6 /, H. Powers 



N'ARIATIONS IN GENERAL BODY FORM 



These are very considerable in the species ; very slender as well 

 as very robust individuals are frequent, while now and then occur 

 forms of surprisingly extreme proportions. This variation is, 

 moreover, quite independent of the plumpness or fatness of the 

 animal. The most massive specimen may be gaunt and lean, 

 while the slenderest specimens occur in the highest condition. In 

 judging the robustness of body form it is necessary completely to 

 ignore transverse diameters in the midbody region; they are 

 wholly worthless. In the feeding season, after breeding has 

 taken place, many males of the species linger in richly stocked 

 feeding grounds for weeks and often eat themselves out of all 

 semblance to normal proportions. Not only the digestive tract, 

 but the body walls distend more and more, becoming loose and 

 flabby unless the animal is gorged with food. Artificial feeding 

 may frequently produce still more distention. Such animals have 

 at times a curiously unsalamandrine and even pathological ap- 

 pearance. They are perfectly healthy, however. But this dis- 

 tention of body has nothing to do with permanent body form such 

 as I consider as variation in this paper. These highly fed animals 

 never maintain their condition for more than a few weeks, the 

 feeding impulse always tending to rythm or irregularity, espe- 

 cially when the food supply is abundant. However constant this 

 latter may be, the animals suddenly, without loss of health,- lose 

 their appetites entirely. Following this the body walls suddenly 

 undergo a marked contraction, even while the animal is still in 

 the water. The soft-bodied, greedy meat-eater, of walrus-like 

 proportions, that reveled in aquatic life, becomes in a day or two 

 slender-bodied, wiry, restless, and absolutely bent upon shifting 

 its abode from the best of aquatic feeding grounds to even the 

 most undesirable of terrestrial conditions. 



It is necessary, then, wholly to discard transverse body meas- 

 urements in the middle region. There remain the possibilities of 

 transverse measurements in the pelvic or the pectoral regions or 

 of the head. None of these is satisfactory, the former two be- 

 cause they vary most with the. animal's condition ; the latter be- 



