APPENDIX. 151 



in the leapt amount of water and 50 cc of stronger alcohol. The mixture is boiled 

 for half an hour with inverted condenser. (I have found saponifieation by this 

 method insufficient in the case of waxes, notably beeswax.) After saponification 50 

 ccofhot water are added, the, mixture is cooled, and shaken out with petroleum 

 ether. 



Page 30, 12. Commenting upon this method, I will state that while Wiley's method 

 is undoubtedly the most accurate for determination of the melting point, the capillary- 

 tube method is so much more convenient that most of the chemists in this section 

 use it in their commercial work. I would suggest that that method be reinstated as 

 an alternate method. A. /S T . Mitchell. 



Page 32, 14. The directions for the Maumene figure are doubtless designed espe- 

 cially for olive oil; but inasmuch as cotton-seed, sesame, or poppy seed and the like 

 might be included among edible oils, would it not be well to consider whether 

 " strongest" sulphuric acid can be added to the latter oils without such frothing and 

 evolution of S0 2 as would vitiate the results? Some experiments on the Maumene 

 test have recently been made in this laboratory a . The details are now being prepared 

 for publication. H. C. Sherman. 



Comment by Mr. Tolman. This suggestion regarding the method for determining 

 the Maumene figure is very appropriate and it will be necessary to modify the method 

 when working with such oils as are mentioned above. Sherman a uses an acid which 

 will give a rise of temperature with water of 33-34 C., and calculates the "specific 

 temperature reaction" as ;,'iven in the methods for oil analysis. But he finds that 

 slightly lower results are obtained by this method. Each analyst should determine 

 the constants for the conditions under which he is working. 



Page 33, 17. I should prefer to use the fatty acids for the Bechi test rather than 

 the oil or fat itself. A. G. Woodman. 



DAIRY PRODUCTS. 



Page 35, 2. The dish used should be platinum or light porcelain. I should bar 

 the use of either aluminum or tin. They are too easily subject to corrosion and 

 change of weight. A small rod, about 1 inch by 3 or 4 millimeters, should be weighed 

 with the dish, and the milk and sand should be stirred up when milk is added. 

 This will greatly facilitate drying and no crust will form on top of the sand. In this 

 manner more milk can be advantageously used, and I should advise increasing the 

 amount to an optional 5 or even 10 cc. Anyone using the rod once will never after 

 do without it. Charles L. Parson*. 



Page 36, 6. From the expression, "if the milk contains one or more parts per 

 10,000 of formaldehyde," one would gather that the test will detect any amount of 

 formaldehyde greater than one part per 10,000, whereas I believe that if the amount 

 of formaldehyde be increased above a certain point the test will no longer show it. 



.1. d'. ]\'<><><lin<tn. 



Page 36, 6. I think the test with phloroglucinol by far the easiest of use accom- 

 panied by certain identification, and should be given direct. Charles L. Parsons. 



Page 36, 6. Under the detection of preservatives, I do not think it well to give 

 the sulphuric acid test because there is such a variation between different chemists 

 as to the reliability of this method and the manner of operating it. I would suggest 

 the hydrochloric acid with ferric chlorid as more satisfactory and reliable, and its 

 use is confined to the detection of formaldehyde in milk. 



' Science," WU1, p. 30. 



