T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
40
natural t o look for a drop in the expansion. The curves obtained with a rising temperature failed t o show any drop so samples were tried using a constant temperature or, in other words, subjecting the sample t o the same condition as an ordinary cure (Fig. 4). The graphs failed t o show any drop with the exception of the cases already mentioned. Even in those cases t h e drop was very gradual and did not start a t t h e point of vulcanization. Theref ore, the method could not be used as a method of regulating or determining the cure of a sample of rubber.
FIO.5-CURVE
SHOWlNQ
EFFECT OF DIFFERENT PRESSURES
ON T W O
R U B B E R S , ONE OF WHICHSHOWS A LARGE AND THE OTHERA SMALL DIFFERENCE BETWEEN R A W AND C U R E D SPZCIFIC
GRAVITIES
Since the curves failed t o show any volume contraction during the heating process, the increase in specific gravity which always occurs must come as t h e result of the pressure and the subsequent contraction on cooling. Since rubber is nearly incompressible, the decrease must come as the result of the squeezing of air out of the raw rubber. If this be true, then a compound showing a big difference between raw and cured specific gravity will expand more against a light spring than against a strong one. This was tried out and found t o be the case (Fig. 5 ) . -EXPANSION
AGAINST
STOCK 50 lb. 1520 ............ 3.81 X 10-4 355 . . . . . . . . . . . . 2.17 x 10-4
SPRING-
100 lb. 3.81 X 1 0 - 4 1.36 x 10-4
SPBCIFIC GRAVITIES
STOCK 1520.... 355..
..
RAW 0.95 1.457
CUREDIN EXP. APPARATUS
CUREDIN
- -
HYDRAULIC PRESS 50-lb. sDrina 0.957 0.956 1.568 1.568 ,
~-
100-lb. sorina 0.956 1.577
EXPANSION V S . CONTRACTION
The increase in specific gravity shows t h a t the contraction is greater than the expansion. Therefore t h e question arises whether or not the values obtained apply t o the cured rubber. I n order t o determine this, measurements were made on samples cured in a disc mold.
....................
Aluminum mold. No. 6 Sidewall Diameter of mold, h o t . . . . . . . . . . . . . . . 3 8 1 / 8 2 in. Diameter of rubber, c o l d . . . . . . . . . . . . . . 320/s2 in.
a =
Lt - Lo -
Lo
f
0.0625
39063
a (cubical)
= 2.20
(218)
X IO-^
12,
No.
I
This compares with 2.8 X IO-^ obtained the other way. This is a check considering the fact t h a t t h e rubber was of a different sample and was from mill stock, while t h e expansiometer sample was from calendered stock. The conclusion may be drawn then t h a t the coefficient determined between 200' and 300' may be taken as sufficiently accurate for the coefficient for vulcanized rubber from the same sample. CONCLUSIONS
Although i t is not yet possible from the results obtained with this apparatus t o state definitely whether or not a stock will mold well in factory practice, i t does give us a means of distinguishing between batches of the same stock which may have different properties, due either t o rubber, milling, or other conditions. The results may be summarized as follows: I-The values of the coefficient of cubical expansion for different grades of rubber have been determined. 11-The higher the rubber content the greater the expansion. 111-The harder the crude rubber the less the expansion. IV-The more the rubber is milled the greater t h e expansion. V-There is no break in the expansion a t the point of vulcanization. VI-The increase in t h e specific gravity is caused by t h e pressure and not by physical change or internal contraction of volume.
THE MOISTURE CONTENT OF CEREALS By 0. A. Nelson and G. A. Hulett LABORATORY O F PHYSICAL
CHEMISTRY. PRINCETON UNIVERSITY, TON, N E W JERSEY
PRINCE-
Received July 29, 1919
The determination of t h e moisture content in colloidal organic substances presents peculiar difficulties, as has been pointed out in previous work on the moisture content of coa1s.l The layer of water adsorbed on the surface of colloids has quite different properties from t h a t of ordinary water. Its vapor pressure is much lower, and if this layer is of molecular dimensions it cannot be removed by the best dehydrating agents in a vacuum desiccator.2 I n order t o get an idea of the situation, consider a cubic centimeter of t h e organic materials t h a t go to make up coals, plant materials, cereals, etc., divided into little cubes 10-6 cm. on each edge. The area of the faces of these little cubes would be 6 0 0 sq. m. If these surfaces were covered with a layer of water approximately I x 10-8 cm. thick, there would be 0.06 cc. of water, or over 5 per cent of the weight of the substance, and this water would be in a condensed condition so that it would show practically no vapor pressure. If the layer were one-tenth the thickness of the assumed cubes, i t would make up 50 per cent of the weight, and still the vapor pressure would be less than t h a t of normal water, so the sub1
a = o.oooo733g
Vol.
