October 31, 1907] 



NA TURE 



67: 



Respecting the relative efficiency as protecting agents 

 of the Nipher jacl;et and the Wild fence, experimental 

 observations have been, and are apparently siill being, 

 conducted in Russia, a country, be it noted, where, owing 

 to the large quantities of snow that fall during the colder 

 months of the year, the sources of error in rain-gauge 

 records through wind assume greater magnitude than they 

 do, for instance, in England, where the winter precipita- 

 tion consists largely of rain. It may be said that, while 

 the Nipher protector is generally well adapted to its pur- 

 pose for ordinary situations of rain-gauges, it may be 

 with advantage replaced by the fence enclosure in the case 

 of gauges which are unduly exposed to the full force of 

 the wind in unsheltered locations. If the accuracy of the 

 Wild fence be taken as too, that of the Nipher jacket may 

 approximate to loo in more sheltered positions, but may 

 be as low as 80 or even lower in such as are quite open 

 to the violence of the wind. It should be added that a 

 Nipher gauge ought to be fitted with some form of heating 

 apparatus adequate to prevent accumulations of snow in 

 winter upon the protecting jacket, from which into the 

 rain-gauge portions of such accumulated snow are liable 

 to be blown. 



This brief abstract of the chief methods of diminishing 



nr oiiininating the wind error due to ra'n-gauges would 



be incomplete were it omitted to mention a process of 



calculation ' by which the rainfall figures for a sufficiently 



long period, as indicated bv a gauge at a pl.-tce which 



suffers undue exposure to the wind in comparison with 



another gauge in the neighbourhood, say a few miles 



distant, at a more sheltered spot that mav be regarded 



normal, may be corrected. The method depends upon the 



relation subsisting between the amounts of discrepancy in 



the records of two such gauges during periods of rain and 



of snow. If the rainfall for a specified time at the 



sheltered station be represented as 100, and that at the 



exposed station as 100 — A for periods of rain and 



X* -H A 

 100 — .4 — B for periods of snow, the equation K =- — j- — , 



when solved for v, affords the correction required. The 

 value of K, which for a few localities in Germany has 

 been found to range from 0-13 to 022, must be empirically 

 determined for a particular district by establishing two 

 similar rain-gauges close together, or, if possible, side by 

 side, one of which is fitted with an efficient wind-pro- 

 tection contrivance, the other being left free ; for a pair 

 of gauges in such close proximity .1: may be considered to 

 vanish, so that K becomes = .'\/B. The value of K for the 

 locality being thus found, x is solved = KB — .'\, which will, 

 of course, be a ^\us or minus quantity according as the 

 true rainfall is slightly greater or less at the exposed than 

 at the sheltered station. This method of calculation, which 

 is applicable in many instances, has been tested bv another 

 more direct one involving anemometer readings, wherebv 

 the measured quantities of precipitation could be reduced 

 to equal mean wind velocities, and as the two have given 

 most concordant results, it may be concluded that the one 

 briefly delineated above is correct. 



Thus at Karzig, in Neumark, the rainfall at an open, 

 wind-swept spot on the outskirts of a forest, though in- 

 dicated by a rain-gauge as considerably less than that of 

 a glade more than 2 kilometres distant, was found by 

 both processes of calculation -to be actually 2 per cent, 

 greater. 



To summarise the contents of this article : — 

 (il Experimental observations extensively carried out 

 during the nineteenth centurv in manv countries have 

 established the fact that in the measurement of rainfall 

 errors of considerable magnitude accrue from the presence 

 of the rain-gauge during the prevalence of wind, and point 

 to the conclusion that such errors arise from the eddying 

 or rebounding of wind about or from the mouth of the 

 rain-gauge. 



(2) The readings of a rain-gauge in a free, open situation 

 may be corrected by means of a method involving their 

 comparison with those of another similar gauge placed 

 at twice the height above the ground. 



^3) The most efficient wind-protection contrivances for 

 t .^^rfc-ayf/nfisc/'^ Zeitschrift, Pand xxiii., igcfi, ^. 444, " WalH und 

 Nied»r-schbs in Westpreussen I'nd Pos n und die Reeinfiussung der Regen 

 und Schnee-messunt: durch der Wind," von J. .Schubert. 



