act as a shutter against the mouth of the venturi of the mixing chamber to control the uptake of primary air. Thus an annular inlet, open at the bottom, is formed for the primary air between the jacket of the nozzle chamber and the skirt of the mixing chamber.

Stoces, or support assemblies.-The functions of the stove, figures 94 and 95, are to support the boiling vessel and burner in constant locations and thus maintain the correct spatial relations between them; to shield the flames and hot combustion gases from the disturbances caused by cross drafts (provided they are not too strong) and otherwise to assist in directing the hot gases against the bottom and sides of the vessel below the surface level of the solution; and finally, to divert the hot gases away from the sides of the vessel above the solution.

The top plate, of transite board with centrally located hole into which the beaker fits, is supported on three strap-iron legs which also support the truncated conical flame shield of polished sheet aluminum. The lower ends of the legs are bolted to the wide iron base upon which the burner is mounted, which in turn is supported upon one fixed foot and two leveling screws.

In stoves for use with glass beakers, which for any particular nominal size vary slightly in diameter, the central hole is cylindrical and just large enough to avoid any binding when the largest beaker in the set is inserted. The beakers are supported at the correct height above the burner (and at the correct insertion through the plate) by means of three thin Monel metal or stainless steel brackets attached to the under side of the plate. The boiling-vessel assembly should be set into the stove with the center of the thermometer bulb in a vertical plane bisecting the angle between two adjacent rays of the burner.

Inclined manometers (or gages). For the convenient observation of the service pressure and the regulated pressure of the gas in the making and inspection of the adjustment of the flow manostat and burners, provide a pair of oil- or water-filled hydrostatic gages with scales about 100 cm long and inclined at suitable (and therefore different) angles. Other types of heating equipment.-Electric-resistance heaters, for example, may be used in place of the special star burner with flow manostatic control described above, provided ample comparative candy tests on samples of a sufficient range of quality have demonstrated that they yield results equivalent to the results obtained with the burner.

Adjustment of the heating device. The heating device is adjusted in the following manner: A minimum of four standard simple barleysugar tests (basic method, p. 372) are run on a pure sucrose sample (p. 385), and the heating device is so adjusted that the mean duration of the time of cooking, between the temperatures 30° and 176.7° C, shall be 20 minutes 20 seconds. Make similar tests at sufficiently frequent intervals to assure that adjustment is not changing.

The first (or coarse) adjustment is made by installing the proper disk nozzle in the burner. The nozzle is of such a size as to yield nearly the required interval. Final adjustment is made by changing the level of hydrostatic liquid in the flow manostat. As a preliminary criterion of adjustment, it is convenient to heat 300.0 g of water in the

323414°. -42- 26

600-ml beaker through the 50-degree interval from 30° to 80° C (or the 90-degree interval from 86° to 176° F). The cooking interval will be approximately six times the time required to heat the water through the specified interval, i. e., the water should heat in approximately 3 minutes 20 seconds as the mean of a minimum of four tests. (3) THERMOMETERS.-Mercury-in-glass thermometers are suitable for all ordinary candy-test procedures, provided the instruments are properly constructed. They may be graduated in either the centigrade or the Fahrenheit scale, as specified by the user, but must not be graduated in both scales on the same thermometer. They should conform preferably with the specifications presented below, which provide an instrument of low lag. Alternatively, thermometers conforming with ASTM Specifications D 183-25 or thermometers for general use may be employed.

TYPE.

SPECIFICATIONS FOR CANDY-TEST THERMOMETERS

Etched stem, glass.

LIQUID. Mercury.

RANGE AND SUBDIVISION.-Minus 10° to plus 200° C, in 1.0°, with expansion chamber at the top to permit heating to 50° above the upper limit of the scale, or plus 14° to 390° F, in 2.0°, with expansion chamber at the top to permit heating to 90° above the upper limit of the scale.

TOTAL LENGTH.-340 to 360 mm.

STEM.-Plain front, enamel back, made of suitable thermometer tubing. Diameter 6 to 7 mm.

BULB. Corning normal or equally suitable thermometric glass. Diameter not less than 4 mm and not greater than 5 mm. Length not less than 10 mm and not greater than 15 mm.

