June 3, 1897] 



NATURE 



99 



would have profited had he taken a more detached 

 view of his subject, and had attempted to define the 

 grounds of our belief in the prevalence of universal law 

 biological phenomena without special regard to Roux 

 >r any other author. 



In the second part of this book, Dr. Hertwig deals with 

 le methods of developmental mechanics as put forth by 

 loux and his school. Roux is one of those who specially 

 'presents a modern tendency to magnify the importance 

 experiment in biological investigation, and to mini- 

 ise the value of what is called " mere observation." It 

 has been pointed out, over and over again, that there is 

 no sharp distinction between experiment and observation^ 

 hat some of the most complete, the most "mechanical" 

 of the sciences — as, for example, astronomy — are, from the 

 nature of things, sciences of observation, and that what 

 we call experiment is but a means to an end, the end in 

 every case being observation of the change of state follow- 

 ing the disturbance which was the first step introduced 

 by the experiment. Here, says Dr. Hertwig, there is a 

 great difference between the inorganic and the organic 

 world. Inorganic substances are relatively stable ; the 

 simple contemplation of them would lead to no result. 

 It is only by disturbing their condition of equilibrium, by 

 causing changes of state in them, that we can observe a 

 succession of phenomena which will enable us to assert 

 relations of cause and effect. With living things the case 

 is different ; they are in a constant state of flux, and 

 observation presents us with an unbroken succession of 

 changes whose laws may be traced by the aid of reflec- 

 tion. What is true of biology in general is especially 

 true of the study of the development of organisms, since 

 it is during their development that organisms exhibit the 

 most numerous, the most constant, and the most unbroken 

 succession of changes. In depreciating the results of 

 observation, and in drawing a brilliant picture of the 

 results which are to follow from the use of experiment in 

 biology, Roux, and those who, like Yves Delage, have 

 followed him, have fallen into the error of supposing that 

 observation is necessarily an act of contemplation with- 

 out reflection, and that reflection added to observation 

 raises the last-named to a level which has neither been 

 attained nor sought after by the majority of workers. 

 Such an error is easily refuted by Dr. Hertwig, who 

 shows that in the domain of embryology the method of 

 observation has been strikingly fruitful of results ; that 

 embryologists have been distinguished for the readi- 

 ness, sometimes by the over-eagerness, with which 

 they have drawn conclusions as to cause and effect from 

 the results of their observations ; and, finally, that the 

 much-praised method of experiment, when applied to em- 

 bryological research, has been comparatively barren of 

 result. We have, indeed, learnt some new and striking 

 facts from the experiments of Driesch, Wilson, Zoja, 

 and of Hertwig and Roux themselves ; but the main 

 result of these experiments has been to overthrow the 

 speculations of the school which plumes itself on its 

 insight into the mechanical processes of development. 



To most persons the most interesting part of Dr. O. 

 Hertwig's book will be the appendix, which contains 

 critical observations on the mechanical laws of develop- 

 ment promulgated by Roux, This author, it will be 

 remembered, as a result of experiments and observations, 

 NO. 1440, VOL. 56] 



made by him on the developing ova of the frog, formiu 

 lated a series of developmental laws, of which the chief 

 were : that the plane of the first segmentation coincides 

 with the sagittal plane of the future embryo ; that the 

 four first blastomeres correspond to four definite regions, 

 right and left anterior and right and left posterior 

 of the future embryo ; and so forth. From these so-called 

 laws theoretical conclusions of wide applications have 

 been drawn. It has been asserted that the organs and 

 tissues of the adult are represented by equivalent ultra- 

 microscopical particles in the ovum, and that these 

 particles are distributed in the ovum in a manner corre- 

 sponding to their prospective situation in the adult. The 

 particular manner in which these hypothetical particles 

 give rise, in the course of development, to the tissues 

 and organs of the adult, is set forth in Roux's theories of 

 self-differentiation and mosaic-work. Dr. Oscar Hertwig 

 deals in detail with the developmental laws of Roux and 

 their consequences, and shows that, in the first place, the 

 laws have no validity, and that, therefore, in the second 

 place, the conclusions drawn Vfrom them are without 

 foundation. Roux destroyed one of the first two or four 

 blastomeres of the frog's ovum, and stated that the 

 embryo resulting from the development of the remaining 

 blastomere or blastomeres was a half embryo or quarter 

 embryo, deficient in those parts which were contained in 

 the blastomeres which had been destroyed. Hertwig, 

 repeating the same experiments, showed that destruction of 

 any of the earliest blastomeres does not, in fact, lead to the 

 results which were described by Roux, and that the early 

 blastomeres might be, so to speak, shuffled up by means 

 of pressure, and yet a perfectly normal embryo be formed. 

 Hertwig's experiments are confirmed by those of Driesch, 

 Wilson and Zoja, who have shown that if the first 

 four blastomeres of echinoderms, of amphioxus, or of 

 hydromedusffi are separated from one another, each gives 

 rise to an embryo, normal in all respects, except that it 

 is a quarter of the usual size. Such experiments point 

 conclusively to the fact that the material of the egg is 

 not qualitatively, but quantitatively divided in the first 

 stages of segmentation, and that Roux's theories of 

 mosaic-work and self-differentiation are, therefore, without 

 foundation. In the appendix to this book the experiments 

 and conclusions of Roux, and the conflicting observations 

 of Hertwig, Driesch and others, are given in a concise 

 and easily intelligible form ; and the reader will hardly 

 fail to be persuaded that the balance of evidence is in 

 favour of Hertwig's opinion. At the same time, it must 

 not be forgotten that there is a considerable mass of 

 evidence in favour of Roux's main proposition, that in 

 normal development the first segmentation planes do 

 mark out particular regions of the future embryo. Dr. 

 Hertwig has handled Roux somewhat severely for his 

 definitions of normal and abnormal development ; but 

 the truth seems to be this. That the ovum of any animal, 

 if left to itself, will go through a course of development 

 which may be called normal, of which one of the features 

 is that the symmetry of the adult is expressed by the 

 earliest cleavage planes. But this course of development, 

 though normal, is not necessary. If the blastomeres be 

 displaced, or even if they be separated, development will 

 complete itself, and the end result is the same, though the 

 first steps have been different. The conclusion is that 



