242 



NA TURE 



\7uly 30, 1874 



been established, in 1873 there were 178, so that at pre- 

 sent there is more than one gauge to every square mile. 

 By this system the conditions of local rainfall have been, 

 as it were, put under the microscope ; and the store of 

 information obtained after the suggestive manner in which 

 it has been analysed in this Report, will be not only 

 valuable to the sugar-planter, but interesting to the 

 meteorologist. 



The north-east trade-wind prevails at Barbados during 

 three-fourths of the year, and most of the rain comes 

 from that quarter. Heavy showers come at certain sea- 

 sons from south-west and north-west, but generally fail to 

 reach the eastern districts. Indeed it very rarely happens 

 that rain falls at the same time, or in equal proportions, 

 over the whole island ; it has, therefore, been divided 

 into two main districts, the windward, and chiefly high- 

 land, and the leeward, or lowland section. 



With regard to the yearly rainfall of the whole island 

 from I S47 to 1 87 1, it has been found, among other results 

 (i), that the rainfall of fifteen years was above the 

 mean, that of ten years below it ; (2) that the deficiencies 

 were generally greater than the excesses above the mean, 

 that is, droughts, when they happen, are heavy ; (3) that, 

 taking the thirty years 1843-72, no succession of wet or 

 dry years in cycles can be traced, but rather an alter- 

 nation of wet and dry years. No more than two dry 

 years have occurred together, but as many as three and 

 four wet years. 



With regard to the monthly rainfall : the mean of all 

 the months is under 5 in. ; March is driest, October 

 wettest. In wet years May contributes most to the ex- 

 cess, March least. March is the only month of which 

 the mean rainfall in dry years has exceeded the average. 

 In dry years the deficiency is generally spread over the 

 whole'year ; in wet years the excess is generally confined 

 to the rainy season (autumn). On the other hand, taking 

 the seven wettest and the seven driest years of the period, 

 we observe that in the wet years two-thirds of the excess 

 proceeds from heavy rains in the dry season, and that in 

 the dry years more than two-thirds of the deficiency is 

 caused by three out of the same four months, viz. June, 

 July, and May, and by October and May. 



A comparison between the three highest and the three 

 lowest rainfalls belonging to each month during the 

 whole period of twenty-five years shows a remarkable 

 uniformity of the relation between the percentages of the 

 extremes ; thus the difference between each average of 

 the lowest months and each average of the highest 

 amounts in none of the twelve months to more than 90 

 (May) or less than 65 per cent. (August and September). 



Of the two stations, Binfield and Hahon, lying respec- 

 tively at 1,065 and 280 ft. above the sea, the former 

 received on an average nearly 1 1 in. more rain in the year, 

 and showed a greater monthly variability. The influence 

 of elevation is interesting. A table of rainfall in 1870 and 

 i87i,at various heights from 100 ft. up to 1,000 ft., shows 

 an increase at every step of 100 ft. but one, and the total 

 increase at 1,000 ft. amounts to 2073 in. on the mean of 

 two years. Two exceptions to this regular increment in 

 the means for 1871-73, in supplement No. 2, are ascribed 

 to the lower stations catching the westerly rains, which 

 do not penetrate far inland. 



In March, one-half more rain fell at night than by day. 



From June to November, the days are slightly the wetter, 

 from December to May, the nights. 



One of the objects of this inquiry is to assist those who 

 are interested in calculating the character of coming 

 seasons. For such a purpose the annual averages of, 

 each month are taken, after eliminating the exceptional 

 months of very great or of very slight rainfall. The 

 original averages are not affected more than 5 per cent, by 

 this removal. Having the ordinary limits of monthly 

 rainfall tabulated, and observing the general appearance 

 of the weather, every planter can form some conjecture 

 whether the coming month will be wet or dry. 



Appendix No. 36 shows the influence of each month 

 according to the rainfall, upon the crop of the same year, 

 and upon that of the following year. Thus a man might 

 fairly bet 9 to 3 that a wet February will be followed by a 

 bad crop, and S to i, the highest odds of all, that a wet 

 September will give a good crop next year. 



A wet year is followed almost invariably by a good crop 

 in the following year ; and it is found that by multiplying 

 the total rainfall of the preceding year by 800 and adding 

 yi per cent, if that year was a dry one, or subtracting 7^ 

 per cent, if that year was a wet one, the crop may be 

 calculated in most instances within 3,000 hhds., the 

 average yield of the island being 45,000 hhds. The 

 good chance of predicting so nearly the total exports of 

 Barbados for the coming year cannot fail to be of value, 

 and further experience will no doubt reduce the proba- 

 bility of error. Let us hope that other States may be led 

 to undertakings of the same kind by this example. 



F. A. R. Russell 



OUR BOOK SHELF 



Manual of British Botany, containing the Flowering 

 Plants and Ferns arram^ed accordins; to Natural 

 Orders. By C. C. Babington, M.A., F.R.S., F.L.S. 

 Seventh Edition ; corrected throughout. (London : J. 

 Van Voorst, 1874). 

 We cordially welcome this new edition of a " Manual 

 of British Botany " which continues to hold its ground 

 against all its competitors. We do not propose to discuss 

 the rival merits of Hooker's, Bentham's, and Babington's 

 hand-books ; each has specialities in which the others are 

 wanting ; and each will, no doubt, long have its advocates 

 and admirers. A special claim to popularity as a field- 

 book is advanced by the present work on the ground of 

 its portability ; and a great advantage is alleged by those 

 who use it to be presented by the practice of placing in 

 italics a few words in the description of each species 

 referring to the character by which it is more readily 

 distinguished from its nearest allies. Comparing the 

 work with the most recent of the earlier editions which we 

 have at hand— the 4th, published in 1856 — we find that it 

 extends only to twenty-six pages more, notwithstanding 

 the numerous additions made since that time to British 

 botany, of which ample account has been taken in the 

 present edition. The only alteration made in the prima' y 

 classification (comparing these two editions) is the sepa- 

 ration of CannabinaceK from Urticacea;. The number of 

 natural orders is six more than in Hooker's " Student's 

 Flora," notwithstanding the union of Salicacere, Myri- 

 cacea?, Betulacea?, and Cupuhfene into the somewhat 

 artificial group of Amentifera-. The location of indi- 

 vidual genera has also been in some cases revised, as the 

 removalof Nartliccium from Juncacex (Babington,4th ed.) 

 or Liliaceae (Hooker) to Melanthacea?. Referring to some 

 of the more difficult genera, in which Prof. Babington is 



