by a process of regurgitation, the nectar being continually thrown out from the honey sac on the partly doubled tongue and then drawn in again until, by the movement of the air and the heat of the hive, the nectar is sufficiently reduced to be deposited in the cells of the comb.

Another change of considerable importance which takes place while the nectar is in the honey sac of the bee, and also probably during evaporation and storage in the comb, is the inversion of a considerable part of the sucrose in the nectar through the action of an inverting enzym secreted by the bee. Bonnier gives the following percentages of sugars of sainfoin nectar and sainfoin honey: Percentage of total sugar in sainfoin nectar and sainfoin honey (Bonnier).

a

These figures show that over 85 per cent of the sucrose in the original nectar had been inverted.

Another modification produced in the nectar by the bees is the introduction of a minute quantity of formic acid. This acid is wanting in the pollen and nectar of flowers and is supposed to be introduced into the honey by the bee just previous to capping the cell. The formic acid thus introduced by the bee is supposed to act as a preservative and prevent the honey from fermenting.

HONEYDEW.

While nectar is the principal and purest source of food for bees, there are other exudations from which they may gather honey and which must always be taken into account by the chemist in honey analysis. The most important and common of such exudations is what is known as honeydew. Bonnier, writing on sources of honey,

says:

Honeydew exudes not from special nectaries, but from the general surface of young and even full-grown leaves. Honeydew may be produced annually on very hot days, and it is exceedingly abundant in very dry seasons, when it forms a valuable supply for the bees, although it yields an inferior honey owing to the presence of gum and dextrin. The trees and shrubs which most fre quently produce honeydew in France are the oak, ash, linden, maple, poplar, birch, hazel, mountain ash, blackberry, and barberry. It is occasionally found also on certain herbaceous plants, including salsify, scorzonera, and some cruciferæ. Usually the honeydew is excreted or rejected by plant lice (aphides), which extract the sweet sap of leaves, but digest only a small portion of it.

a Loc. cit.

König, Chem. Nahr. Genussm., 2:995.

This variety is known in France as miellat. But honeydew is sometimes exuded directly from the stomata of leaves in the absence of plant lice. This variety is true honeydew, or miellée.

Honeydew differs from floral nectar in composition in several important particulars. E. v. Raumer gives the following composition of honeydew obtained from maple leaves:

Unger gives the following percentage for the dry matter of two honeydews:

Nitrogenous matter as albumin__.

Analyses by Boussingault give the following data for the dry substance of two honeydews:

Analyses of two honeydews by Boussingault.

d

Analyses by Wiley of honeydew from the pine tree gave the following results:

Water

Sucrose.

Invert sugar

Dextrin, etc. [(α) » = +105]

Per cent.

54.41

8.16

17.44

19.99

The honeydew itself was dextrorotatory, but the resulting honey levorotatory, from which it was inferred that the bees had brought about an inversion.

The various analyses of honeydew quoted all show a high preponderance of ash and of gum or dextrin as compared with floral nectar. These dextrins may be derived in part from the juice of

a Zts. anal. Chem., 1894, 33: 397.

Ber. Kais. Acad. Wissensch. Wien, Mathematisch-Naturwissenschaftliche Classe, 1857, 25: 449.

C Compt. rend., 1872, 74: 87.

Amer. Chem. J., 13:24.

the leaf; it is more probable, however, that they are produced after exudation by fermentation through the agency of gum or dextranproducing bacteria.

a

Among other constituents occurring in honeydew and not in floral nectar should be mentioned dulcite, found by Maquenne in honeydew from Euonymus japonica, and melezitose, found by the same authority to the extent of 40 per cent in honeydew from the linden. The question of honeydew and its bearing upon the composition of honey will be discussed more fully under the subject of abnormal honey.

DESCRIPTION OF SAMPLES ANALYZED.

Over one hundred samples of pure honey were analyzed in this investigation. The honeys represent the product from about fifty different varieties of flowers and were widely distributed, having been gathered from thirty-two States and Territories. Of the samples analyzed several were excluded from the table either on account of doubt as to origin or because the samples had changed in composition through long standing. Eighty-five of the samples were part of a honey exhibit made at the St. Louis Exposition in 1903 by the National Bee Keepers' Association, this exhibit being afterwards donated to the Bureau of Entomology. The following account of this honey display is furnished by Mr. France, of Platteville, Wis., general manager of the association:

To make a display of honey both attractive and educational, I asked the National Bee Keepers' Association members in every State to save carefully and separately the various kinds of honey produced in their locality. Many bee keepers expressed their willingness to help me, but said it was a poor season with them, and there were but few kinds of honey they could secure. A large United States map was fastened to the north wall of the Convention Hall, then bracket shelves were placed across each State, Canada, and Cuba, and on them I put most of the honey, in 157 1-pound clear glass jars with aluminum screw top.

