192 



NATURE 



I April b, 191 1 



PofUMRisiNG Astronomy. — From the Rochdale Times 

 wt: see that the Rev. W. G. Pritchard i<« making an excel- 

 lent innovation, for the {x>pulari!»ation of astronomy, in 

 connection with the Kdiuation Ciuiid of the town. The 

 <juiid meets frequently for the discussion of art, bcience, 

 and literary topics, and the programme for Tuesday night 

 was an open-air talk on the stars. The weather being 

 favourable, the members were to gather in the vicarage 

 iield and there discuss the various celestial objects, under 

 the leadership of Mr. Pritchard. We would commend this 

 programme for general adoption among similar associa- 

 tions. 



I^KI.ATIONS OF rilENOLOGICAL AND 

 CUM A 11 C VA RIA TION . » 

 ''PHE monograph referred to below deals with the flower- 

 ing date of thirty-nine plants for the years 1896- 

 1909. Unfortunately, only ten, or 25 per cent., were noted 

 for every year, and six were observed in seven to nine 

 only of the fourteen years. 



The observations were of herbaceous plants, shrubs, and 

 -trees that had grown at least two years in the meteor- 

 ological enclosure of the Royal Observatory grounds at 

 Uccle, two miles south of Brussels, on level ground, 340 

 feet above sea-level, all on clay soil. Observations were 

 in each case made on the same individual plant, such as 

 were well exposed being selected. In five chief essentials 

 ihey were therefore ideal, identity of well-placed speci- 

 mens, of soil, of location, of elevation, and of the observer, 

 jM. Jean Vincent, always on the spot. The series opens 

 with Corylus avellana, February 25, and closes with Aster 

 horizontalis (1900-7 only), September 15. The months re- 

 presented are February i, March 2, April 10, May 14, 

 June 7, July 4, and September 9. 



After an introductory survey, in which reference is made 

 to the far greater number of factors now known to 

 influence the flowering date than was once supposed, M. 

 Vanderlinden notes the increasing importance assigned to 

 the completion by each species of its "period of repose," 

 on which, largely, the mean date of flowering depends. 

 The research dealt largely with the influence, in associa- 

 tion with this, of meteorological influences, as shown both 

 by observation and experiment. 



The test for " flowering," as customary also in this 

 country, was the exposure of the stamens. It is not, 

 however, stated whether for the hazel the pistils were 

 observed instead, these being far less erratic than the 

 staminiferous flowers. 



Flowering is much more definite, and therefore better 

 suited for such observations, than other phases, such as 

 leafage, fruiting, and defoliation. The first tables give 

 the flowering date for each year, with the mean for the 

 years observed. It would surely have been well to inter- 

 calate dates for the missing years. To do this salis- 

 factoiily is indeed rather complicated, and such values are 

 not equal to actual observed dates. But it would be safe 

 to count the error as at most a quarter of that where 

 such precaution is omitted. Thus the mean date given for 

 Ribes nigrum, April 10, is that from 1903 on. The corre- 

 sponding years for Corylus avellana give February 19, but 

 that tabulated (on thirteen years of the fourteen) is 

 February 25. For the other years, 1896 and 1898-1902, the 

 average is March 3. The divergence at the later dates, we 

 «hall see, would be much less, but the argument would be 

 equally afTected. Thus Ribes rubrum is given April 10 

 <i903-9), R. alpinum April 14 (1898-1909). But the latter, 

 on the mean of 1903-9, should be April 10-9. These dis- 

 crepancies are less important for the investigations in hand 

 than had these dealt more specially with relative dates of 

 flowering, but they can hardly be neglected. 



In looking at the dates, it is noticeable how much less 

 range there is from the mean in the case of the earlier 

 flowers than in British observations of the few for which 

 there are common records. Thus for the hazel, for Uccle 

 and Purley (Surrey), respectively, the range since 1906 is 

 from February 17 to March 10, and February i to 



^ " Etude sur les ph inomencs P^riodique<; de la Vieitation dans leurs 

 Rapports avec la Variations climatique';." By Dr. E. Vanderlinden. Ex- 

 T-^L'i '^" ^«<^ueil de ri nstitut botanique L^o Errera, tome viii. Pp. 67, with 

 Tabl-sand 16 plates. (Bruxelles : Hayez, Imprimeur des Academies Royales 

 «9io.) 



