July 6, 1888.] 



SCIENCE. 



1 1 



water, — I examined all of the data since 1870, to coi'rect, so far as 

 possible, all other errors of the same kind. The record of snow for 

 the winters of 1870-71 and 1871-72 were found to be given in this 

 way, and comparison with the Signal Service observations also in- 

 dicated that the reduction had been neglected in a f^w instances in 

 subsequent years. This critical e.xamination of the original obser- 

 vations has led to the construction of the accompanying table of 

 monthly totals : — 



Monthly Precipitation at Fort Leavenworth, Kan. 



cause, — I do not know ; but it is fairly manifest that the conclusions 

 based on the assumed comparability of the two scries are quite 

 worthless. Geo. E. Curtis. 



riirmingluim, Conn., June 30. 



* Observations by Mr. F. Hawn. t Observations by Signal Service. 



Important changes in the values for April, May, July, and Au- 

 gust, 1 87 1, are corrections of serious errors e.xisting in the pub- 

 lished observations, the corrected values having been furnished by 

 Major WoodhuU. For those months in which the record at the 

 fort is missing, namely, February and July, 1872, and October to 

 December, 1883, the Signal Service observations have been inserted 

 to complete the series. 



The Signal Service record has also been substituted in Novem- 

 ber. 1871, and December, 1S80, — months in which the fort record 

 is manifestly recorded improperly, but for which the correct record 

 cannot safely be inferred ; and also in December, 1S73, November, 

 1874, and February, 1S75, for portions of which the fort record of 

 snow is apparently measured carelessly, or recorded without reduc- 

 tion, but of whose error the evidence now at hand is not entirely 

 conclusive. 



Although in these several instances the fort record has been 

 completed by the use of Signal Service observations, the series still 

 remainsessentially homogeneous and comparable from lS37to 1883. 



Combining the whole series in ten-year means, we have the ma- 

 terial for ascertaining the existence of any secular change : — 



The increase of seven inches shown by the combined Fort Leav- 

 enworth and Signal Service records has largely disappeared. Ex- 

 amining, now, the average annual rainfall from 1872 to 1883 given 

 by the Signal Service record and the record at the fort, we find that 

 the former is 38. 5 inches, and the latter 33.0 inches, show'ing a dis- 

 crepancy between the two of five and a half inches. 



To what this discrepancy is due, — whether to differences in the 

 rules of observation or to an error of ten per cent in the Signal 

 Service gauge (as was the case at Providence, R.L), or to some other 



Photographs of Lightning-Flashes. 



Possibly some of your readers may be interested in the follow- 

 ing report : — 



In the month of June, 1887, a committee of the Royal Meteoro- 

 logical Society, London, issued about two hundred circulars to the 

 secretaries of photographic societies in various parts of Europe and 

 America, and also to other likely persons, requesting them to fur- 

 nish the society with photographs of lightning-flashes. 



About sixty photographs of lightning-flashes were received in 

 answer to this invitation ; and these were e.xhibited at the meeting 

 of the society in March, 1888, where they received much attention. 



From the evidence now obtained, it is evident that lightning as- 

 sumes various typical forms, under conditions which are at present 

 unknown. 



The following appear to be some of the most typical forms of 

 lightning-flashes: — 



1. Stream lightning, or a plain, broad, rather smooth streak of 

 light. Only two or three specimens of this form have been re- 

 ceived. The committee are disposed to consider this a distinct 

 type of a single stream-like character, without distinct irregularities, 

 or branches, and not merely the result of bad focusing, because 

 other objects, such as trees, are extremely sharp. 



2. Sinuous lightning, when the flash keeps in some one general! 

 direction, but the line is sinuous, bending from side to side in aver)', 

 irregular manner. This is by far the commonest type. 



It is very noticeable that the thickness of the line varies during 

 the course of discharge. Sometimes the thinnest part of the white 

 streak is the highest, and the flash appears to get thicker as it ap- 

 proaches the earth ; at other times a flash in the air begins thin,, 

 broadens out in the middle, and fines away again at the farther 

 extremity. 



The committee can offer no explanation of this at present, but 

 they would call attention to the fact, that, in some photographs of 

 electric sparks taken from an induction-coil, those of high tension 

 are thinner than those of low tension. 



3. Ramified lightning, in which part of the flash appears to 

 branch off from the main streak like the fibres from the root of a tree. 

 Of course, there is no evidence as to whether these fibres branch 

 off from, or run into, the main flash. 



4. Meandering lightning. Sometimes the flash appears to mean- 

 der about in the air without any definite course, and forms small, 

 irregular loops. The thickness of the same flash may vary consid- 

 erably in different parts of the course, as mentioned above; and a 

 flash may go pretty straight in one portion of its path, but meander 

 considerably in another. 



5. Beaded or chapleted lightning. Sometimes a series of bright 

 beads appear in the general white streak of lightning on the photo- 

 graphic plate. Occasionally these brighter spots appear to coincide 

 with bends in a meandering type, but often the beads appear 

 without any evident looping of the flash. 



But as a flash is moving in space, while two directions only can 

 be shown on the plane of the paper, there is every reason to believe 

 that the brighter spots on the positive picture may be points where 

 the flash was zigzagging, either directly towards, or away from, the 

 observer, and thereby giving a somewhat longer exposure to these 

 spots. 



6. Ribbon lightning. Nearly one-sixth of the photographs received 

 by the society show flashes exhibiting more or less of a ribbon-like 

 form. One edge of the ribbon is usually much whiter and firmer 

 than the other. 



Occasionally in the same picture some flashes appear normal, and 

 others ribboned ; but the flashes in a picture need not have oc- 

 curred simultaneously. The committee have not yet in their pos- 

 session any conclusive evidence as to whether the same flash may 

 be normal in one portion, and ribboned in another portion, of its 

 course. 



In one picture there is a bright streak on the lop of the flash : 

 then about an eighth of an inch of ribbon-like light, the folds fol- 