E. Mack and G . A. Hulett, Am. J . Scz.,
174. 2 LOC.
Cil., p. 92
4S (1917), 89, and46 (19181,
Jan., 1920
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
stance would take up moisture from moist air and still feel “dry.” Such is the case of peat containing as much as 75 per cent water.’ Colloidal substances containing a large amount of moisture readily give up water t o dehydrating agents C., b u t or when heated in a n air bath t o I O ~ ’ - I I O ’ due t o the fact t h a t the vapor pressure continually drops with the removal of successive layers of water the system soon reaches such a low vapor pressure t h a t a measurable amount of water is no longer obtained, even though considerable moisture is still present. Attempts have been made t o remove the water by chemical means, e. g., by mixing the substance with calcium carbide and measuring the acetylene formed.2 This method is not applicable in all cases and has not found very much favor. The majority of methods are based upon the increasing vapor pressure of the water by raising the temperature, and depending on evaporation or dehydrating agents for its removal. But with organic substances one is limited as t o temperature on account of chemical decompositions which form water. Heating in an air bath or in an atmosphere of indifferent gas t o a temperature of 105’-110‘ C. is t h e common practice for moisture determinations, and gives a useful commercial test method, but does not give the true moisture contents. I n the case of coals, e . g., i t was found t h a t the true moisture contents were from 30 per cent t o 50 per cent greater than t h a t given by the official “moisture” method. I t was also found possible t o heat coals in a vacuum t o above 250’ C. (in many cases t o 300’ C.) without appreciable amounts of decomposition. But in the case of coals the cellular structure of the original material all disappeared in the early stages of decomposition and, also in the long process of forming coals the more decomposable matters have presumably disappeared. So we are dealing with a comparatively favorable substance. I n the case of cereals and natural organic substances which have their cellular structure and contain easily decomposable matter, the problem of determining the moisture which is present as such is obviously a far more difficult one than for coals. Generally organic substances decompose non-reversibly, giving gases, e . g., COZ, CO, CHI, HP, moisture, and a residue, and we have reason t o believe t h a t there is no definite temperature of decomposition such as a boiling point or a freezing point, but t h a t we are dealing with chemical reactions which go on a t all temperatures, but with widely different rates. A substance t h a t decomposes readily a t 200’ or 300’ C. also decomposes a t lower temperatures, but often SO slowly t h a t no measurable results are obtained in experimental time. When an organic substance with water as its external phase is heated in a vacuum, water is given off. If this water is removed as fast as i t is liberated a t a constant temperature, the rate of removal is fast a t first, then falls off rapidly and in time entirely ceases. 1
Am. . I . Sc’., 43, 89. H. C. McNeil, Bureau of Chemistry, Bulletin is2 (1912).
41
If we plot the per cent of moisture against time we get a curve like the 105’ C. curve of Fig. I . Now if the experiment is carried out a t a higher temperature, another curve is obtained lying just above the one considered. The reason for the relative positions of these curves may readily be understood from the viewpoint of the lowering of t h e vapor pressure due t o thinner and thinner layers of water as outlined above.
20
40
60 EO FIQ. 1
140
120
When the curve runs parallel t o the time axis i t means t h a t no more water is being liberated a t t h a t particular temperature. Water may still be present but with such a low vapor pressure t h a t sensible amounts do not leave the substance in experimental time, although a t a higher temperature with increased vapor pressure more water is obtained and continues t o be liberated until the vapor pressure again drops, and in time the curve runs parallel t o the time axis. When, however, we reach a temperature a t which the rate of decomposition is rapid enough t o furnish measurable amounts of water, the flat part of the curve will not be parallel t o the time axis, but will gradually rise. Such a series of curves gives us information as t o the temperature and time we may heat the substance in question without encountering measurable amounts of water arising from decomposition of the material. METHOD AND APPARATUS
The apparatus used was a modification of the one used in previous work‘ on coals and is represented diagrammatically in Fig. 2. Glass tube A was about 2 0 0 mm. long and 1 2 mm. in diameter, fitted with a ground stopper having a capillary side tube containing the stopcock s. Glass tube B was closed a t both ends and fitted inside of A for the purpose of reducing the “dead space” of the system. The boiler C contained the con2tant boiling liquid. D was an electric heater, and E was a condenser made of thin glass, about 80 mm. long, with a diameter of 7 mm. Solid carbon dioxide, F, was used for the purpose of maintaining a temperature of -78’ C. It was found 1
LOG.
cit.
