SCIENCE. 



21 



with the absolute size of the bird, some flight-hindering 

 element not yet therein contained, increases. We might 

 therefore put the question, whether equally rapid, and 

 (comparatively) equally great contraction in a small bird. 

 In fact, too, it is chiefly the larger birds that present the 

 phenomenon of soaring, a condition in which, the body 

 being maintained at the same height for a certain time, 

 muscular work is saved by special arrangements. It 

 soaring be an advantage, it must, in alternation with 

 periods of active rise by means of rudder-like mechan- 

 ism, be extensively uiilized for the problem of a flying 

 machine. 



COLOR RELATIONS OF METALS. 



In a paper on the color relations of copper, nickel, 

 cobalt, iron, manganese, and chromium, lately read be- 

 fore the Chemical Society, Mr. T. Bayley records some 

 remarkable relations between solutions of these metals. 

 It appears that iron, cobalt, and copper form a natural 

 color group, for if solutions of their sulphates are mixed 

 together in the proportions of 20 parts of copper, 7 of 

 iron, and 6 of cobalt, the resulting liquid is free from 

 color, but is gray, and partially opaque. It follows from 

 this that a mixture of any two of these elements is com- 

 plementary to the third, if the above proponions are main- 

 tained. Thus a solution of cobalt (pink) is complement- 

 ary to a mixture of iron and copper (bluish green) ; a 

 solution of iron (yellow) to a mixture of copper and 

 cobalt (violet) ; and a solution of copper (blue) to a mix- 

 ture of iron and cobalt (red). But, as Mr. Bayley shows, 

 a solution of copper is exactly complementary to the red 

 reflection from copper, and a polished plate of this metal, 

 viewed through a solution of copper salt of a certain 

 thickness, is silver-white. As a further consequence, it 

 follows that a mixture of iron (7 parts) and cobalt (6 

 pans) is identical in color with a plate of copper. The 

 resemblance is so striking that a silver or platinum vessel 

 covered to the proper depth with such a solution is indis- 

 tinguishable from copper. 



There is a curious fact regarding Dickel also worthy of 

 attention. This metal forms solutions, which can be ex- 

 actly simulated by a mixture of iron and copper solu- 

 tions ; but this mixture contains more iron than that 

 which is complementary to cobalt. Nickel solutions are 

 almost complementary to cobalt solutions ; but they 

 transmit an excess of very yellow light. Now, the ato- 

 mic weight ot nickel is nearly the mean of the atomic 

 weight of iron and copper ; but it is a little lower, that 

 is, nearer to iron. There is thus a perfect analogy be- 

 tween the atomic weights and the color properties in this 

 case. This analogy is even more general, for Mr. Bayley 

 states that in the case of iron, cobalt, and copper, the 

 mean wave length of the light absorbed is proportional 

 to the atomic weight. The specific chromatic power of 

 the metals varies, being least for copper. The specific 

 chromatic power increases with the affinity of the metal 

 for oxygen. Chromium forms three kinds of saltjS. Pink 

 salts, identical in color with the cobalt salts ; blue salts, 

 identical in color with copper salts ; and green salts, 

 complementary to the red salts. 



Manganese, in like manner, forms more than one kind 

 of salt. The red salts of manganese are identical in 

 color with the cobalt salts, and with the red chromium 

 salts. The salts of chromium and manganese, according 

 to the author, are with difficulty attainable in a state of 

 chromatic purity. He thinks these properties of the 

 metals lead up to some very interesting considerations. 



Fire and Water-Proof Paper. — A mixture is made of 

 two-thirds ordinary paper pulp, and one-third asbestos. 

 The whole is then steeped in a solution of common salt and 

 alum, and after being made into paper is coated with 

 an alcoholic solution of shellac. 



DETECTION OF STARCH-SUGAR MECHANI- 

 CALLY MIXED WITH COMMERCIAL CANE- 

 SUGAR * 



By P. Casamajor. 



