438 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1918. 



tablished law of the necessary regressive evolution of nonfunctional 

 parts and organs. The reversibility of the ascending evolution of a 

 complex organ, when it depends on a reduced part, is therefore not 

 impossible (we can suppose, for instance, that the secondary ventrals 

 of teleosts will return in the future to their original condition 

 through the complete disappearance of the clavicular ligaments). 



In the case of a nonrecluced part, whose form has, however, 

 changed during ascending evolution, the indestructibility of the past 

 .again does not exist in the strict sense, the nonreduced part 

 being able to change its form again by a new progressive evolution, 

 although the original condition of this part, and consequently the 

 original condition of the organ in question can not be reestab- 

 lished. The pelvis of Triceratops may be taken as an example. The 

 postpubis of this pelvis exists in a very rudimentary condition, and 

 as rudimentary parts tend to disappear, the postpubis certainly 

 would have disappeared if Triceratops had lived long enough. It 

 is therefore only its recurved ischium, very different in form from 

 the ischium of its distant tetrapod ancestors, which was able to pre- 

 vent Triceratops from recovering its original pelvis. 



Finally, if there is an ascending evolution of nonreduced parts 

 (pelvis of Stegosaurus) it is the change of function which saves these 

 parts from a regressive evolution ; the indestructibility of the past 

 does not exist here either. And it is clear that the same reasoning is 

 also applicable under the third law to the evolution of a complex 

 organ. 



To sum up : The irreversibility of the evolution of a complex or- 

 gan depends entirely on the irreversibility of the evolution of the 

 reduced or nonreduced individual parts which enter into its com- 

 position, and the second and third laws are not without exceptions 

 in this respect any more than the first; as we have seen, it is the 

 germinal base of the first law r w r hich underlies the entire subject. 



As I said at the beginning, most naturalists know D olio's first law 

 only. This is only a part of his general law, although the most im- 

 portant and most certain part. 1 This general law, in spite of the pos- 

 sible exceptions, has an extraordinary importance for biological 

 philosophy and evolutionary philosophy in general. Dollo will al- 



1 Besides the law of irreversible evolution Dollo has formulated (see Dollo, 4, p. 165) 

 two other laws that of discontinuous evolution (before II. de Vries) and that of limited 

 evolution. In his subsequent writings Dollo has only rarely touched on these two other 

 laws (on the law of limited evolution see Dollo, 7, p. 9, Dollo. S, p. 813 and p. 820, 

 Dollo, 9, p. 131 ; on the law of discontinuous evolution see Dollo, .">. rem. (66), p. 120; 

 Dollo. 7, rem. (11), p. 9; and Dollo, 17, pp. 139-140). 



