J™,^^' MORPHOLOGY IN ZOOLOGICAL SCIENCE. 21 



stitutes ontogeny. It is therefore to be looked on as a new kind of 

 dissection. Many embryologists have attributed to the visible units 

 observed in a segmenting egg, independence of growth and a definite 

 function in building up the organism. The experimental embryo- 

 logist, however, by forcing these units into abnormal positions, or by 

 isolating them, is able to demonstrate that such attributes are 

 impossible, and that the real factors into which development is 

 separable are not identical with the visible differentiation which the 

 germ exhibits. Driesch distinctly states that this dissection is the 

 whole aim of experimental embryology, but some of his followers 

 seem to imagine that by such means it is possible to discover the 

 ultimate causes which bring about variation. This idea is, I think, a 

 totally mistaken one. No examination of the changes which, by the 

 application of physical or chemical means, we are enabled to bring 

 about in the outer structure of the organism, will bring us one step 

 nearer the discovery of those causes which are able to modify its 

 inner hereditary potentiality. We may, like Driesch, force that part 

 of the egg which normally produces the head to give rise to the tail, 

 or, like Herbst, turn endoderm into ectoderm ; all we arrive at is a 

 resultant of the combined working of hereditary tendency and effect 

 of environment. 



The study of individual variations stands on a different basis. 

 The modifications which species undergo are ex hypothesi made up of a 

 summary of individual variations, and it seems quite proper to begin 

 with a consideration of these ; but at the outset we are met with a 

 most serious difficulty. We have no means of distinguishing modifi- 

 cations, which have been produced by the action of environment on 

 the particular individual we examine, from those due to variations in 

 its hereditary qualities. Still further, the individual variations which 

 appear most prominent to us are often those which there is strong 

 reason to believe have never taken part in the evolution of species. 

 In his " Materials for the Study of Variation," Mr. Bateson has 

 collected numerous instances of supernumerary digits in Vertebrata 

 and of branched legs among the Insecta ; yet, in all the hundreds of 

 species of Mammalia, and the hundreds of thousands of species of 

 Insecta, no solitary instance is known where a variation of this kind 

 has become the distinguishing mark of a species. Even of those 

 changes due to variation in the hereditary powers, it is clear that only 

 a small proportion concern us. For, in order to become a real factor 

 in evolution, it is necessary that a variation should not only be trans- 

 missible to the offspring, but further, that it should occur sufficiently 

 frequently to give natural selection an opportunity of taking advan- 

 tage of it. 



The Statistical Study of Variations, which we owe to 

 Dr. Galton and Professor Weldon, cannot, I think, be assailed 

 on theoretical grounds. This method aims at representing in a curve, 

 not only the extent of variation in a given character met with in a 