In a previous communication on the same subject, t read 

 before the American Chemical Society at ihe meeting of 

 March, 1880, I gave several processes for the detection of 

 starch-sugar in commercial sugars. One of these con- 

 sisted in adding to the suspected sugar a quantity of cold 

 water, scmewhat less than its own weight, and stirring 

 the mixture for a few seconds. If staich-sugar is present, 

 it will be seen in the shape of white chalky specks. 



Quite lately a sample of yellow refined sugar was given 

 to me which was supposed to be adulterated by being 

 mixed with starch glucose. By applying the test just men- 

 tioned, there seemed to be left a tew small chalky specks, 

 which dissolved after standing a minute or two, making it 

 very uncertain whether any starch glucose was present. 

 Upon repeatedly trying the same test the result was al- 

 ways doubtful. 



I was then lead to treat the suspected sugar by a liquid 

 capable of dissolving sugar, but without any solvent action 

 on starch-glucose. After many trials, I found that methy- 

 lic alcohol of such density as to mark 50° by Gay-Lussac's 

 alcohometer answered the purpose very well, if previously 

 saturated with starch-sugar, as this solution dissolves cane- 

 sugar, either white or yellow, very readily, but does not 

 dissolve starch-glucose. 



Methylic alcohol at 50 , saturated with starch-sugar, 

 gives a solution of specific gravity = 1.25. 100 c.c. of 

 methylic alcohol at 50° dissolves 57 grms. of dry starch- 

 sugar, the volume of the solution being 133 c.c. A solu- 

 tion of starch-sugar in ethylic alcohol does not answer so 

 well, because ethylic alcohol does not dissolve so readily 

 the gummy matters found in soft sugars, which are those 

 generally chosen for adulteration with glucose. 



To test the presence of starch-sugar in a commerc'al 

 cane-sugar, the suspected sugar should, in the first place, 

 be thoroughly dried, as otherwise any water present will 

 weaken the alcohol, and enable it to dissolve more starch- 

 sugar. It should then be stirred for about two minutes 

 with the saturated solution of starch-sugar in methylic 

 alcohol. After this, the residue is allowed to settle, and 

 the clear solution poured off. The residue may then be 

 washed with a fresh quantity of the same solution. After 

 stirring again and allowing the residue to settle, there 

 will remain, if any starch-sugar is present, a certain quan- 

 tity of chalky white specks, accompanied by a fine deposit, 

 formed by the starch-sugar present in power of fine grains. 

 These finer pat tides are never seen when water is used 

 for detecting the presence of starch-sugar, as they dissolve 

 in water very readily. It seems probable that by using 

 this solution of starch-sugar in weak methylic alcohol, the 

 starch-sugar in an adulterated sample could be estimated 

 quantitatively by a process somewhat analogous to that 

 of Payen for estimating cane-sugar. 



Not having had any occasion for such a process I have 

 not determined experimentally the degree of approxima- 

 tion obtainable in th.s way. 



The methylic solu'ion of starch-sugar should be poured 

 on a filter, after it has dissolved all it can from a commer- 

 cial sugar, and the residue should be washed out with the 

 same solution, and everything poured on a weighed filter. 

 After all the liquid has run off, the filter and the residue 

 may be rapidly washed with the strongest methylic alcohol 

 fuund in commerce, which tests 92^° by Guy-Lussac's 

 alcohometer, and which dissolves starch-sugar with great 

 difficulty. 



By a dexterous use of this process it seems probable 

 that very approximate results may be obtained, although 

 what is said here is merely in the nature of a suggestion 

 to those who may have use for quantitative results. 



*A paper read before the American Chemical Society, Nov. 4, 1880. 

 t Chemical News, vol. xli., p. 221 ; Journal 0/ the American Chemical 

 Society, vol. ii., p. 111 ; Sugar-Cane, vol. xii., p. 283. 



