowadays of twine, but formerly of sinew. On the kelp-beds these nets are spread out, and the natives withdraw and watch. The Otters come to sleep or rest on these places, and get entangled in the meshes of the nets, seeming to make little or no effort to escape, paralysed, as it were, by fear, and fall in this way easily into the hands of the trappers, who have caught as many as six at one time in one of these small nets, and frequently get three.... No injury whatever is done to these frail nets by the Sea Otters, strong animals as they are; only stray Sea Lions destroy them.... The salt water and kelp seem to act as a disinfectant to the net, so that the smell of it does not repel or alarm the shy animal.”[195] GENERAL RELATIONS OF THE LAND CARNIVORA, RECENT AND FOSSIL. From very obvious reasons we have been compelled to describe the various forms of Land Carnivora of which we have been able to take account, one by one, beginning with Cats, and ending with the Otters. But the reader will already have discovered that a linear arrangement like this gives no true conception of the relations existing between the various families of which the sub-order is composed, or of the various genera which are included in the families. For cross-relationships of the most puzzling and often complicated description are perpetually turning up: among the Æluroids, for instance, we found Cryptoprocta to be intermediate between Cats and Civets, and yet, if we had followed the order indicated by this relationship, we should have had to ignore the close connection between Cats and Hyænas, and that between Hyænas and Civets, through the intermediation of the Aard Wolf. It is necessary, then, to devise some method of writing down the names of the families, other than that of placing them one under the other, if we are to get anything like a clear notion of their mutual relationships. The method adopted by Professor Flower is perhaps the most convenient, and following him, we arrange the groups thus:-- FELIDÆ. HYÆNIDÆ. URSIDÆ. CRYPTOPROCTIDÆ. PROTELIDÆ. CANIDÆ. PROCYONIDÆ. AILURIDÆ. VIVERRIDÆ. MUSTELIDÆ.