NO. 2162, VOL. 86] 



.March 20; blackthorn, April ti to May i, and March !~ 

 to April 27; but for doy row.-, May 26 10 Jun- 4, ., 

 June I to to. 



The next subject dealt with is the rehr 

 departures from the mean flov -■ • - 'ue ana i; 

 sponding variations in the m- .1 factor*, 



parative curves are given for t _ ., namely, ni 

 and minimum temperature, radiation as shown by I 

 alcohol radiometer, humidity, and rainfall. The 1. 

 Were found to be relatively negligible, and, in the aailf 1 

 months, the same is true of the third. Fk>rew<.nc* «1 - 

 |)ends, then, mainly on temperature, as the • ' 'hnI 



stores are already present. I-ater on, fc^iage dc 



flowering to supply chlorophyll, which nee* .-> , "iii»t. 



But from June on heat is again the predominating factor, 

 since the interval between foliage and flower is so long. 

 These curves are given year by year. It is by careful 

 examination of these, and confirmation of results, where 

 possible, by e.xperiment, that M. Vanderlinden reaches his 

 conclusions. Naturally, there is always a certain amount 

 of lag, but this is less with the herbaceous planu. The 

 most effective combination is high temperature and radia- 

 tion, with feeble humidity lasting for several days before 

 the normal flowering date. 



The chief experiments ccmsisted in subjecting the plants 

 to warm baths, to moist warm air, to various light con- 

 ditions, to special warmth treatment over a definite time, 

 followed by ordinary conditions. Some twenty conclusion* 

 are drawn, among which, besides those already mentioned, 

 are the following : — 



When blooming has been retarded it follows upon less 

 stimulus. 



Phenology is practically evidential for temperature and 

 sunshine alone, and then only for approximate values. 



Effects may remain latent (and so cumulative) over short 

 periods. 



Heredity determines the normal date. 



Autumn and early winter have no influence [in Belgium]. 



Groups flowering concurrently vary concurrently. The 

 evidence given for this is perhaps too limited for the con- 

 clusion. 



It is certainly so in another case. .Anyone who has 

 dealt much with sun-spots would hardly venture on any 

 conclusion from data confined to fourteen years. Hence 

 the statement that no relation shows itself should rathf-r 

 have been that, as the observations have not yet be<*n 

 carried on for fifty or one hundred years, it is too early 

 to investigate the matter. The sun-spot table was hardly 

 the best way to utilise the space. Is not this true also of 

 the fourteen tables of daily temperatures and radiation? 

 One would have been content with a summary to cwn- 

 pare with the valuable plates, based upon the figures, if 

 instead we could have had further investigations worked 

 out from the observations. It would have been especially 

 interesting to have had relative results month by month, 

 as, for instance, the relation between annual variation and 

 the cumulative values above some minimum, below which 

 the given plant showed no response to the effective factors 

 of heat and light. 



But, in asking for more, it must be understood that 

 this is because of the excellent value of that which is 

 given. J. Edmund Clark. 



NON-E UC LI DEAN GEO ME TR Y. 



T T is now well known by all mathematicians that 

 Euclid's theory of parallels is not indispensable for 

 the construction of a self-consistent geometry, but that, 

 on the contrary, there are three coordinate systems, of 

 which Euclid's is one, equally entitled to consideration, 

 and equally general. So far as we can see at present, the 

 strict proof of this statement must be analytical ; at any 

 rate, when we suppose that the elements — lines, points, 

 planes, &c. — are, in the space considered, exactly 

 analogous to the corresponding elements in Euclidean 

 space. However, it fortunately happens that we can con- 

 struct a non-Euclidean geometry in ordinary space by 

 suitably changing the definitions of its elements, and this 

 is, at any rate, of considerable help in convincing a student 

 of the possibility of the non-Euclidean systems. Prof. 

 H. S. Carslaw has recently explained one such method 



