a

sucrose content of a honey may considerably exceed 10 per cent. Lippmann gives the analysis of a honey gathered by bees near a sugar factory which had a sucrose content of 16.38 per cent. Cuban honeys have frequently been found to exceed 10 per cent of sucrose, a circumstance due perhaps to the bees having gained access to the products of sugar houses.

The average percentage of ash in the ninety-two samples of levorotatory honeys was 0.18 per cent, the range extending from 0.03 to 0.90. With the dextrorotatory or honeydew honeys the range extended from 0.29 to 1.29 per cent, with an average of 0.81. The limit for ash allowed by the Standards Committee is 0.25. The German " Vereinbarungen" set no fixed limit upon ash, but state that the latter varies from 0.1 per cent to 0.8 per cent; pure honeys vary usually only from 0.10 to 0.35 per cent. Twenty of the levorotatory honeys in the table exceeded 0.25 per cent, and nine exceeded 0.35 per cent. It is evident that the limits fixed for ash under the United States standards do not include a number of samples of otherwise pure honeys. Honeys from certain sections may contain much more ash than those from others. Algarroba honey, for example, in the Hawaiian Islands contains 0.34 per cent ash, while in Arizona and Texas the mesquit honey contains less than 0.15 per cent. All honeys from the Hawaiian Islands, in fact, even of the purest floral type, are distinguished by a high ash content, the average of the seven levorotatory Hawaiian honeys in the table being 0.44 per cent. Honeys containing much honeydew are also characterized by a high percentage of ash.

b

DEXTRIN

In the matter of dextrin content the honeys showed considerable variation, ranging from 0.04 to 7.58 per cent for the levorotatory honeys and from 6.02 to 12.95 per cent for the dextrorotatory. König gives as the average of 173 levorotatory honeys 2.89 per cent dextrin, which is about 1 per cent higher than the average of this class of honeys in the table. Comparisons of the dextrin content of honeys analyzed by different chemists have but little value, however, on account of differences in the methods used for estimating this constituent.

The dextrins which are found in honey are apparently not derived to any appreciable extent from floral nectar, but are more probably gathered by bees from other sources, such as the gummy exudations

• Zts. angew. Chem., 1: 633; Chem. der Zuckerarten, 3d ed., p. 1048.

See note p. 74.

of young buds or the wounded surfaces of shrubs and trees. The juice which exudes from the wounded parts very frequently undergoes a viscous fermentation, with the formation of considerable dextran, a gummy product very closely related to dextrin and possessing a high optical activity.

The juice of leaves may be still further modified through the agency of the plant-louse or aphis, the gummy secretions (honeydew) which these and other leaf-feeding insects exude having already been mentioned (p. 12). This is, in fact, the chief source of the abnormalities in the composition of honey. Sporadic outbreaks of abnormal honey in apiaries, especially in dry summers, are, in fact, not infrequent. Honey from trees such as the basswood, oak, poplar, and hickory seems especially liable to contamination with honeydew. Probably all honey contains a certain amount of this and other extraneous exudations, from the merest traces up, and the normal quantities as shown in the table merge so gradually into the abnormal that it is difficult to establish a fixed and definite line between the two classes of products. It is the usual practice to put all dextrorotatory honeys in a class by themselves, and in the present work this precedent has been followed, all honeys polarizing to the right having been included in the abnormal or honeydew class. It must be said, however, that while the point of optical inactivity offers a most convenient line of demarcation to the chemist, such a division is purely arbitrary in character and has no strict scientific basis.

UNDETERMINED MATTER.

The undetermined matter of the honeys reported in the table, taken by difference, represents (in addition to pollen grains, wax particles, and other insoluble matter) the albuminoids, fixed and volatile acids, coloring matter, aromatic bodies (terpenes, etc.), higher alcohols (mannitol, etc.), and various other bodies of indefinite or unknown character. The percentage of such undetermined matter was considerable in certain of the honeys, such as the apple blossom and several of the clovers, and a closer examination of these honeys would have been undertaken had larger amounts of the samples been available.

NITROGEN.

a

The percentage of nitrogen was determined in a number of the honeys, with the following results:

a Determinations made by T. C. Trescot.

The above results for protein are all below the average percentage of nitrogenous matter in the honeys listed by König, which was 1.08 per cent, the results ranging from 0.30 per cent to 2.42 per cent. As a general rule, honeydew honeys have a higher nitrogen content than those of a purer type.

ACIDITY.

