September 2$, 1883.] 



SCIENCE. 



445 



function, and sJolves several of tbe moie fundanieii- 

 tal problems connected with Ibe determination under 

 various conditions of tbe variation of the integral. In 

 particular, he shows how the problem of relative maxi- 

 ma or minima can be conducted to that of absolute 

 raa.xin)a or minima. — ( A'our. ann. math., Feb.) t. c. 



[266 

 PHYSICS. 

 (P/totoffraphi/.) 



Photographing Reichenbach's flames. — Tbe 

 question of the actual existence of these flames, sur- 

 rounding the poles of powerful magnets, has again 

 been brought up for discussion in scientific circles. 

 Numerous persons have claimed to be able to see 

 them, and some even to be able to distinguish be- 

 tween the poles and the color of the flames. Eei- 

 chenb.ich himself attempted to photograph them by 

 the daguerrotype process, but was apparently dis- 

 satisfied with the results he obtained. Jlr. William 

 Brooks has taken the matter up. and thinks he has 

 obtained actual impressions of the flames, by means 

 of photography, on sensitive dry plates prepared 

 especially for the purpose. In total darkness a per- 

 forated blackened card was placed one-eighth of an 

 inch above the poles of a permanent horseshoe- 

 magnet, and a sensitive plate placed an eighth of an 

 inch above the card. With five minutes exposure he 

 obtained a result; and this was repeated many times, 

 the most remarkable thing being, that sometimes 

 he obtained a positive and sometimes a negative 

 image, underprecisely the same conditions. Another 

 curious effect obtained was, that some printed matter, 

 which was under the w-ash of Indian ink used to 

 blacken the card, was perfectly readable wlien the 

 plate was developed. This latter result, however, 

 was obtained on only one occasion. He also suc- 

 ceeded in obtaining prints through a glass plate on 

 which were painted figures in black varnish. This 

 was contrary to the experience of Reichenbach, 

 who considered that the rays were not transmitted 

 through glass. — w. H. i>. [267 



Hydrokinone. — Of this new developer, first in- 

 troducetl by Capt. Abney, Mr. Charles Ehrmann says, 

 " Tbe best results I have obtained with ten grains 

 of hydrokinone to eight ounces of water, and caustic 

 ammonia (1 to ") added gradually as the develop- 

 ment progressed. The negatives are of a non-actinic 

 color, similar in tone to one slightly intensified with 

 uranium and prussiate of potash; therefore the devel- 

 opment need not be carried on very far, thus preserv- 

 ing all finer modulations. An injudicious amount 

 of alkali will produce green fog," — {Phot, times, 

 July.) 



Of this same developer, Mr. Edwin Banks claims 

 that it is much more powerful than pyro, and that it 

 will bring out a fully developed picture with at least 

 half the exposure that is necessary when pyro is em- 

 ployed. At first sight this seems strange, when it is 

 observed how much more powerfully the latter absorbs 

 oxygen; but the explanation probably lies in the fact 

 that hydrokinone is more gradual in its action, and 

 has a greater selective power, than pyro. With a 

 collodio-bromide film, for instance, which is not so 



much protected from chemical action as one of gela- 

 tine, pyrogallic acts with such energy when mixed 

 with an alkali, that the whole film is reduced imme- 

 diately, and no image, or only a faint one enveloped 

 in fog, appears : hence a powerful restrainer must be 

 used to keep this action within bounds, A soluble 

 bromide, which is the salt commonly used, has this 

 effect, but, unfortunately, at the same time partially 

 undoes the work which the light has done, rendering 

 it necessary to give a longer exposure. But with hy- 

 drokinone no restrainer is necessary, unless a great 

 error in exposure has been made. It does its work 

 rapidly and cleanly, in this respect resembling ferrous 

 oxalate. It does not discolor during development so 

 much as pyro, and consequently does not stain the film 

 so much, whilst full printing vigor is very easily 

 obtained without having to resort to intensification. 

 The color and general appearance of the negative are 

 more like those of a wet plate, since the shadows 

 remain quite clear, and free from fog. It seems 

 almost impossible to fog a plate with it. One grain 

 of hydrokinone to the ounce is strong enough for 

 most purposes. With some samples of hard gelatine 

 it is advisable to use two; but with most kinds and 

 with collodion, one grain is suflicient. Two or three 

 drops of a saturated solution of washing-soda to the 

 ounce of the hydrokinone solution rapidly develops 

 the image, and the addition of a few drops more to 

 complete development is all that is needed. A 

 soluble bromide acts very powerfully as a retarder 

 and restrainer. With a mere trace added, develop- 

 ment is very much slower. — {Brit, journ. phot., 

 .July C.) w. H. 1". [268 



ENGINEERING. 

 Sotirces of error in spirit-levelling. — Precise 

 levelling in this country has been done by the U. S. 

 lake survey, which has determined the elevation of 

 all the great lakes with a probable error of less than 

 one foot; by the coast and geodetic survey, which is 

 carrying a line of levels across the continent from 

 Chesapeake Bay to San Francisco ; and by the Mis- 

 sissippi Kiver commission, which lias a line from the 

 Gulf as far north as central Iowa, to be connected with 

 Lake Michigan, and thence with the sea-level at New 

 York. Mr. J. B. Johnson has been connected with 

 some nine hundred miles of this work, and discusses 

 tbe sources of error. He first classifies errors into com- 

 pensating and cumulative. Then he treats them as, 

 1°, errors of observation, in the instrument or in the 

 rod; 2°, errors from instrumental adjustment; 3°, 

 errors from unstable supports ; 4°, atmospheric errors, 

 from wind, from tremulousncss of the air caused by 

 difference of temperature, and from variable refrac- 

 tion. He concludes, that, with good instruments and 

 pnipercare, thirty miles of line should be' duplicated a 

 month with one Y-level and a target-rod, and all 

 discrepancies brought within five-hundredths of a 

 foot into tlie square root of the distance in miles; or 

 with the U. S. precise levels and speaking-rods, read- 

 ing three horizontal wires, one instrument should 

 bring the discrepancies within two-hundredths of a 

 foot into the square root of the distance in miles. 