4 2
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
Vol.
12,
No.
I
t
--D
FIG.2
FIG.2a
quite possible t o weigh tube A in the balance, so the sample t o be used was weighed directly. I n case the sample was a fine powder a wad of glass wool was placed on top before tube B was adjusted, as i t was found t h a t without this some of the substance would pass along with the gases when stopcock S was opened t o a vacuum. After the condenser E had been sealed in place the system was evacuated as far as the stopcock s, and the condenser warmed in order t o remove any moisture t h a t might be adsorbed on its inner surface. This water was taken up by the Pzo5, in K. After the condenser had been cooled t o --.78' C. the stopcock S was opened, and the air, adsorbed gases, and moisture passed rapidly out of A. The moisture was condensed as ice in tube E while the gases passed on and were collected in T and reserved for analysis. This modification of the original "coal apparatus" made it possible t o make determinations of the moist u r e liberated a t any stage of the experiment, since it was an easy matter t o close stopcocks S and R, cut out the condenser and seal on another, without interrupting the heating, a n operation t h a t required only I O or 1 5 min. The vapor baths which we used t o obtain constant and reproducible temperatures were: water for 100' C.;pyridine for 115' C.; chlorbenaene for 1 3 2 ' C.; xylene for 1 4 2 O C.; brombenzene for 155' C.; aniline
for 184' C.; naphthalene for 218' C.; and $-mononitrotoluene for 2 3 7 ' C. Since we were using a Topler pump and toward the end of the experiment had a very high vacuum (less t h a n 0.0001 mm. above the vapor pressure of mercury a t room temperature), i t became a question whether all the moisture was retained in E, or whether some was carried over into the Pzos tube, K. This was tested in the following ways: Two condensers, E, were sealed together in series and sealed in place in the system. The one next t o A was left empty while 0.1472 g. of water was weighed into the second. A moisture determination was then made on a sample of wheat flour, and after heating and evacuation the required length of time (4 t o 5 hrs.), the amount of water in each condenser was weighed separately. The water remaining in the second tube weighed 0.1472 g., exactly the amount t h a t had been weighed into it, thus indicating t h a t either no water had escaped or t h a t this amount was the same as t h a t which had come over from the first condenser. I n order t o make sure t h a t the second possibility was not the true one t h e following test was made: A small amount of water was weighed into a condenser, E, but instead of sealing it t o A it was guarded by a Pd& tube with a stopcock, and evacuated. Then I O or 1 5 cc. of air were allowed t o pass slowly (during 4 or 5 hrs.) through
Jan., 1920
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
the Pz05tube over the “ice” and collected in T. The weight of t h e water filled into the condenser originally was 0.1887 g. and a t the end of the experiment the water left in the tube weighed 0.1886 g. This indicated t h a t all t h e water had been retained in E. Our Topler pump, fitted with a calibrated fall tube, served also as a McLeod gauge, and we have calculated t h a t the pressure in the system was often less t h a n 0.0001 mm. Water a t -78’ C. has a vapor pressure of 0.0005 mm.‘ and we feared a loss, b u t evidently the rate of evaporation a t these low pressures, 0.0005 mm. t o 0.0001 mm. or less, is so slow t h a t not a measurable amount is lost in experimental time. The method as now employed permitted us t o pump t h e moisture and gases from the samples a t any given temperature, retain the moisture in E as ice, and collect and measure the gases. Furthermore, t h e small amounts of gases liberated, on continued heating a t a given temperature, were measured in the fall tube L and it was possible t o measure the minutest traces with high precision. I t was exactly this part of the gas-time curve t h a t was important.2 As already pointed out, we are concerned with the moisture-time curve when heating a t a given temperature, b u t we have found t h a t gases are also given off with adsorbed water, and a series of gas-time curves have the same character as the moisture-time curves of Fig. I . It was a n easy matter t o follow the gases