Miss Agnes Crane — Evolution of the BracMopoda. 105 



Let us now summarize the principal results of these combined 

 studies from -'the life" and from "still life," for the palseobiologist 

 begins his inquiries at the point where those of the embryologist 

 usually terminate. 



We find that complex or synthetic types are by no means rare. 

 The farther back we search the geological records the greater is the 

 number of types uniting characters which subsequently become 

 differentiated and accentuated, and serve to distinguish the diverse 

 members of widely separated families. 



The occurrence of s:enera with such mixed characters that 

 descriptive paleeontologists of the older school found it difficult to 

 determine to which of two families they should be referred ^ — as 

 Davidson sometimes frankly lamented, — ci)upled with the frequently 

 expressed difficulties of separating many so-called species, should 

 have originated doubts as to successive specific or generic creations. 



Reversionary types are also met with which exhibit a tendency 

 to go back to some earlier adult phase of genetic characterization. 

 Such living "atavistic" types are Gwynin, Cistella among Megathy- 

 rids, and DijscoUa among Terebratuloids, Atretia (17) of the 

 Ehynchonelloids, Bhinobokis among the fossil Trimerelloids, and 

 Anlosteges among the Productoids. 



The study of the individual development of larval forms of 

 different species reveals past phases in the history of the evolution 

 of the genus to which they belong, and its ancestral relationslups. 

 It becomes evident that these successive transitional stages were 

 often stereotyped as adult genera fossilized in the records of the 

 rocks. If this can be demonstrated of one branch, one order of 

 Brachiopoda, we maintain that it follows logically that the principle 

 is true of all bi'anches, although the methods and lines of variation 

 may have deviated. In other words, the same goal is reached by 

 travelling different ways, as can be actually proved in more than 

 one instance. 



Now, it is impossible to deny that such a successful demonstration 

 has been accomplished through the combined results of various 

 researches on the structure and development of the recent Terebra- 

 tuloids by Friele (;^6), Deslongchamps (34), Morse (57), Dall (24), 

 the CEhlerts (59, 61), and Beecher (4), and for the Palaeozoic 

 Brachiopoda by Hall and Clarke (38). From Clarke and Beecher's 

 (14) studies of a large series of individuals of several Silurian 

 species we may also learn how easily abnormal forms originate at 

 different periods of individual growth, and that these direct and 

 indirect divergents, due to physiological causes accelei'ated by the 

 environment, may give rise to new generic or subgeneric forms. 

 The descendants of these generic variants, again prematurely 



* This difficulty becomes naturally most apparent when treating of the Primordial 

 and later Pala30zoic genera, and is less marked among later developments. Hall 

 and Clarke, as is well known, consider the employment of family groups, and even 

 broader divisions, as "so palpably a violation of nature's method as to make itself 

 felt as an incumbrance," and therefore use them merely as conventional and 

 convenient. 



