IOWA ACADEMY OF SCIENCES. 73 



COMPOSITE MTLK-SA.MPLES IN THE LABORATORY. 



BY G. E. PARTICK. 



Composite milk- samples for use at creameries, as a means of savins: labor in the 

 valuing of milk by any of the "oil tests," I first proposed (in detail) in Bulletin 

 No. 9, of the Iowa Experiment Station, May 1890. The preserving agent there rec- 

 ommended for preserving the samples was corrosive sublimate, HgCl 2, numer- 

 ous experiments having shown that it preserves the mechanical, as well as the 

 chemical, condition of milk better than any other common antiseptic. For use in 

 creameries I insisted that the sublimate have mixed with it some suitable aniline 

 color, as a guard against accidental poisoning; and to hasten solution in the milk, 

 admixture of common salt, NaCl, was recommended. , 



For six months past I have employed the same principle m the laboratory, in 

 analyzing the milk of experimental cows, not only for fat (by one of the "oil 

 tests") but also for solids, gravi metrically. (See Iowa Station Bulletin No. 13, 

 page 29, May, 1891.) 



For this purpose the preservative is of course used without admixture of aniline 

 color in common salt, as these would bring error in the results on solids. 



The corrosive sublimate is powdered finely and passed through a very fine 

 gauze sieve. Only a very small amount is needed to preserve milk-samples five or 

 six days without material change; and five days is as long as such keeping is 

 desirable in most experiments on milk production. For keeping five days, .125 

 gm. of the HgCl-i is sufficient in cool weather, and .200 gm. in summer weather, 

 provided the daily samples are 50 c.c. each, making the complete composite 

 sample 250 c c. The theoretical error thus introduced in the result of solids is 

 only .05 per cent with the smaller amount, and .08 per cent with the larger; the 

 former figure is within the " limits of error " in ordinary routine work, and the 

 latter nearly so if not quite. Many comparative trials have, however, convinced 

 me that there is a ten/ slight loss in the solids of milk preserved for five or six 

 days, but that it rarely exceeds .05 per cent; therefore it is my custom to neglect 

 correction for the HgCl-,' when it amounts to only .05 per cent; and when it 

 amounts to .08 per cent to correct by deducting .03 per cent. These corrections 

 are accurate enough for use in routine work, by the method of drying in air on 

 fine asbestos in open watch-glasses — the method which 1 h^ve thus far employed; 

 doubtless finer work could have been done, and perhaps more accurate corrections 

 found, by the method of drying in hydrogen, had time permitted the employment 

 of this method. 



The following test determinations were made by Mr. E. X. Eaton, assistant 

 chemist in this station. 



