PRESIDENTIAL ADDRESS. 37 



if it occurs at corresponding developmental periods in all cases; and 

 still more if it occurs equally in forms that hatch early as free larvae, 

 and in forms with large eggs, which develop directly into the adult. 

 As examples of such characters may be cited the mode of formation 

 and relations of the notochord, and of the gill clefts of vertebrates, 

 which satisfy all the conditions mentioned. 



Characters that are transitory in certain groups, but retained 

 throughout life in allied groups, may, with tolerable certainty, be 

 regarded as ancestral for the former ; for instance, the symmetrical 

 position of the eyes in young flat fish, the spiral shell of the young 

 limpet ; the superficial positions of the madreporite in Elasipodous 

 Holothurians, or the suckerless condition of the ambulacral feet in 

 many Echinoderms. 



A more important consideration is that if the developmental changes 

 are to be interpreted as a correct record of ancestral history, then the 

 several stages must be possible ones, the history must be one that 

 could actually have occurred, i.e., the several steps of the history as 

 reconstructed must form a series, all the stages of which are practicable 

 ones. 



Natural selection explains the actual structure of a complex organ 

 as having been acquired by the preservation of a series of stages, 

 each a distinct, if slight, advance on the stage immediately preceding 

 it, an advance so distinct as to confer on its possessor an appreciable 

 advantage in the struggle for existence. It is not enough that the 

 ultimate stage should be more advantageous than the initial or earlier 

 condition, but each intermediate stage must also be a distinct advance. 

 If then the development of an organ is strictly recapitulatory, it 

 should present to us a series of stages, each of which is not merely 

 functional, but a distinct advance on the stage mmediately preceding 

 it. Intermediate stages, e.g., the solid oesophagus of the tadpole, which 

 which are not and could not be functional, can form no part of an 

 ancestral series ; a consideration well expressed by Sedgwick ' thus : 

 ' Any phylogenetic hypothesis which presents difficulties from a 

 physiological standpoint must be regarded as very provisional indeed.' 



1 Sedgwick, ' On the Early Development of the Anterior Part of the 

 Wolffian Duct and Body in the Chick,' Quarterly Journal of Microscopical 

 Science, vol. xxi., 1881, p. 456. 



