NOVEMBEK 3, 1922] 



SCIENCE 



513 



eream, and we have stated very plainly that 

 there is little or no error if the glymol is added 

 slowly, with the tip of the pipette resting 

 against the neck of the bottle near the butter 

 fat layer but (and this is the important point) 

 that by the manner in which the glymol is like- 

 ly to be added by the average dairy helper 

 there is nearly always an error in the direction 

 of low readings. 



That this conclusion was entirely justified 

 will, we believe, be apparent when we state that 

 the results quoted in the original paper were 

 obtained in Professor Gregory's own labora- 

 tory, both of the authors watching the tests as 

 they were made in a purely routine manner, by 

 the regular tester and upon creams as they 

 were received in the laboratory. We made no 

 suggestions to the tester, merely asking per- 

 mission to take readings before and after each 

 addition of glymol. Our observations were 

 confirmed by the tes.ter at the time they were 

 made. 



That this is not to be taken as a criticism of 

 the work of the tester, or as a charge of care- 

 lessness or wilful negligence, or of lax admini- 

 stration of any laws bearing upon the subject, 

 should be evident when we state that it was 

 only after a considerable amount of subsequent 

 experimentation that we atumbled upon the ex- 

 planation of the error, which is to be found in 

 the fact that the tester usually follows the very 

 natural method of placing the tip of the pipette 

 just inside the top of the bottle neck, then al- 

 lowing the oil to flow at the full delivery speed 

 of the pipette into (the bottle. This cause of 

 error was discovered in March, 1922, and it 

 seems obvious that Professor Gregory's in- 

 spectors could not have found and checked the 

 incorrect method for 1,800 or more testers dur- 

 ing the past year, when neither they nor we 

 nor (so far as is known) any one else knew 

 that the method was incorrect. 



We regret very much that our statement re- 

 garding the importance of the error in Indiana 

 was interpreted as a charge that Indiana 

 creameries "are beating our producers out of 

 $20,000 worth of cream per year." The case 

 of ten Indiana creameries was cited merely be- 

 cause we happened to have approximate figures 

 on production for 1917 and no charge of this 



kind was made or even remotely intended. 

 However, because of possible similar misin- 

 terpretation by others, we shall gladly delete 

 this paragraph from the paper, as it adds noth- 

 ing to the scientific value of the latter. 



One further misquoitation should be noted. 

 In the discussion of the use of various non- 

 miscible oils for this purpose we mentioned 

 the trial of several of these, citing the work 

 of Eckles on fat-saturated amyl alcohol. Of 

 all of the liquids tried, we found that amyl 

 alcohol so prepared was the only one that did 

 not cause a change in readings and we so stated 

 but, far from recommending the use of this 

 liquid, we stated various objections to it and 

 finally concluded that the use of all such fluids 

 should be abandoned. Professor Gregory has 

 cited an experiment in which amyl alcohol 

 lowered the fat reading after standing ten 

 minutes but in this experiment, according to 

 his own oral statement, he omitted the very 

 important detail of first saturating the alcohol 

 with butter fat. The conclusion seems fairly 

 obvious. The advance abstract of our paper 

 did not mention this point, but, as already 

 noted, an abstract can not go into experimental 

 details. 



We understand that quite lately the labora- 

 tory in which our observations were made has 

 begun the use of a much lighter hydrocarbon 

 oil than the one formerly used. We do not 

 know whether this has diminished the error but 

 we think it quite likely. If so, we have here 

 another variable, in the variation in specific 

 gravity and viscosity of the oils in general use 

 for the pui'pose, where limits are not speci- 

 fied for these properties. 



In conclusion, we do .not believe that this, 

 or any other scientific question, can be settled 

 by denials or display of feeling, but that the 

 laboratory is the only place where a decision 

 can be reached and we offer, as the simple 

 solution of the matter, that each one who is 

 interested shall try the 'two methods of adding 

 glymol, making observations for himself. We 

 have no doubt that evidence will be found both 

 for and against our conclusions because the 

 essential defect of the glymol method lies in its 

 variable possibility of error. But if our state- 

 ments are "incorrect truths," polemics will not 



