12 



SCIENCE. 



[Vol. V., No. 100. 



portion is small. Since, now, the heat of combus- 

 tion of cellulose is the same as that of starch, ac- 

 cording to von Rechenberg's determinations, 1 the 

 difference in the nutritive value of the two must be 

 measured by the heat of combustion of the marsh- 

 gas and hydrogen evolved. 



The well-known experiments of Henneberg and 

 Stohmann on the respiration of sheep showed no 

 considerable excretion of either hydrogen or marsh- 

 gas. In one of them, for example, the animal ate 

 per day 1,216 grams of hay, and excreted 1.5 grams 

 of marsh-gas. Not having at hand the original ac- 

 count of the experiment, we will assume that the 

 hay contained only twenty-five per cent of crude 

 fibre, of which one-half was digested. This amounts 

 to 152 grams per day. This quantity of cellulose, if 

 oxidized to carbonic acid and water, would yield 676,- 

 704 cal. 2 From this we have to deduct the amount 

 of heat carried off in 1.5 grams of marsh-gas, which, 

 according to Favre and Silbermann, amounts to 19,595 

 cal. There remain 657,109 cal., representing the 

 worth of the 152 grams of cellulose to the animal. 

 The same weight of starch, if completely oxidized, 

 would yield 680,808 cal. : in other words, the cellu- 

 lose set free in the body of the animal ninety-six and 

 a half per cent of the energy which the same weight 

 of starch would have done. 



Naturally these calculations are not exact ; but 

 they serve to show, that, if the heat liberated during 

 the fermentation of the cellulose is of use to the ani- 

 mal, the nutritive value of cellulose does not fall so 

 much below that of other carbohydrates as some are 

 inclined to believe. H. P. Armsby. 



IS THE RAINFALL OF KANSAS IN- 

 CREASING?* 



Thirty years ago the territory of Kansas was not 

 occupied by the white man, and, if we except a few 

 acres cultivated by the Delaware Indians, no portion 

 of her soil had been turned up by the plough. Her 

 entire area was included within the vast and almost 

 unknown region of the ' treeless plains ' and the 

 ' great American desert.' During that brief inter- 

 vening period, more than a million people, chiefly of 

 the agricultural class, have taken possession of her 

 domain, and have already brought her to the very 

 front rank of the states of the Union in the extent 

 and value of her agricultural products. History 

 affords no other instance of the permanent occupation 

 of so extensive an area, previously unoccupied by man, 

 by so large an agricultural population, in so short a 

 space of time. Here, certainly, if human agency 

 could anywhere affect climate, would such an effect 

 be produced. Here, assuredly, if settlement ever in- 

 creases rainfall, will such increase be most marked 

 and most unmistakable. That such increase has ac- 



1 Journ. prakt. chem., n.f., xxii. 1 and 223. 



- 1 cal. = the amount of heat required to raise the tempera- 

 ture of 1 gram of water 1° C. 



3 Lecture before the Kansas academy of sciences, Nov. 25, 

 by l'rof. F. H. Snow. 



tually taken place, I believe to be established beyond 

 a doubt. It is a circumstance peculiarly favorable 

 to the determination of the point in question, that, 

 although the general settlement of Kansas by cultiva- 

 tors of the soil is of such recent date, reliable obser- 

 vations upon the rainfall had been made at the military 

 posts upon the eastern borders for a sufficient period to 

 make possible a satisfactory comparison between the 

 rainfall before settlement and after settlement. The 

 records at Fort Leavenworth cover the longest period, 

 and enable us to compare the nineteen years imme- 

 diately preceding the occupation of Kansas by white 

 settlers with the nineteen years immediately follow- 

 ing such occupation. During the first period the 

 average rainfall was 30.96 inches; during the second 

 period it was 36.21 inches; giving an average increase 

 of 5.21 inches per annum, — an increase of nearly 

 twenty per cent. The Fort Leavenworth records 

 cover so long a period of time (nearly forty years), that 

 the increased average of the second half of the period 

 cannot be attributed to a mere ' accidental variation.' 

 In the issue of Science for April 18, 1884, it is stated 

 that " the supposed increase in the rainfall in the dry 

 region beyond the Mississippi is not borne out by the 

 returns of the signal-service." But the records of 

 the signal-service upon which this statement was 

 based include a period of only twelve years of obser- 

 vation (from 1871 to 1882), which is undoubtedly too 

 short a period for either establishing or disproving 

 the fact of a ' secular ' variation. 



But the fact of an increased Kansas rainfall does 

 not rest entirely upon the Fort Leavenworth obser- 

 vations. There are other stations in Kansas whose 

 records cover a much longer period than that of the 

 longest established regular station of the signal- 

 service. There are the twenty years' records of the 

 U. S. military post at Fort Riley, the twenty-four 

 years' records of the State agricultural college at 

 Manhattan, and the seventeen years' records of the 

 State university at Lawrence. If these several periods 

 of observation be divided into two equal parts, in 

 each case it is found that the average rainfall of the 

 second half is notably greater than that of the first 

 half. At Fort Riley the increase amounts to 3.05 

 inches per annum, and at Manhattan to 5.61 inches per 

 annum, and at Lawrence to 3.06 inches per annum. 

 Expressed in per cent, the rainfall of these three sta- 

 tions has increased in the second half of each period 

 of observation, at Fort Riley, thirteen per cent; at 

 Manhattan, twenty per cent; and at Lawrence, over 

 nine per cent. If the increased rainfall could be 

 shown by the records of a single station only, or if 

 the several stations with sufficiently long periods of 

 observation exhibited discordant results (some indi- 

 cating a decrease, while others indicate an increase), 

 or if even a single station indicated a diminished 

 rainfall, the fact of a general increase would lack 

 satisfactory demonstration. But the entire agreement 

 of the four stations whose records have been used in 

 a discussion of this question seems to establish be- 

 yond doubt the fact of an increased rainfall in the 

 eastern half of Kansas. 



There can be no reasonable doubt that the general 



