OTHER GRADIENT THEORIES 171 



on a number of occasions, and their importance for morpho- 

 genesis has been emphasized by such workers as Boveri, von 

 Ubisch, D'Arcy Thompson, and notably Child, 1 who has 

 systematized the theory more thoroughly than other workers. 

 The more general conception of the morphogenetic field, of 

 which I believe gradient phenomena to be a particular case, 

 has been analysed from somewhat different points of view by 

 Weiss (1926), Guyenot (see Guyenot and Ponse, 1930), Hirsch 

 (1931) and Bertalanffy (1928). This is not the place to enter 

 into a general discussion of the subject, and I propose merely 

 to refer to a few relevant facts and ideas. 



In the first place, two sets of essentially morphological facts 

 concerning gradients have long been known and recognized. 

 The first is generally subsumed under the title of the law 

 of antero-posterior development. It points out that during 

 development differentiation begins anteriorly and gradually 

 spreads posteriorly. Often the development of the head is 

 far advanced when that of the hinder end is not yet begun. 

 In some forms the undifferentiated posterior region may per- 

 sist throughout life ; or, as in many Crustacea, it may persist 

 through a considerable phase of free-swimming existence 

 although lost in the adult. In addition, there exists in bilater- 

 ally symmetrical animals a similar gradient in time of develop- 

 ment between dorsal and ventral surface. In Vertebrates the 

 region which leads the way is the dorsal mid-line, in Inverte- 

 brates in general the ventral mid-line. Subsidiary graded 

 effects of similar nature also occur within the appendages. 



Secondly, a gradient also almost invariably occurs within 

 the ovum, along the main axis. This may be revealed in the 

 stratification of yolk or other materials, or in the greater rate 

 of segmentation at the animal pole, or in both ways. It is 

 the merit of Child that he has linked up these two sets of 

 facts in one general physiological theory. He has further 

 shown that the physiological gradient effects of which these 



1 I prefer not to use the term metabolic gradients, also sometimes 

 used by Child. Child has not conclusively demonstrated that his 

 gradients are fundamentally metabolic in character ; but he has 

 demonstrated that, as regards morphogenesis, gradient-systems do 

 exist and are operative. The most important facts about these mor- 

 phogenetic gradient-systems are (a) that they are field systems in 

 which all the parts are interdependent within one plastic system, and 

 (b) that in some respects at least they are quantitatively graded. 

 Personally I would prefer the phrase morphogenetic gradient-fields, but 

 there is no need at the moment to complicate terminology thus. For 

 an excellent discussion, see Needham, 1931, p. 582 seq. 