DISTANCE FROM BOTTOM OF BULB TO 10° C OR 14° F MARK.-80 to 100 mm. DISTANCE FROM 200° C OR 390° F MARK TO TOP OF THERMOMETER.-30 to 70 mm. LENGTH OF UNCHANGED CAPILLARY between top of bulb and first graduation mark, 60-mm minimum; and between the last graduation mark and the expansion chamber at the top, 10-mm minimum.

TOP FINISH.-Glass ring.

SPACE ABOVE MERCURY.-Filled with nitrogen or other suitable gas. GRADUATION. All lines, figures and letters are to be clear cut and distinct. Maximum width of graduation lines, 0.1 mm. On the centigrade scale, all graduation lines at multiples of 5° are to be longer than the remaining graduation marks. Graduations are to be numbered at each multiple of 10°. On the Fahrenheit scale, all graduation lines at multiples of 10° are to be longer than the remaining graduation marks. Graduations are to be numbered at intervals of 20° beginning at 20°.

IMMERSION, 76 mm.-The words "76-mm immersion" and a line around the stem at a distance of 75 to 77 mm above the bottom of the bulb are to be etched on the stem.

SPECIAL MARKING.-The words "NBS CANDY TEST", a serial number and the manufacturer's name or trade-mark shall be etched upon the stem. The marking "C" or "F", as the case may be, shall be etched on the front of the stem above the scale.

SCALE ERROR. The error at any point on the scale, when the thermometer is standardized at 76-mm immersion in a room at 25° C temperature, shall not exceed 1° C or 2° F.

CASE. The thermometer shall be supplied in a suitable case, on which shall appear the marking "NBS CANDY TEST, -10° TO 200° c" or "NBS CANDY TEST, 14 TO 390 F", as the description may require.

NOTE. For the purpose of interpreting these specifications, the following definitions apply:

The total length is the over-all length of the instrument, as supplied.
The diameter is that measured with a ring gage.

The length of the bulb is the distance from the bottom of the bulb to the beginning of the enamel backing.

The top of the thermometer is the top of the finished instrument.

(4) COOLING SLABS. Water-cooled slabs for standard plaques. As constructed according to the drawing, figure 96, two circular plates of aluminum are drawn down at their peripheries to form short skirts which can be placed edge to edge to form the water chamber. These are notched in such a way that when so placed, a pair of circular openings can be formed for the insertion of the inlet and outlet nipples. Within the chamber, with its inlet nipple protruding through one of the openings, is placed a flat spiral coil of aluminum tubing in contact with both plates and having its open discharge end at the center of the assembly. The outlet nipple is inserted in the other hole. The coil is tack- or spot-welded to one of the plates at a sufficient number of points to assure its remaining in place, and the joints between the skirts, as also between the nipples and the skirts, are carefully welded.

Plan view and side elevation. The latter is quarter-sectioned to illustrate the construction described in the text.

The faces marked ƒ are to be carefully polished and they must not be marred in use. Alternatively (but with some disadvantages), the slab may be constructed of copper, with brazing or silver soldering in place of welding.

(This slab may be constructed in any other manner which achieves approximately the same result and cools the candy evenly by a flow of water directed to absorb the heat most rapidly at the center of the plaque, where the temperature is highest.)

In service, the slab should be level. Water should enter the coil at a temperature of 25° C and should flow at a rate of about 5 liters per minute into the center of the water chamber, thence through the spiral path between contiguous turns of the coil to waste through the outlet nipple.

Quenching plate. For the prompt cooling of the superheated bottom of the beaker (to prevent charring of the residual sirup, which fouls the glass and complicates the washing), provide a square piece of 4-inch aluminum or sheet copper measuring 12 to 14 inches square. Aluminum is preferable to copper because of its lighter weight.

(d) HEATING CYCLE

The heating cycle of any candy sugar test includes the cooling interval (during which the material is restored to room temperature) as well as the cooking interval, which itself includes three main

Temperature, °C

180

160

140

120

100

80

regimes, or phases, of heating, as indicated by the temperature gradients of the cooking interval presented in figure 97. These three phases comprise (1) The bringing of the mixture to the boiling point (accomplished under two regimes. in the NBS method (2) the concentration (usually under cover) to proximate saturation at the current temperature, and (3) the further concentration of the sirup under continuously increasing supersaturation to the terminal temperature. From what is said below of the utility and reliability of the candy test, it is evident that the most advantageous duration to select for each of these intervals is that which differentiates FIGURE 97.-Temperature gradients the samples most precisely and most developed during the 20-minute cook- reliably in respect to their stability ing interval of a simple barley-sugar properties. The same thing is true test conducted according to National Bureau of Standards procedure, of the terminal temperatures. If the using a 600-ml glass beaker on the heat effect is carried beyond an stove assembly illustrated in figure 7. established optimum in either reTwo regimes of heating prevail in the first (or spect, the separation becomes blurred.

