to 1000 m rt. of wk'' air p r m i . The initid air is a t Q pressure of 150 lbs. per 74, in. What size of stamitard steel pilye must be used in order that the drap in prewme shall nup exceed 10 Ibs. per sq. in.?

S ~ u r ~ o w - ~ o m Fig. 14,s - 0.0188 at '80' F.

From Pf. 17, .@ = 000118 a t SOo F. and Z atm. Since the aver-

*age pressme is -= 9.87 atm , s - 8.87 X 0.00118 - 0.01165.

S i n c e h e air refers ta air a t 1 atrn. and 70" F. the actual volume at SOo F.

150 + la0 2 X 14.7

((540- F. absolute) a..d 9 87 atni. pressure is 'Oo0 9 . 8 7 X 530

j40 = 103.2 cia. ft. per min.,

103*2 60 X 0.785 D2 D2

144 = 'L5 ft. per sec. a t 80' F. and a mean pressure and w =

315 D v s D X F X 0.01165 195

5 0,0188 D men -= =-. (4

Taking the equivalent length of each elbow as 30 pipe diameters,

3 X 3 0 D - 1 = 200 + -- 200 + 7.5 D. 12

0.323 f I s u2 But 9 = ,

0.323 f [ Z O O + 7.5 Dl 0.01165 X (315)* * a = D5

@) 374 f(200 + 7.5 D] 5.

DS 54. in. o =

As noted ahwe , cone may assume a value of D actually obtainable, find

from Fig 12 *be rorresponding value of '2, and calculate p . The result

of such pmcedure is sliown below. Z

D v s # - x (Fiom Dr P

Nominal Actual (From Fig. (Table (From Dia. Dia. (D) Eq. a) 12) 11) Eq. b) 1 . 5 1.61 1211 0.0041 10.82 30 .1 2 2 . 0 7 9 4 . 3 0.0043 37.7 9 . 2

1 The friction factor for this case was taken from Fig. 10, as Fig. 12 does D v s

not show f for values of - beyond 100. 5

It should be noted that although the first trial diameter (1.61 in.) called for a pressure drop of 30 lbs. per sq. in., a diameter only slightly larger (2.07 in.) gave practically the desired pressure drop. This is because p varies inversely as the j f f f h power of diameter. Hence a 2-in. standnrd steel pipe must be used to meet the specified conditions.

, ACKNOWLEDGMENT

In conclusion the writers desire to express their appre- ciation of the many helpful suggestions of Dr. W. K. Lewis, a t whose instigation this work was first undertaken; of the accurate work of Mr. V. V. Kendall, who aided in the con- duction and direction of much of the experimental work; and of that of Messrs. Kite and Kennedy, who made the runs in the turbulent and critical flow regions. Acknowledgmentp are also due to the various cooperating companies mentioned in the introduction; and to Mr. R. S. Danforth of The Kinney Manufacturing Company, and Dr. W. H. Herschel of the Bureau of Standards, for valuable suggestions.

The Determination of Carbon Dioxide in Baking Powder''z By C. S. Robinsons and Selma L. Bandemer

MICHIGAN AGRICULTURAL COLLEGE EXPERIMENT STATION, EAST LANSING, MICHIGAN

The gasometric method for the determination of carbon dioxide in baking powder is as accurate as the absorption methods. It is much more rapid and requires but a single piece of apparatus. It can be used for the determination of both total and residual carbon dioxide.

.............. At the 1920 meeting of the Associat,ion of Official Agricul-

tural Chemists the senior author suggested, as a substitute for the present official methods for determining total carbon dioxide in baking powder, a modification of the gasometric method originally devised by Van Slyke for the determina- tion of carbon dioxide in blood p l a ~ m a . ~

Its use has since been extended to the estimation of residual carbon dioxide. The object of the present paper is to de- scribe this procedure and report the results obtained with it.

PROCEDURE TOTAL CARBON DIOXIDE-The procedure for the determi-

nation of total carbon dioxide in baking powder is identical with that for the analysis of marl and limestone as given in the original description of the method, except that 100-mg. samples are used.

RESIDUAL CARBON DIOXIDE-A 2-g. sample of the well- mixed material is weighed into a small beaker and 20 cc. of distilled water are added. The mixture is thoroughly stirred, allowed to stand for 20 min. a t room temperature, and set in a bath of boiling water for the same length of time. Fi-

1 Received November 3, 1921. 2 Published as Journal Article No. 20 from the Chemical Laboratory

Published by of the Michigan Agricultural College Experiment Station. permission of the Director of the Experiment Station.

8 Research Associate in Chemistry. 4 J. Bzol. Chem., SO (1917), 347, Sod Sa., 10 (1920), 41.

nally it is boiled for 1 min. The addition of a drop or two of octyl alcohol to reduce foaming is sometimes advisable a t this point. One sample is diluted to 25 cc. in a volumetric flask, and 1-cc. samples are used for the determination.

The accompanying table shows the results obtained with this method and with the two official methods. They are typical of the agreement of results which have been obtained in this laboratory with all sorts of carbonates during the past year. COMPARISON OF OFFICIAL. AND GASOMETRIC METHODS FOR DETERMINING

Total Carbon Dzoxide CARBON DIOXIDE I N BAKING POWDER

I n IIa IIIa IVa V u (CaCOs)

Official method

Gasometric

43.34 13 .54 13.48 16 .73 14 .62 43 .81 13 .29 13 .11 1 7 . 0 3 14 .35 43 .70 13 .30 1.1.48 16.90 1 4 . 2 9 43 .78 13 .55 13 .34 17 .03 14 .52 43.70 13 .52 13 .31 16.89 1 8 . 5 2 43 .78 13 .52 13 .29 16 .92 14 .53

Residual Carbon Dioxide I b I I b IIIb IVb Vb ~~

Official method, Heidenhain 0 . 7 3 1 . 1 5 1 . 7 4 5 . 5 1 1 . 2 7 Knorr 0.85 1 . 2 6 1 . 5 1 5 . 2 6 1 . 3 9 Gasometric 0 . 7 5 1 . 1 6 1 . 6 3 5 . 4 5 1 . 3 6

0 . 7 5 1 . 3 1 1.59 5 . 6 2 1.27

In order to test the method when a product contained a. water-insoluble carbonate, Sample IVB was made up by mix- ing Sample Vb and calcium carbonate. "Duplicate results agreed as closely as did those with other products.

It has been found necessary to lubricate the stopcocks of the apparatus rather frequently with a heavy grease to prevent leakage during the maintenance of the high vacuum used. If the carbon dioxide is absorbed each time and the residual air measured instead of being calculated from the table, this source of error is avoided. With proper attention t o stopcocks, however, this is unnecessary.

Only one source of trouble has been encountered.