Outerbridge.] ^^^ [May 15, 



blighter aud longer in the finer. But on compai-ing the alloys 876.5 and 

 883.5, (reducing the variation to seven thousandths) I was both surprised 

 and disappointed to find the visible difference of result but slightly appre- 

 ciable. And the same with regard to the alloys 883.5 and 888.3, and the 

 same with other alloys with equal or less comparative variation of fine- 

 ness. A variation of one-thousandth, required an effort of the imagina- 

 tion as well as of the eye to detect any difference whatever. And, al- 

 though I endeavored to map an apj^arent difference between alloys vary- 

 ing two-thousandths, it would certainly not have been a safe test on which 

 to base an assay. Frequent repetitions with changes of adjustment were 

 tried, the battery power varying from one to six Bunsen cells, in connec- 

 tion with Leyden-jars varying from one very small jar (improvised out of 

 a test-tube) to fifty large jars, (representing a metallic superficies 

 of many square feet) with variations of the distance of the electrodes 

 apart, and with and without the use of a condensing lens, but all these 

 failed to give closer results. 



It is true, that these changes of conditions produced certain variations 

 in the effects observed — as, for instance, it was noticed that an increase 

 in the Leyden jar surface always lengthened the lines — the distance be- 

 tween the electrodes and all other conditions remaining the same — while 

 a decrease in the condensing surface had an opposite effect. Thus, 

 to take the extreme cases, with the single small Leyden jar above referred 

 to, and one cell of battery, the lines broke when the electrodes were not 

 more than j\ of an inch apart, and disappeared entirely on separating 

 the points ^ of an inch. 



With fifty Leyden jars and six cells of battery, it was found impossible 

 to break the lines at all, even by removing the electrodes to the extreme 

 limit of the spark, and in this case new lines also appeared. 



Other variations occurred ; such as a momentary ii-regularity in the 

 length and brightness of the lines, under a strong battery power, owing 

 to the unequal action of the spark ; — a difference in the action of the gold 

 lines deiiendent upon the nature of the alloy, silver tending to lengthen 

 them more than an equal admixture of copper ; — the length of the lines 

 is also dependent upon the distance between the spark and the slit (when 

 the latter is used without the intervening condensing lens) ; — moreover, 

 the eye itself is liable to become confused by continued comparisons of 

 very slight differences. The above and other modifications, so far from 

 solving the problem of close work, rather indicated possible aoui-ces of 

 error. 



Another element of the process suggested itself to me as likely to ren- 

 der the results uncertain for the practical purpose of assaying, vis : 

 whether the quantity of metal vaporized and giving the spectrum is not 

 too infinitesimal to give safe results for a large melt. This would be af- 

 fected by the least v/ant of homogeneity in the metal. This is a serious 

 consideration, and with the view partly to search for unknown sources of 

 error and partly to ascertain generally the quantity of metal operated on 