liberated with time. Furthermore, when organic substances decompose t o form water, gases are also formed and generally in more noticeable amounts t h a n water. So this method is a particularly sensitive one for determining decomposition. As mentioned above, the moisture liberated during a run was condensed as ice in E. Ice does not dissolve gases, so aside from moisture only vapors which have a pressure of the order of water a t -78’ C. would be retained. The tube E was cut out, wiped and weighed carefully, and the water removed and examined. It was then dried inside by blowing d r y air through, and again weighed, thus the total amount of water liberated during the run was directly and accurately determined. We removed and replaced E with a fresh tube a t intervals during the experiment, and thus the total amount of water was “analyzed” as t o rate of liberation. We may now take the total amount of water liberated in each run and plot against the corresponding temperatures and get moisture-temperature curves which show distinct breaks (Fig. 3). If we evacuated a sample and measured the liberation of moisture and gases with increasing temperatures, t h e results gave little information, for the liberation of moisture and gases was so slow t h a t no characteristic points appeared. I n fact ft was found necessary to maintain the temperature from 3 t o 6 hrs. t o find out all t h a t would happen a t t h a t given temperature. The moisture-temperature curves indicate t o us the amount of moisture liberated a t each temperature, and 1
Dushman, Phys. Rev., 6, 223.
of gas in the fall tube, the Tapler pump must be free from all traces of moisture, so the PsOs in tube K must be kept in first-class condition. a In measuring these small amounts
43
the break shows the point of temperature a t which deposition becomes measurable when the sample is heated for 3 hrs. or more. I t does not necessarily show the total amount of moisture t h a t is actually present in the substance but it does show the time and temperature t o which we may safely heat the substance. By extrapolating the flat part of the curves beyond the breaks we may get an idea of the total amount of moisture present as such. I n addition t o determining the moisture contents of wheat flours and corn meal we have also made moisture determinations on cornstarch, celluloses and one protein (Edestin). Since these substances are the main constituents of the greater number of cereals i t was of importance t o know not only the moisture contents of each but more so t o know the temperature a t which decomposition became rapid. Some of t h e results of our experiments are given in the tables. Vol. Gas Temp Deg. C.
SAMPLE
Corn Starch (S-1).
,
.. . . . , . .
Cellulose (Swedish Filter Paper) .., . ..
....
. . . ...
Protein (Edestin)
.. ..
. .. .... . . . ,
218 100
115 132 142 115 184 218 115 132 155 184 218 237
100 115 132 156 184 218
Per cent Water 10.80 10.87 10.98 11.01 11.12 11.41 13.37 11.34 11.37 11.70 11.93 12.12 12.17 12.32 11.80 12.09 12.16 12.27 12.30 12.32 12.46
per G. Time of Sample Sample Heated Cc. Hrs. Min. 0.007 4 00 0.017 6 20 0.030 5 00 0.077 6 20 0.125 6 00 4 00 0.400 2.762 3 20 0.000 4 40 0.042 6 45 6 00 0.172 6 15 0.231 0.402 4 00 0.421 4 30 2.188 2 00 0.000 4 40 0.001 4 00 0.017 4 00 0.031 4 45 0.124 4 00 0.685 3 00 1 00 2.423
2.63 2.67 2.71 2.72 2.93 3.61 5.49 5.55 5.57 5.63 5.74 6.11 10.40 10.63 10.82 11.08 11.67 14.06
0.001 0.037 0 IO98 0.234 0.572 1.269 0.000 0.010 0.018 0.058 0.070 0.321 0.000 0.010 0.018 0.057 0.153 1.845
3 4 4
3 2 2 2 3 3 5 3 2 2 2 2 3 3 3
30 10 20 45 00 00 00
00 00 00 00 40 00 15 00 10 45 00
DISCUSSION OF RESULTS
One point of importance to be noted about t h e moisture-temperature curves, Fig. 3, is t h a t none show any distinct breaks before a temperature of 184’ C. is reached and, with the exception of wheat flour and protein, the deviation from a straight line is very slight until 218’ C. is reached. This means t h a t with these two exceptions the water arising from decomposition while heating for over 3 hrs. in a vacuum is very slight even a t this high temperature. Proteins. and substances containing large amounts of these compounds decompose a t temperatures somewhat l&er t h a n do starches and celluloses, a thing t h a t would probably be expected if we consider the very complex structure of the molecules t h a t go t o make up these substances.
