1897. HERTWIG ON MECHANICS AND BIOLOGY. 413 



due more to the patient collection of facts and the stringing together 

 of conclusions based upon observed sequences of these, than to 

 experimental work alone. On p. 75, Professor Hertwig gives an in- 

 stance from Roux's own work of the way in which one-sided experi- 

 mental work may lead to erroneous or unsound conclusions ; while he 

 draws attention to the work of Dareste, Gerlach, Weismann, Boveri, 

 Wilson, Morgan, and many others, both as a proof that he does not 

 underrate experimental work in itself and that the best results are 

 attained when observation and experiment are not divorced. 



The last pages of the volume are devoted to a series of critical 

 studies of Roux's laws of mechanical development, and deal with (i) 

 the mosaic theory, (2) the so-called " copulatory track," (3) certain 

 definitions, and (4) cytotropism. 



The first study treats of Roux's view that the development of the 

 gastrula in the frog is, from the first segmentation-furrow onwards, 

 a piece of mosaic work (in that a perfect whole arises out of various 

 independently differentiated parts), and, further, of the problem 

 whether there is, or is not, a constant relation between the direction 

 of the first plane of segmentation and the direction of the median 

 plane in the resulting embryo. Professor Hertwig regards Roux's 

 conclusion that there is such a relation as a preconceived theory, into 

 which Roux has, in some Mikado-like manner, contrived to make his 

 results " fit." In the case in question the results arrived at may be 

 valuable or the reverse, but we think Professor Hertwig deviates 

 somewhat from the path of fairness in the view he takes of them. 

 Surely it is too much to expect of any investigator that he should 

 start without a hypothesis at all ; this would be the very experiment- 

 ing "in's Geradewohl " which Professor Hertwig, in his neatly-chosen 

 quotation from Johannes Miiller, so deprecates. 



In the second study Roux's theory as to the influence of the pro- 

 cess of fertilization in determining the direction of the first segmenta- 

 tion-furrow is discussed, and somewhat severely handled. 



The third study calls for little special mention, many of the 

 questions therein raised having been already more fully dealt with 

 in Part I. of this work, and already described in Natural Science. 



The last chapter is devoted to a newly-discovered phenomenon 

 described by Roux as " cytotropism," which Hertwig understands 

 to mean a process, partially chemiotactic in nature, by which neigh- 

 bouring cells in a segmenting ovum so react upon one another as to 

 draw each other gradually into closer contact. Professor Hertwig 

 makes a careful survey of the facts which have led to these conclu- 

 sions, and believes that the former should bear a different interpreta- 

 tion. His comment is somewhat ironically expressed in a quotation 

 from Roux's own writings, in which it is said to be "easier to establish 

 a new fact than to arrive at a true estimate of its significance." In 

 conclusion. Professor Hertwig sums up his own position with regard 

 to these theories in three pithy paragraphs. 



