422 ANNUAL EEPORT SMITHSONIAN INSTITUTION, 192 8 



for their skulls are situated behind the center of gravity, so that the 

 head has to be constantly held in position by powerful muscles; 

 finally, the faciaj part of their skulls is relatively large and their 

 brain cases are relatively small. 



Properly to qualify as a " missing link," a " find " would have to 

 show that it was part of a creature which had some of the essential 

 characteristics of both humans and apes. It would not be necessary 

 that all parts should be intermediate to exactly the same degree, for, 

 as Sir Arthur Keith very properly points out, evolution has its dis- 

 cordancies, and in the course of racial development one organ com- 

 monly alters its structure at a different rate from another. But 

 there are two requisites which could not be dispensed with — first 

 that more than one part of the animal should be discovered, and 

 second, that these parts should have unquestionably belonged to- 

 gether. Why this is necessary may need explaining. Fossil skele- 

 tons are sometimes found with the bones practically undisturbed, 

 or, if somewhat scattered, all clearly of one kind with no admixture 

 of any other sorts. On the other hand, they may be found jumbled 

 together in what were once quicksands, pits, or crevices where ani- 

 mals of several kinds have perished, or in what were once stream 

 beds where loose bones of many sorts have been irregularly spread 

 or collected by the action of the water. When the bones of a fossil 

 skeleton retain their original position in the body of the animal the 

 evidence is clear; when they have become widely separated, and par- 

 ticularly when they have been mixed with remains of other creatures, 

 the evidence is perplexing. The reason why more than one part of 

 the supposed link must be known is this — namely, that one part of 

 an animal not infrequently resembles the corresponding part of an- 

 other to which it is not nearly related. This likeness without rela- 

 tionship is usually called " parallelism " or " convergence," and it 

 is so common throughout nature that everyone who is occupied with 

 the classification of animals must be constantly ready to take it into 

 account. The similarities in general form between a porpoise and 

 a fish or between a snake and an eel are well-known examples of 

 parallelism. In their fundamental structure the animals of each 

 pair are profoundly different; but in their external form they are 

 remarkably alike. On the other hand, there are often found striking 

 resemblances between particular parts of two animals which are 

 unlike in general form. For instance, the crowns of the molar teeth 

 in one of the extinct North American peccaries deceptively resemble 

 the crowns of human molars, as has recently been brought out during 