40

20

2 4 6 8 10 12 Hundreds of Seconds

preboiling) phase of this procedure. The resulting four regimes of heating: (1) with

stirring of the covered sirup prior to boiling,

(4) with boiling of the sirup in the open beaker

is terminated; (b) at the transition from quies

(1) DURATION OF THE COOKING stirring of the uncovered sirup, (2) without INTERVAL. In all standard proce(3) with boiling of the sirup under cover, and dures conforming with these general are set off by the three abrupt flexures of the methods of candy testing, the correct temperature curve (a) at 70° C, where stirring cooking interval is assumed to be cent heating to vigorous ebullition of the sirup; that which results when the test is and (c) at 120° C, where the sirup is uncovered. conducted in apparatus which has been carefully standardized and the heating device adjusted with reference samples of pure sucrose (p. 385) in the basic simple barleysugar test, as directed on page 372. As long as the specified adjustment of the apparatus is maintained the correct cooking interval will be yielded in any standard test on a sample of any appropriate composition, although the actual time elapsing between 30° C (86° F) and the terminal boiling temperature may be not quite the 20 minutes 20 seconds, required under reference conditions.

From this it follows that if the mean duration of the cooking interval in a minimum of four standard simple barley-sugar tests on an

"unknown" sample of commercial sugar should deviate from the mean duration observed in calibration with the prototype sample by very much in excess of 20 seconds, such deviation from the mean could be regarded as a characteristic property of the commercial sample.

(2) DURATION OF THE COOLING INTERVAL. Although the decomposition of the ingredients of a candy test is most rapid at the terminal boiling temperature, there seems to be less promise of a gain in differentiation of samples [10, 11] by either delay or prolongation of the cooling interval than in shortening it and making it more uniform. This fact sufficiently explains the specification of the water-cooled slab in standard tests [13].

(e) SAMPLES

(1) TYPES OF MATERIALS TESTED.--This discussion of the National Bureau of Standards improved simple barley-sugar test is limited to its use (a) in the standardization of equipment for candy tests in general, which involves only reference samples of specially purified sucrose (see below); and (b) in the routine inspection of commercial white sugars, where ordinarily only samples of char-refined sugars (and direct-consumption sugars of comparable purity) will be involved. A discussion of the modifications of the basic procedure permitted for the setting up of other types of candy tests will be presented elsewhere [13].

(2) REFERENCE SAMPLES.-Reference samples for the purposes of candy-test procedure include two general classes of materials: (a) Specially purified lots of an individual chemical substance, such as sucrose, which, for practical purposes, are essentially reproducible. These may be designated primary or prototype samples, (b) materials such as commercial sucrose, dextrose, and corn sirup. As a class, these may vary considerably in composition.

For the most critical reference purposes, prototype sucrose prepared according to the method of Bates and Jackson [3] is specified. For the standardization of equipment (p. 381), grade A char-refined sugar generally serves for the reference sample which, although specially purified as a commercial product, really belongs to the second class of reference material.

(f) IMPROVEMENTS OF THE NATIONAL BUREAU OF STANDARDS METHOD OVER THE HOOKER PROCEDURE

For the reasons stated in the introduction, the National Bureau of Standards basic method for the simple barley-sugar test differs from the Hooker candy test in the following details:

The boiling vessel is a 600-ml chemical-resistant glass beaker with detachable thermometer support in place of a copper casserole with no thermometer support.

The weight of sample is 250.00 g of sugar on a dry basis, instead of 1⁄2 pound without reference to moisture content.

The water is added to make up to a total weight of 350.00 g, i. e., 100.00 g of total water, including the moisture in the sample, instead of 3 oz (87 or 89 ml) of water by volume. This involves a decrease of about 2 percent in the ratio of sugar to water.