METHODS OF PALAEONTOLOGICAL INQUIRY. 75 



proportion to prevent such a combination. The way of effecting 

 it is as follows : 



Suppose that Specimen I of a given species consists of a 

 skull, vertebral column, and hind limbs, the problem being to 

 supply the missing fore limbs; let us assume further that Speci- 

 men II of the same species has the skull, femur, and fore limb. 

 The femur is thus common to both, and we have the proportion 

 (calling the humerus of the second specimen//", the femur F", 

 and the femur of the first specimen F') as follows : 



F" ://" : :F' \x. F" x = H" F' .: 



_ H" F' 

 * z ^fr- 



If the comparison is made from a few specimens, individual 

 variations in size and proportions are apt to give a more or less 

 grotesque result, but this may be corrected by employing a 

 large number of individuals and making the calculations as 

 often as possible and by the aid of as many different bones as 

 possible, and averaging the result. Such a method is tedious 

 and requires great supplies of material, but it has the advantage 

 of being trustworthy. 



In making the figure of a restored or composite skeleton I 

 find the following mode of procedure useful : The most com- 

 plete individual is laid out in a natural position in a box of sand, 

 and photographed; the photograph is then drawn in outline on 

 a large sheet of bristol board by the aid of the camera lucida, 

 the missing parts are calculated from other specimens and 

 drawn in their proper places, the whole adjusted, and, if neces- 

 sary, redrawn. It often happens that more or fewer bones are 

 missing from all the specimens, especially the more fragile and 

 loosely connected bones, such as the scapula, ribs, sternum, 

 and caudal vertebrae. These, if not too numerous, are supplied 

 conjecturally, and this fact is indicated by leaving the missing 

 bones unshaded in the drawing. With all due care, however, 

 and with seemingly abundant supplies of material, restorations 

 sometimes go ludicrously astray, and probably none is ever 

 made entirely free from faults. Still, the making of them 

 serves a useful purpose, for I find that even after studying the 