The acid of the honey expressed as formic varied from 0.04 per cent to 0.25 per cent, with an average of 0.09 per cent; the acidity of the honeys listed by König ranged from 0.03 per cent to 0.21 per cent, with an average of 0.11 per cent. The high acidity of several of the honeys, as No. 40, was due to a slight souring of the sample through fermentation.

PENTOSANS.

One sample of pure algarroba honey and five samples containing more or less honeydew were distilled with hydrochloric acid and the furfural in the distillate precipitated with phloroglucin. The furfural-phloroglucid was filtered, dried, and weighed, and the corresponding percentage of pentosans determined according to the method of Kröber and Tollens." The following values were obtained:

Determination of pentosans in different samples of honey.

The honeydew honeys, notwithstanding their high percentage of non-sugars, show no difference in pentosans as compared with the algarroba honey. Inasmuch as pure sucrose will give sufficient fur

U. S. Dept. Agr., Bureau of Chemistry, Bul. 107, p. 54. 20230-Bull. 110-08- 4

fural on distillation with hydrochloric acid to correspond to 1.15 per cent of pentosans," the conclusion may be drawn that pentosans are not present in honey unless in very minute traces.

TANNIN BODIES.

All of the honeys were examined for tannin by testing a solution. of the honey with ferric-chlorid solution; twenty-five of the samples gave a positive reaction with this reagent; six, very marked; thirteen, distinct; and six, faint. As a general rule the honeys giving pronounced tannin reactions were high in acid, as Nos. 71, 78, and 51. In flavor, the high tannin honeys were usually characterized by a marked astringency.

COMPARISON OF HONEYS BY VARIETIES AND GROUPS.

A comparison of the honeys by varieties shows usually a welldefined agreement in composition between the individual honeys of each particular class. The alfalfa honeys, for example, are usually marked by a lower content in dextrin and undetermined matter, and a higher sucrose content than any of the other varieties. In fact, two of the eight alfalfa honeys analyzed exceeded 8 per cent of sucrose, the limit set by the Standards Committee. The relatively high purity (low content in dextrins and undetermined matter) of the alfalfa honeys was shared, but to a less degree, by other members of the Leguminosa.

The honeys of the Composite were about the average as regards organic non-sugars. The Rosacea were low in dextrin, but all high in undetermined matter. The buckwheats seemed characterized by an almost entire absence of sucrose and by the presence of tannin bodies. The basswood honeys were all relatively high in dextrin. This was also true of the sumacs, the poplar, oak, hickory, and other tree honeys, all these containing considerable quantities of honeydew. In addition to a high dextrin content, the latter were also characterized by a relatively high amount of ash. Honey gathered from plants or blossoms containing tannin, as the sumac and hop, usually gave pronounced reactions for tannin. The tupelo, mangrove, and sage honeys were all distinguished by their high content of levulose.

Individual differences and peculiarities such as those noted can only be referred to in a very general way; the number of samples analyzed was so small for many of the varieties, and the admixture of other kinds of nectar in certain instances so uncertain that it was impossible to draw any sweeping conclusions as to the composition of each particular honey type.

Tollens, Kröber, and Rimbach; Zts. angew. Chemie, 1902, p. 508.

INFLUENCE OF ENVIRONMENT UPON COMPOSITION.

EFFECT OF HUMIDITY.

Environment or locality has frequently a very marked influence upon the composition and properties of honey. The data contained in the table, however, are not sufficient to warrant any definite conclusions being drawn upon this subject, except as regards two or three constituents. To determine with certainty the influence of environment upon the composition of the solid constituents of honey comparisons of analyses, as a rule, should be made only upon the product derived from a single variety of flowers in order to eliminate the differences which might be due to composition of the floral

nectars.

In the matter of water content, however, variations in the composition of nectars would play a very insignificant part. The percentage of moisture in a honey is determined largely by the degree of ripeness or concentration to which the bees have reduced the nectar in the hive, and the work of the bees in this respect would be promoted or retarded by local atmospheric conditions of dryness or humidity. Honeys from those sections of the country which are marked by low average humidity and precipitation would, therefore, naturally have a lower percentage of water than honeys from sections of higher humidity and rainfall. A comparison of the results in the following table shows, as a matter of fact, a very decided relationship of this kind, the average humidity of the atmosphere, the average annual rainfall, and the water content of honeys from the very arid States of the Southwest and from the more humid States of the upper Mississippi being given:

Comparative table showing the relation between humidity of locality and moisture in honey.

The annual average of stations in the several States from the report of the United States Weather Bureau, 1904-5.