120 



CHEMISTRY. 



of 200 ('. I :!!>,' F.) till the disengagement o f car . 

 buretted hydrogen ceases, even when water or vapor 

 is added. The mass soon solidifies. It is treated 

 several times again with water or vapor at 200 C. 

 till it becomes clear. The mass then contains from 

 .(:; to .").> per cent, of oxalic acid, or 140 parts of 

 oxalic acid tor 100 parts of sawdust. The oxalate 

 of soda which is obtained by the process is pure 

 white. Adding sulphuric acid to it, we obtain 

 oxalic acid, which separates by crystallization. 

 The soda lye can be utili/.ed directly after concen- 

 tration for another operation. 



For the detection of small proportions of acety- 

 lene in the air Prof. Frank Clowes uses a small 

 hydrogen flame set at such height as may be neces- 

 sary, which shows a pale but well-defined "cap" in 

 air containing any proportion of the gas less than 

 the lowest explosive proportion. When the hydro- 

 gen (lame is exposed to the air to be tested for ace- 

 tylene in a darkened space it is at once tinged 

 yellowish green. In determining the limits of ex- 

 plosibility when acetylene is mixed in gradually 

 increasing proportion with air and kindled it was 

 found that air must contain at least 3 per cent, of 

 acetylene before it can be fired by a flame and the 

 mixture caused to burn throughout. As the pro- 

 portion of acetylene is increased, the explosive 

 character is augmented; when 22 per cent, of acety- 

 lene is present, carbon begins to separate during 

 the burning. The amount of carbon that separates 

 increases until the explosive character of the mix- 

 ture disappears. This point is reached when 83 pet- 

 cent, of acetylene is present in the air. The limit- 

 ing percentages in air which are explosible of the 

 several gases named are as follow : Acetylene, 3 to 

 82 ; hydrogen, 5 to 72 ; carbon monoxide, 13 to 75 ; 

 ethylene, 4 to 22 ; methane, 5 to 13. Acetylene thus 

 gives a wider range of explosive proportions than 

 any other of these gases. Prof. Clowes also recom- 

 mends the flame-cap test for carbonic oxide as 

 being at once quick of execution, sufficiently deli- 

 cate, and wide in its range of indications. The 

 only drawback to the method is that all combus- 

 tible gases give flame caps, and these are indistin- 

 guishable from that furnished by carbonic oxide. 

 Hence the flame-cap test is suitable only when other 

 combustible bodies are known to be absent. 



Boron is now made commercially for use in the 

 preparation of boronized graphite for electric bat- 

 teries. In the process as described by H. N. War- 

 ren, boracic acid, after calcination to deprive it of 

 its water of crystallization and thus convert it into 

 b >n>n trioxide, is ground to a fine powder and inti- 

 mately mixed with a suitable proportion of magne- 

 sium sodium chloride ; the mixture is rapidly heat- 

 ed to fusion, and metallic sodium is introduced from 

 time to time in large pieces, the mass being well 

 stirred after the introduction of each piece. By a 

 moderate reaction, but with the evolution of intense 

 heat, metallic magnesium is set free and attacks the 

 boron trioxide, with formation of magnesia and free 

 boron. The mass is thrown into an excess of hy- 

 drochloric acid, which dissolves out all except the 

 b iron. The boron procured by this process is pe- 

 culiarly active; and 'In- carbons formed with it 

 have been found ver.v flVctive, and are in active 

 demand. 



A> a nitrogen absorbent for the liberation of 

 ar_'on. 11. X. \Varreii suggests that either quick- 

 lime or barium be. saturated with a strong solution 

 of lithia, ignited, and mixed \\jth a sufficiency of 

 magnesium powder, and the mixture be. allowed to 

 ice in an atmosphere of hydrogen at as low a 



. The resulting mass < - 



tains metallic lithium in a n extremely divided 

 stat, small quantities of barium or cal- 



cium he [.ower of absorbing nitrogen with 



the utmost facility, so that the mixture often be- 

 comes incandescent during the reaction. 



Atomic Weights. Additional light is thrown 

 upon the constitution of tellurium by the rcdeter- 

 mination by Masumi Chikashige of the atomic weight 

 of specimens derived from Japanese minerals. The 

 determination was made by means of tetrabromide, 

 by the advice of Dr. Divers, in order to ascer- 

 tain whether this metal, found under mineralogical 

 conditions quite unlike those pertaining to the tel- 

 lurium employed by others who have investigated 

 the subject, has the same atomic weight as the 

 other. Brauner's method was closely followed, and 

 the results of the three experiments 127'57, 127'61, 

 and 127-58 agree with his in making the atomic 

 weight 127'6. The tellurium of previous experi- 

 menters occurred in union with bismuth, gold, sil- 

 ver, etc., while that employed in this research was 

 obtained from the red native sulphur, or telluro- 

 sulphur, of Japan. It is extremely improbable that 

 if the substance known as tellurium is compound, 

 as Brauner supposed, its composition should be 

 identical when occurring in association with sul- 

 phur as a sulphurlike body in Japan and when 

 occurring in metallic combination in Europe and 

 America. As the result of these and of other 

 experiments in which the conditions were dissimi- 

 lar has been constantly the same, the point may 

 be regarded as settled that its atomic weight ex- 

 ceeds that of iodine. Its occurrence with sulphur 

 and selenium in the Japanese mineral at the same 

 time furnishes additional proof that it belongs to 

 the sulphur group. Dr. Divers adds to Mr. Chika- 

 shige's paper a note implying that the case of tellu- 

 rium and iodine with atomic weights in the reverse 

 order of their places in the periodic scale is not the 

 only one. Cobalt belongs undoubtedly to the sec- 

 ond division of Group VIII, and nickel to the third 

 division, according to both Mendeleeff and Lot liar 

 Meyer ; yet all the elaborate work done on the sub- 

 ject has left cobalt with an atomic weight slightly 

 higher than that of nickel, or at least equal to it. 



The name "twin elements" is applied by Richard 

 Lorenz to simple bodies, whose atomic weights ap- 

 proach each other very closely. The properties of 

 such elements display manifold mutual relations. As 

 a type of such twins, cobalt and nickel are men- 

 tioned. Their atomic weights are nearly equal, 

 their chemical behavior and occurrence show great 

 similarity, and their separation presents difficulties. 

 But while this couple have often been viewed as in- 

 timately connected, less attention has been paid to 

 the, fact that many other elements also form twin 

 pairs in a similar manner the differences between 

 their atomic weights being not greater than 1-4 

 units, and in many cases smaller than 1. The au- 

 thor's list of "twins" includes boron and carbon, 

 sodium and magnesium, aluminium and silicon, 

 phosphorus and sulphur, potassium and calcium, 

 vanadium and chromium, manganese and iron, 

 nickel and cobalt, selenium and bromine, palladium 

 and silver, tin and antimony, iodine and tellurium, 

 tantalum and tungsten, and lead and bismuth. The 

 author further finds that in his scheme the atomic 

 weights of each pair of twins differs from those of 

 the foregoing and succeeding pairs (in round num- 

 bers) by 4 or a .multiple of 4, and that the single 

 elements situate between the pairs of twins are 

 found in the places required by the rule. This 

 holds good for almost all the known elements, and 

 most decidedly when the atomic weights are most 

 certainly known, and does not hold good (with few 

 exceptions) when the atomic weights are less accu- 

 rately known. 



It has already been shown, remarks M. Carey Lea, 

 in a paper on numerical relations existing between 

 the atomic weights of the elements, that elements 



