NATURAL SCIENCES OF PHILADELPHIA. 17 



character. In the first place, it is impossible to always measure 

 the tail from the same point. There is no visible point of separa- 

 tion of tail from body ; the latter narrows regularly and gradually 

 into the former, so that we cannot say where one ends and the 

 other begins. Measurement from the end of the fur is usually 

 supposed to be meant; and about where the true fur ends the 

 scaly plates begin. But it will be quite safe to allow for a varia- 

 tion of at least two inches in this regard. I have certainly seen 

 so much ditference as this, as well as I could judge. This refers 

 to external measurements of the animal in the flesh; the dried, 

 especially if skinned, tail can afford no data whatever. Finally, 

 appeal to the vertebro? themselves is not infallible, for these may 

 vary. The first coccygeal vertebra (as I hold it to be) resembles 

 the others in physical characters, but is really a part of the sacrum, 

 being anchylosed therewith. At the extremity of the tail, one 

 vertebra or more may be lost without evident trace. The termi- 

 nal dozen or more bones have no neural canal, nor any notice- 

 able parts beyond the impervious centra by which they articulate ; 

 they resemble the internodes of a digit. The tail is singularly 

 liable to ulceration upon injury, when one or more bones may 

 exfoliate from the end, and the subsequent cicatrix be scarcely 

 recognizable.^ I met with one such case. It would not be safe 

 to base a specific difference upon less than three or four inches 

 in length of tail ; and this should be taking age for age of the 

 animal, of course, and be only declared of specimens measured 

 in the flesh. In color, the opossum's tail is usually in largest 

 part whitish or flesh-colored, blotched with a dark livid or 

 blackish hue toward the base. The degree of blotching and the 

 shapes of the dark spaces vary with almost every individual. 

 Sometimes most of the tail is dark colored ; sometimes there is 

 almost no pigiiient in the scales. Absolutely nothing can be pre- 

 dicated of this feature. 



The ears of this creature are very thin and membranous, and 



' It is also highly probable that the coccygeal vertebriB normally varj^ 

 one or two in number, that is to say, we may find 21, 22, 23 bones, without 

 any having been lost by accident or disease. This, however, is an opinion 

 of mine, not an observation. Some small long-tailed animals vary more 

 than this ; e.g., Mr. Allen says (Bull. Mus. Comp. Zool. No. 8, p. 228) that 

 Hesperomys leucopus may have from 24-5 to above 30 caudal vertebroe. 

 Other Muridce appear to vary quite as much. 

 1871.] 



