INTRODUCTION. 19 



typical of the adults of more primitive allies, but characters which are usually eliminated in 

 development. Arrested variants are shown abundantly in the consideration of ocular plates, 

 when fewer plates may become insert than is characteristic of the species (Tripneustes, text- 

 figs. 123-125, p. 124). 



2. Progressive variation, in which the variant has characters not typical of the species, 

 but which are further evolved on the direct line of differential development, and are seen typi- 

 cally in more evolved nearly allied species or genera. Such progressive variants are also shown 

 abundantly in the consideration of ocular plates, where more plates become insert than is 

 typical in the species (Centrechinus, text-figs. 93-95, p. 107 and 176, p. 153). 



3. Regressive variation, in which the variant takes on characters of the adult of some simple 

 and more primitive type of the group. Such characters are not necessarily a repetition of youth- 

 ful characters but may go back to a remote ancestry. An arrested variant in a sense is one form 

 of regressive variation, but a regressive variant includes much more than arrested variation. 

 To distinguish them, an arrested variant is one that has developed to a certain point as usual, 

 and then failed to take on the later added characters typical of the species, so that, although 

 an adult, it has immature characters. A regressive variant is one that has attained full char- 

 acters and then in later life has reverted to youthful or primitive characters as an individual 

 variation, or it is a variant that from youth has primitive characters not normally seen in the 

 development of the species. A modern horse with extra digits as in Tertiary times could be 

 considered a regressive variant, but could not be considered an arrested variant. In Echini, 

 cases of regressive variation are shown in the simple ambulacrum in Melonechinus (Plate 57, 

 fig. 3), and in the single column of interambulacral plates in Arbacia (Plate 4, fig. 11) and Trip- 

 neustes (Plate 6, fig. 4). 



4. Parallel variation is where a character is taken on exceptionally which may be com- 

 pared with characters normally occurring in some type of the group not closely connected, so 

 that it cannot be genetically compared. A case apparently is the extra genital pores fre- 

 quently developed in Recent Echini (Echinus, text-fig. 115, p. 117). 



5. Aberrant variation is where a character is taken on which is quite abnormal, not to 

 be correlated with the typical condition in associated forms. Cases of aberrant variation 

 are sea-urchins which have four or six areas developed (Plate 6, figs. 1-4). All the evidence 

 goes to show that aberrant variation is rare, and most variants can be considered as arrested, 

 progressive, regressive, or parallel variants, and as such can be correlated with species more 

 or less nearly allied which typically possess the character which is a variant of the case in hand. 

 This holds true as far as known in both animals and plants, as seen in the present paper, and 

 as I showed in detailed studies of plants (1899). 



Next to stages in development, variation is the most suggestive line of study in attempt- 

 ing to work out the genetic relations of plants and animals. When a specimen of one species 



