Ma^ ig; v^M] 



MATURE 



203 



common' in the older text-books of cellulose as a 

 very stable substance, (i) that cellulose is con- 

 stitutionally modified by any and every treatment 

 with reagents, and (2) that profound changes 

 affecting the reactivity of its individual groups are 

 determined by treatments which are not marked by 

 change of weight of the cellulose or by visible 

 structural modifications. 



It thus follows that surgical cotton-wool or 

 chemical filter-papers, for example, which are often 

 considered as "pure cellulose " and the criteria of 

 purity adopted by various manufacturers, are 

 selected on an empirical basis, and the authors, 

 therefore, attempt to define the "normal standard." 

 Their definition of this as cotton purified from 

 its raw condition by such treatments as attack and 

 remove its non-cellulosic components with the as- 

 certained minimum of action upon the cellulose 

 itself obviously raises difficulties in the verifica- 

 tion of the so-called " standard " product. Their 

 conclusions are most valuable as a stimulus to 

 further research, and their statement that "every, 

 process of treating the vegetable fibres in the arts 

 produces some constitutional change " shows how 

 necessary is systematic research work organised 

 on a co-operative basis for the continued well- 

 being of all the textile and. paper-making trades. 

 Robert H. Pickard. 



The Year-Book of the Scientific and Learned 

 Societies of Great Britain and Ireland. Com- 

 piled from official sources. Pp. vii: + 334. 

 (London: Charles Griffin and Co., Ltd., 1917.) 

 Price gs. 

 The thirty-fourth annual issue of this useful work 

 of reference provides a convenient record of the 

 work done in science, literature, and art during 

 the year 1916-17. Not only are the activities 

 of the scientific societies chronicled, but an account 

 is provided also of the researches carried out by 

 the Meteorological Office, the National Physical 

 Laboratory, the Royal Observatory at Greenwich, 

 the Geological Survey, Kew Gardens, Rothamsted 

 Experimental Station, and similar organisations 

 of a national character. 



The volume is very comprehensive in scope, but 

 in the science sections we miss references to the 

 Society of Glass Technology and to the Illuminat- 

 ing Engineering Society, both of which are 

 active and important. 



LETTERS TO THE EDITOR, 

 [the Editor does not hold himself responsible for 

 opinions expressed by his correspondents. Neither 

 can he undertake to return, or to correspond with 

 the writers of. rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 taken of anonymous communications.] 



Cotton-growing Statistics. 



The article on the above subject in Nature of 

 April II was welcome to me as directing attention 

 to very important economic possibilities which lie 

 behind the making of precise reports and reasoned 

 forecasts concerning the state .af. all crops, by means 

 NO. 2533, VOL. lOl] 



of thePlant-Development-Curve method which I de- 

 vised'during my service in Egypt. With your permis- 

 sion, however, 1 would like to point out that the 

 writer of the article was under a misapprehension in 

 thinking that the recent Egyptian data he mentioned 

 were adverse to this new method. 



Although it is true that the paper cited from the 

 Agricultural Journal of Egypt (vol. vii.) states that 

 "the object of (i) was to study the life-history of the 

 various types of cotton in the different parts of the 

 country and to ascertain whether it was possible to 

 estimate the yield of the crop several weeks before the 

 cotton was ready for picking," yet the absence of 

 further comment is obviously due merely to a certain 

 looseness of structure in the paper in question, 

 whereby a summary of results obtained under the 

 stated purpose of this section "(i)" is given some- 

 what irrelevantly as follows on p. 52: — '"The above 

 observations seem to point to the fact that watering 

 experiments on the cotton crop are most necessary ; 

 for if the flowering curve could be maintained {sic) 

 during the whole of July instead of dropping con- 

 siderably in the- middle of the month, there should 

 be a considerable increase in the yield of the crop." 



Any decision as to the real result of this section 

 of the paper in its declared object must therefore be 

 taken from the data themselves, which are published 

 in the form of plant-development curves. 



So far as the prediction of boiling from flowering 

 is concerned, with a seven-week interval between, the 

 satisfactory nature of the results may be seen at a- 

 glance by those who care to consult the paper in- 

 question (although each curve represents the be- 

 haviour of only a single observation row of only fifty 

 plants, instead of being the mean of at least five such 

 rows of one hundred plants). In thus comparing the 

 flowering and boiling curves, the reader should dis- 

 card the first weekly ordinate in every boiling curve 

 as there published, since the counting of the bolls 

 was not started soon enough, and this ordinate, there- 

 fore, includes earlier bolls from the second, and even- 

 the third, antecedent week, which naturally spoils the 

 similarity. On placing the curves in superposition 

 with a seven-week shift, these new data {e.g. Fig. 9 

 and Fig. 11) then illustrate most satisfactorily the 

 points which I have discussed in detail in "Analyses 

 of Agricultural Yield" (Phil. Trans. Roy. Soc, 

 B. 327, . 333, 352), and specially demonstrated by 

 Fig. 17 in part iii. of the " .'\nalyses," such aS the 

 incidence of hot days, waterings, and boll-worm in 

 affecting the shedding. 



As regards the forecasting of the flowering itself by 

 the growth-curve, the data under discussion have no 

 significance. The daily fluctuation of flowering' of 

 cotton in Egypt is not predicted by main-stem growth 

 measurements made later than June i (see Fig. 10 in 

 my " Development and Properties of Raw Cotton "), 

 owing to the four weeks' duratipn of this predeter- 

 mination period, and to other causes which I dis- 

 cussed on p. 182 et ante of part iii. of the "Analyses." 

 Since the growth-curves in vol. vii. of the Agricultural 

 Journal of Egypt only began on May 28, they are 

 effectively not for this purpose growth-curves at all. 

 In any case, the growth should be measured daily, 

 and not merely once a fortnight. 



W. Lawrence Balls. . 



St. James's Square, Manchester, April 16. 



Dr. Balls 's comments on the short article on 

 " Cotton-growing Statistics " in the issue' of Nature 

 for April 11 opens up a wide and interesting feature 

 in sdentific research, viz. the value of observed data 

 and their interoretation. 