The samples from Mr. France were received in the exposition jars, with their original labels. The Canadian and Cuban samples were not included in the number analyzed. Of the remaining fifteen samples eleven were Hawaiian honeys furnished by Mr. D. L. Van Dine, entomologist of the Hawaii Agricultural Experiment Station in Honolulu, and four were New England honeys supplied by the Bureau of Entomology.

All of the samples had been previously extracted or strained. In several cases where this had been imperfectly performed, the finely divided fragments of comb were removed by filtering the liquefied honey through cloth.

a Bul. soc. chim. Par., (3) 21: 1082.
Compt. rend., 117: 127.

Amer. Bee J., Jan. 5, 1905, p. 5.

The detailed description of the samples is given in the table beginning on page 22.

PREPARATION OF SAMPLES FOR ANALYSIS.

When the honeys were free from granulation and in a perfectly liquid condition, no preparation was necessary. Samples of honeys

FIG. 1.-Distribution of honeys analyzed, showing areas of low and high average humidity. In the dry area: Average annual humidity, 51 per cent; average annual rainfall, 10.6 inches; average moisture content of honeys, 15.09 per cent.

In the moist area: Average annual humidity, 76 per cent; average annual rainfall, 30.9 inches; average water in honeys, 18.88 per cent.

which had granulated (as was generally the case) were carefully warmed by placing the loosely stoppered bottles in a bath of water warmed to 45° to 50° C. until complete solution of the crystallized dextrose was effected. The samples, after thorough mixing, were cooled, when they were ready for analysis.

Error due to imperfect samples of honey.-A very serious error may be introduced into the analysis of honey by imperfect sampling. This applies not only to the sampling of bottled honeys in the laboratory, but also to the sampling of honeys in bulk by inspectors. Nearly all honeys granulate more or less readily, and the result is that a compact crystalline deposit of dextrose settles at the bottom of the container, leaving in solution a greater preponderance of the less readily crystallizable sugar levulose. Leaky containers for the same reason may cause a serious error in the determination of the analytical constants of a honey. Samples taken from the liquid surface and from the crystalline bottom layer of such honeys will show the greatest difference in composition; hence the necessity of thoroughly mixing the contents of a container before sampling. In

the laboratory this should be done by melting. If honeys are sampled in bulk from casks or barrels melting is impracticable, and the most that the inspector can do is to mix the contents as well as possible by shaking and stirring before taking the sample for analysis.

Errors due to changes in composition during storage. It is exceedingly important in making analyses of honey that the work be begun immediately upon the receipt of the samples. Changes in composition may take place through fermentation by yeasts and bacteria, or even through the agency of inverting enzyms left in the honey by the bees (page 12). It has happened in the experience of the Bureau of Chemistry that samples of honey taken by inspectors from casks and bottled and sealed have decreased considerably in sucrose content within four months. Inspection samples of raw honeys from casks should be sterilized to destroy micro-organisms and enzyms before bottling and sealing.

METHODS OF ANALYSIS.

Unfortunately the quantity of sample available for analysis was so small in nearly every case that a detailed study of the less commonly determined components of honey, such as make up the organic solids, not sugars, could not be undertaken. Only the more usual physical and chemical constants were determined, these including the direct and invert polarizations at both 20° and 87° C., and the percentages of water, invert sugar, sucrose, ash, dextrin, and acidity. In a few cases in which the quantity of the samples was sufficient, determinations of nitrogen and pentosans also were made.

The methods of analysis employed were in general those of the Association of Official Agricultural Chemists, as given in Bulletin 107 of the Bureau of Chemistry.

OPTICAL METHODS.

DIRECT POLARIZATION.

For the direct polarization of the honeys the normal weight (26) grams) of the liquefied sample was dissolved to 100 metric cubic centimeters at 20° C. after the addition of 5 cubic centimeters of alumina cream. The solution was filtered and immediately polarized in a 200-millimeter tube in a Schmidt and Haensch triple shadow saccharimeter at 20° C. The same solution was again polarized after eighteen to twenty hours standing, the second reading being taken as the constant polarization of the honey. The difference between the immediate and constant polarization was taken as the birotation.

After taking the constant reading at 20° C. the solution was brought to a temperature of 87° and again polarized. The temperature was maintained at 87° by means of a current of hot water, and constant readings could usually be secured within five minutes. The field