44
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
Vol.
12,
No.
I
The gas-temperature curves (Fig. 4) show very nicely a t what temperature decomposition becomes rapid. It is t o be noted t h a t there are no very sharp breaks in any of these curves, which is in accordance with the assumption t h a t there are no definite temperatures of decomposition. However, when such a temperature is reached t h a t appreciable amounts of water are liberated due t o decomposition, the gastemperature curves become relatively steep and in course of time almost parallel t o the volume axis. 3.0
29
.
-_
28 ~-
.
._._. .
2.7
I I/
FIG.3 F
r As has been pointed out above, these results do not represent the true moisture contents of the substances tested. To obtain these we have t o extrapolate the flat part of these curves t o some higher temperature. Here, however, we are confronted with the question as t o what temperature we needed in order t h a t no water may remain as adsorbed water in the substance. Obviously i t is impossible t o determine this by direct experiment, because most of the substances found in cereals decompose readily when we get much above 200' C. I n the case of coals it was found quite safe t o use temperatures as high as 300' C. without encountering appreciable amounts of water due t o decomposition, and also t h a t a t this temperature probably all the moisture was given off. Considered from another point of view, we have reason t o believe t h a t very little water can be retained in a n adsorbed and colloidal state a t this temperature. We know t h a t above 365' C. (the critical temperature of water) water cannot exist as a liquid, no matter how much pressure is applied, from which we may infer t h a t a t these high temperatures, where water is incapable of existing except as a gas, colloidal and adsorbed water should be removed much below this temperature in a high vacuum. With substances where the flat part of the curve rises rapidly with temperature it is quite safe t o assume t h a t there cannot be any adsorbed water a t 250' C. The true moisture contents of the substances tested would therefore be as follows: Substance Per cent Moisture Wheat Flour (P C. 20094). . . . . . . . . . . . . . . . . . . 1 1 . 8 Corn Meal (P. C. 20093) ...................... 12.25 Corn Starch (S-1) , . . . . . . . . . . . . 12.4 Protein (Edestin) .............................. 12.3 Cellulose (Swedish Filter Paper).. . . . . . . . . . . . . . 2 . 8 Cellulose (Absorbent Cotton). . . . . . . . . . . . . . . . . 5 . 9
................
100
120 130
140
160
180
200
220
240
FIG.4
As will be seen from the tables, the periods for which we heated t h e samples ranged from 3 t o 6 hrs. At t h e higher temperatures i t is not necessary t o continue heating much over 3 hrs., while a t temperatures between I I O O C. and 184' C. the heating should be kept u p for about 5 hrs. We found in those runs in
Jan., 1920
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
which we weighed the water liberated every hour or oftener t h a t moisture continued t o be given of€ for over 4 hrs. To give a n example: It was found t h a t on heating Swedish filter paper for I hr. a t 184‘ C., 4.79 per cent water was obtained, while on continuing t h e heating for 4 hrs. more an additional 0.11 per cent moisture was liberated. Similar results were obtained in other cases. SUMMARY
The method of determining the moisture contents of cereals and other colloidal organic substances outlined above consists essentially of heating the material in a very high vacuum for ’definite periods of time. T h e amount of moisture liberated is condensed quantitatively in a small tube surrounded by solid carbon dioxide from which i t is accurately weighed. Since the smallest amount of decomposition can be accurately determined, this method permits us t o determine t o what temperature and for what length of time a substance m a y be heated without encountering an appreciable amount of water from decomposition. Observations on t h e evolution of gases and moisture where measurable decomposition was taking place gave us a n idea of t h e rate of decomposition a t these temperatures. LABORATORY EXPERIMENTS ON THE MANUFACTURE OF CHINESE ANG-KHAK IN THE UNITED STATES By Margaret B. Church BUREAU OF CHEMISTRY, u. s. DEPARTMENT OF AGRICULTURE, WASHINGTON, D. c. Received May 29, 1919
Chinese red rice, or ang-khak (ang-quac),’ is produced b y means of a noteworthy fungus, Monascus p u r p u r e u s Went. Red rice evidently originated in one of the provinces of China, and even to-day may be procured only in certain localities of t h a t country. It is well adapted t o its special use, the coloring of food products, such as Chinese cheese, because of its property of breaking into fine particles when rubbed or brought into contact with watery solutions. The Chinese have been very secretive concerning the preparation of red rice, and the literature contains only the following facts on the subject. Ordinary rice is moistened and somehow infected or inoculated with a fungus which under humid conditions produces cottony mycelium, eventually binding the rice grains together. The rice grains, a t the end of a brief period, are thoroughly impregnated with mold hyphae which produce the red color. When ready for the market, t h e red rice, the grains of which may be readily powdered between the fingers, imparts a clear carmine color t o any moist food substances t o which i t is added. This same fungus is, according t o Went,2 responsible for the red rice of the Malay Islands, and is also employed in Formosa in making anchu, a rice drink. Monascus p u r p u r e u s has been reported from silage in America by B ~ c h a n a n . ~We have also obtained i t F. Lafar, “Handbuch der technischen Mykologie,” 2, 265-269, Jena, 1906. * A n n . Sci. Nut. Bot., 1 (1895), 1-16, plates 1 and 2. “Monascus purpureus in Silage,” Mycologia, 2 (1910), 99-106, plates 22 and 23.