NO. 1983. VOL. 7^1 



rain-gauges are the Nipher jacket and the Wild fence 

 enclosure ; the latter, though more accurate and advant- 

 ageous in special circumstances, is generally less used 

 than the former. 



(4) The corrected rainfall for a sufficient length of time 

 of a wind-swept spot may in many instances, if the rain- 

 fall for the corresponding period of a sheltered spot in the 

 same neighbourhood, say a few miles distant, be known, 

 be determined by means of an equation involving as known 

 data (a) the relative amounts of discrepancy in the records 

 of the gauges at the two places during periods of rain and 

 of snow ; (6) an empirically determined constant K. 



L. C. W. BONACINA. 



RECENT DEVELOPMENTS IN THE THEORY 



OF MIMICRY.' 

 'T'HE remarkable resemblances that exist between certain 

 insects belonging to widely different orders have long, 

 been known to naturalists. Wasps and hornets are 

 imitated by the " clear-wing " moths, the resemblance 

 being so close that it has sometimes deceived for the 

 moment a skilled entomologist. Certain two-winged flies, 

 that inhabit the nests of humble-bees are scarcely to be 

 distinguished from their hosts, and the handsome 

 Xylocopas, or carpenter-bees, familiar objects in the 

 tropics, are deceptively copied by two-winged flies found 

 in the same regions. 



But it is not only the bees and wasps that are so 

 imitated, nor are the imitating insects to be found only in 

 the ranks of moths and flies. An ichneumon fly in 

 Borneo, belonging to the same order as the bees and 

 wasps, though not in the same sense a stinging insect, 

 is closely copied by a Reduviid bug. 



Other instances are numerous. So long ago as the 

 year 1836, the French entomologist Boisduval directed 

 attention to the extraordinary resemblance that exists^ 

 between certain butterflies which are not at all closely 

 related to each other, belonging, indeed, to groups which 

 are widely distinct. One of these butterflies is a member 

 of the Danainse, a group of which we have no resident 

 representative in this country ; a second is nearly related 

 to our familiar " swallow-tail " of the Cambridgeshire 

 fens ; while the third is a Nymphaline, not far removed 

 from our British "White Admiral." The structural 

 differences between these butterflies show the want of 

 real affinity between them in spite of their superficial 

 resemblance. The "cell," for example, of the hindwing 

 is open in the Nymphaline, while in the other two it is 

 closed by a transverse vein. This illustrates the point 

 that these resemblances affect only obvious characters ; they 

 are independent of affinity or blood relationship, and leave 

 untouched such morphological features as do not readily 

 meet the eye. 



An insect thus resembled by another is spoken of as 

 its "model," the imitating insect is called a "mimic," 

 and the combination of model and mimic or mimics is 

 known as a " mimetic pair " or " mimetic assemblage,"' 

 as the case may be. 



What is the meaning of these resemblances? Many of 

 them were well known to the older naturalists, who, 

 however, had nothing to offer by way of explanation but 

 vague talk about " repetition " and " analogy " in nature. 

 The well-known entomologists Kirbv and Spence got so 

 far as to suggest that in some cases the resemblance 

 might be of advantage to the mimic, but in their day 

 it was not likely that the subject should be treated from 

 the evolutionary point of view, and the first really scientific 

 explanation of the matter was given by Bates on his 

 return from his famous visit to the Amazon, now nearly 

 fiftv years ago. 



Bates had observed that in these cases of deceptive 

 resemblance between butterflies, one member of the pair 

 or of the group was often characterised by abundance of 

 individuals, while the whole group was marked bv slow- 

 ness of flight, conspicuousness of appearance, and 

 immunity from the attacks of insect-eating birds. On 

 these grounds he put forward the suggestion that the 

 mimicking species enjoyed protection from attack by their 



' An evening discourse Heliveri-d at the Leicester meeting of the British ■ 

 Association on August 5 by Dr. F. A. Dixey. 



