27- 



NA TURE 



[January 6, 1910 



centre. A depression round which tlie distribution of 

 wind force as compared with the prevailing gradient 

 is unsymmetrical will move towards the region of 

 "least resistance," i.e. the region where the winds are 

 most conspicuously in defect. In identifying the 

 region of least resistance the second principle is also 

 used. It is based on the conception of "divergent" 

 winds, kny wind which has a component directed 

 away from a centre of low pressure is divergent for 

 that centre, and as such marks a region of low resist- 

 ance to its advance. Generally speaking, the greater 

 the "divergence" the less the "resistance." Strong 

 northerly or north-westerly winds to the eastward of a 

 depression are looked upon as an extreme case of 

 ■divergence, and as a sure sign of a rapid advance of 

 the depression. 



iM. Guilbert proceeds to elaborate no fewer than 

 twenty-five rules for forecasting which for the most 

 part follow more or less directly from the two funda- 

 mental principles. Their application is illustrated by 

 a large number of examples, taken mostly from cases 

 -when the forecasts issued by the Bureau at Paris 

 proved incorrect. 



.So far as M. Guilbert is concerned, both principles 

 must be looked upon as a direct result of the careful 

 scrutiny of weather maps; they are entirely empirical, 

 and no attempt is made to justify them from general 

 dynamical considerations. Since the book has been 

 in our hands, we have watched the charts published 

 in the Daily Weather Report, and have noticed occa- 

 sions on which the application of the first principle 

 would apparently have been useful. On other occa- 

 sions we have found difficulty in applying the rule. 

 On many maps there are, within one and the same 

 meteorological region, winds which are, some in excess, 

 others in defect, of the normal. M. Guilbert gives no 

 instructions as to how to proceed in such cases. 



The principle of the "divergent" wind is not likely 

 to be accepted without qualification in the form 

 in which it is put forward. M. Guilbert insists very 

 strongly that a single conspicuously divergent surface 

 wind (observations on mountains or at high levels are 

 expressly ruled out as not being comparable with the 

 surface gradient) must be regarded as an almost in- 

 fallible indication of the early advance of a depres- 

 sion. Objections based on the argument that 

 surface winds are very liable to be influenced by local 

 conditions are brushed aside. .Among the examples 

 quoted in the book there are many instances of sur- 

 prisingly daring and successful predictions, some ap- 

 parently ex post facto, others attested by stamped post- 

 cards as being genuine forecasts made before the 

 event. We are, however, entitled to ask whether the 

 rules might not lead to equally daring but unsuccess- 

 ful forecasts? Nearly too examples are quoted; the 

 number is large, but so is the number of charts from 

 which the selection is made, and it is to be supposed 

 that iM. Guilbert, whose style often suggests counsel's 

 address to the jury rather than the judge's summing 

 up, has picked out the cases which best illustrate his 

 points. No doubt he could produce many more 

 instances if called upon to do so, but the multiplication 

 of selected examples does not carry conviction. 

 NO. 2097, VOL. 82] 



Before passing judgment we must hear coun- 

 sel for the other side, who may be able to 

 bring forward a similar number of cases at variance 

 with the principles advanced. M. Guilbert has run up 

 against the old difficulty of expressing his hypothesis 

 in such a manner that it can be tested by an appeal to 

 measurement. The relation between the observed wind 

 velocity and the gradient should lend itself to numeri- 

 cal treatment, seeing that both quantities are the 

 object of regular observation and measurement. 



The book is accompanied by a preface by Prof. 

 Bernard Brunhes, the director of the observatory on 

 the Puy de Dome, who acted as reporter on the occa- 

 sion of the competition at Liege. In a supplement 

 M. Brunhes points out that M. Guilbert's rules are 

 consistent with the results deduced by Lord Kelvin 

 and Bjerknes for the action of a steady current on a 

 vortex, and describes some laboralorv experiments of 

 his own illustrating the phenomena. 



MEDICAL EMBRYOLOGY. 



Text-book of Embryology. By Dr. Frederick R. 



Bailey and .\dam M. Miller. Pp. xvi4-672. 



(London : J. and .\. Churchill, 1909.) Price 21s. 



net. 



THIS bulky volume is the third American text- 

 book of embryology that has appeared in recent 

 years. Like Prof. Heisler's work, of which a revised 

 edition was published two years ago, it is addressed 

 primarily to students of medicine and anatomy, being, 

 in fact, based upon the course in embryology given 

 at the medical school of Columbia University. Con- 

 sequently it differs in its style of treatment from 

 Prof. Lillie's "Development of the Chick," which is 

 apparently intended for zoological students beginning 

 embryology ; and whereas Prof. Lillie confines his 

 attention almost exclusively to a single type, and 

 never passes outside the class Aves, the authors of 

 the present volume, although dealing more particu- 

 larly with human development, have aimed at treating 

 the subject from a comparative standpoint, believing 

 this to be the most efficient way of teaching it. 

 With this opinion most teachers of biology must 

 surely concur. It is satisfactory to note also that the 

 phvsiology of the developmental processes is not 

 entirely passed over. Thus, in an excellent chapter 

 on the nervous system contributed to this volume by 

 Dr. Oliver S. Strong, the author has been able to 

 include much physiological matter which usually 

 finds no place in a text-book of embryology. 



In the second chapter a section is devoted to 

 ovulation and menstruation and the relation which 

 subsists between these processes, but it is to be re- 

 gretted that the comparative method which is so 

 successfully followed in other parts of the work is not 

 extended to the problems dealt with here. That 

 menstruation in the Primates is the physiological 

 homologue of the prooestrum in the lower Mammalia, 

 and that in the latter ovulation occurs normally dur- 

 ing oestrus, are now generally accepted facts, the 

 recognition of which has removed many apparent 

 difficulties which used to perplex the older writers. 