*
45
repeatedly from silage where it grew in a substratum 6 in. below the surface, and also from spoiling cornmeal samples. Notwithstanding the competing organisms, Monascus p u r p u r e u s has always been successfully isolated from Chinese red cheeses, which are colored with red rice. A culture of the same or a closely related mold has also been obtained from moldy carob beans. The most interesting source of all, however, is “freckled” codfish. We have not completed experiments t o prove t h a t the “freckling” is due t o Monascus p u r p u r e u s , b u t we have observed curious spore-like bodies, which in part compose the freckles, develop into the fungus in question when transferred t o highly salted agar plates. The medium used was Czapek solution agar, slightly acidulated with normal lactic acid, and modified by the addition of I j per cent sodium chloride. The spore-like bodies, transferred from the codfish “freckles,”’ remained apparently unchanged on the agar surface of the plate for weeks, during which time the medium evaporated and the salt appeared a s crystals in the agar and on its surface. Careful microscopic examination of these plates showed t h a t the spore-like bodies developed in this concentrated material. The mold growth continued microscopic in area long enough for the plates t o be overlooked. On reexamination this microscopic growth had developed into a local, living, white fluffiness of mold hyphae, bearing the monilia-like spores of Monascus p u r p u r e u s , and showing under the microscope the characteristic ascus-like bodies. The chief interest of this investigation is the production of the red color. Lafar claims for red rice a clear carmine t o gray-red color on pulverization. Buchanan’s Monascus p u r p u r e u s from silage grew on silage agar and broth, and on gelatin solutions, with the production of white t o red t o deep carmine mold growth, according t o the age and the food given. On rice flour medium the surface mold growth was gray, with t h e medium a brilliant carmine. Two strains of Monascus purpureus from silage, sent in by A. R. Lamb, of Iowa State College, were retained and studied2 in the Microbiological Laboratory. One, A, is a more or less floccose fungus of a white, soiled white, or gray color; the other, B, is similar t o Buchanan’s organism in its color reactions, and i t grows in a submerged manner. Four strains of MonasG U S p u r p u r e u s were secured from Chinese products, three from the superficial red coloring on soy bean cheeses and one from red rice. Each shows some variation from the others in the intensity of the color reaction and rate of growth under the same conditions. Nevertheless all are apparently morphologically alike, and duplicate the description of Buchanan, Went, Ikene,3 U ~ e d a and ,~ Har~.~ The laboratory experiments with the pure culture manufacture of red rice were begun in December 1917.
* A. W. Bitting, “Preparation of the Cod and Other Salt Fish for the Market,” U. S. Dept. of Agr., Bureau of Chem., BulZetin 138 (1911), 1-63. 9 Czapek solution agar with 3 per cent sucrose was used as a standard medium. “uber die Sporenbilding und Systematische Stellung von Monascus Durpureus Went,” Ber. botan. Ges , 21 (1903), 259. 4 The Botanical Magazine, Tokyo, 16 (1902), 160. 8 “Physomyces heterosporus,” B o t ~ n .Cent?., 41 (1890), 378, 405.
*
