20 E. R. HOSKINS AND M. M. HOSKINS 



frogs. If two glands of exactly the same size but of different 

 shapes are measured by this method, the volumes determined 

 might differ greatly, while two others of different shape and 

 size might seem to have the same volume. This irregularity 

 in the form of the hypophysis, for instance, is shown in our own 

 figures and in those of Rogers on page 605. Rogers states that 

 the same relative error would apply to all the glands so measured, 

 but he used so few specimens that no two of the glands were 

 of quite the same shape, and hence different errors have been 

 made in his measurements of different specimens. In far greater 

 error are Rogers and Allen when they attempt to establish the 

 relative volume of an organ by dividing the product of its three 

 principal diameters by the length of the animal's body, for the 

 volume of a body varies as the cube of a linear dimension. 

 Moreover, in larvae of different sizes the proportions of the body 

 vary considerably and the difference in the size (volume or 

 weight) of two animals is not directly proportional to the differ- 

 ence in their body length. This is especially true where one 

 attempts to compare tadpoles and frogs. It may be noted in 

 figures 12 and 16 that the length of the real body of the larva 

 (excluding the anal canal) is nearly the same as that of the re- 

 sulting frog and yet the larvae are from two to three times as 

 large as the frogs produced from them (table 2). It may be 

 noted also that the larger thyroidless larvae have relatively 

 long tails, and hence their size is not directly comparable with 

 that of control larvae on the basis of length of body. Still 

 further, larvae contain relatively much more water in their 

 tissues than do frogs. 



Another source of error in work of this sort is that the animals 

 vary in the amount they shrink during fixation. Given two 

 tadpoles fixed in the same fixative, one may shrink nearly twice 

 as much as the other, and hence with a small number of speci- 

 mens large errors may be introduced in attempting to estimate 

 volumes of structures. Where different fixatives are used, still 

 greater differences in shrinkage result. These animals are so 

 small that unless they are fixed, many of the organs cannot be 

 properly dissected or easily handled, and will undergo post- 



