74 NEW YORK STATE MUSEUM 



fused together and greatly modified in the process. The genal 

 spines persist in the adult of the Mesonacidae and often the inter- 

 genal spines, but only in a modified manner. The intergenal spines 

 are seen in a later geological period in the adult Bronteus, 

 where they might be considered as a reversion to a character of their 

 Cambrian ancestors. Hydrocephalus appears to have an 

 intergenal spine and in all of the Proparia (Beecher, 1897, p. 198) 

 the " genal spine " is attached to the space within the facial sutures, 

 and is in fact the prolongation of one of the fused segments 

 of the cephalon, and corresponds in this respect to the intergenal 

 spine of the Mesonacidae. Some of the species of the genus Agnos- 

 tus also show spines that suggest the intergenal spine, notably A . 

 granulatus Barrande and A . rex Barrande." 



Assuming that, as the absence of connecting forms between the 

 Cambrian and these Ordovician and Devonian trilobites suggests, 

 the archaic characters of the later forms are not continuous with 

 those of the Cambrian ancestors, then we would have either a rever- 

 sion to an ancestral character of organs, never suppressed; or a 

 reiterative development of organs that were apparently lost. 



Biologists have of late years thrown considerable doubt on the 

 occurrence of cases of "reversion," "atavism" and reiterative 

 development. The law of " irreversibility of evolution " enunciated 

 by Dollo in 1893 ^^^^ been accepted by most biologists, although 

 severely criticized by several; and Agnes Arber has lately drawn 

 attention to a part of the broader principle recognized by Dollo, 

 which she defines as the " law of loss." This indicates the " general 

 rule that a structure or organ once lost in the course of phylogeny 

 can never be regained; if the organism subsequently has occasion to 

 replace it, it can not be reproduced, but must be constructed afresh in 

 some different mode." Lull (1917, p. 572) would restrict the law of 

 irreversibility of evolution to this impossibility of regaining a lost 

 anatomical structure, defined as the " law of loss " by Arber. 



An important qualification of this law is that it applies only to 

 actual losses and not to apparent losses due to the interpolation of 

 inhibiting factors (Arber, p. 27, footnote). 



In analyzing the criticisms to which Dollo's law has been sub- 

 jected on the part of botanists, notably by Errera who con- 

 sidered the law of irreversibility as disproved by the facts of 

 " reversion " ("that is by cases in which a variation appears which 

 is interpreted as an atavistic throw-back to a hypothetical ancestor, 

 and in which some character since lost by the species makes a 

 renewed appearance"): Mrs Arber found that the cases cited 



