A Biometrical Study of the Influence of Size of Brood Cell Upon the Size and Variability of the Honeybee (Apis mellifera L.) by Roy A. Grout, 1931

The data presented in this paper show conclusively that it is possible to obtain a larger bee through the use of brood combs constructed from artificial foundation having enlarged cell bases. The author, therefore, corroborates the contentions of Baudoux (8) and Pincot, according to Gillet-Croix (26), that the bees reared in enlarged cells are larger than bees reared in normal cells. Even though the data at hand show that an increase in the size of the cells in which the bees are reared is accompanied by an increase in the dry weight, length of right fore wing, width of right fore wing, sum of the length of the third and the fourth tergites and the length of proboscis, the author hesitates to support them in their contentions that the larger bee is a better bee. Since the crucial test is honey production, further experimental data are necessary.

Baudoux's contentions (8) that the bee reared in the enlarged cell could fly further, would have a greater carrying capacity and could acquire nectar from flowers having a deeper corolla are not substantiated by the data of this experiment. The author is of the opinion that an extensive investigation of these characteristics should be made. Since in this experiment, all three sizes of cells were placed in the same colony in order that the bees emerging from the cells would be the progeny of the same mother, no indication of the increase in the ability of the bee to gather nectar was observed.

The data presented in this paper agree with data presented by Michailov (43 and 44), stating that an increase in the size of the cells in which worker bees are reared is accompanied by a corresponding increase in the weight, length of right fore wing, width of right fore wing, sum of the widths of the third and the fourth tergites and length of proboscis.

Since most writers agree that the principal point in the selection of bees and the acquisition of nectar is length of proboscis, it is of interest to note that in the case of colony 25 the length of proboscis was increased 2.07% through the use of enlarged cells. In the case of colony 21 the length of proboscis was increased 1.51% and in the case of colony 18 the length of proboscis was increased 1.40%. The average maximum increase in the length of the proboscis for the bees of all three colonies is 0.11 mm. Whether this increase in the length of proboscis is significantly related to an increase in the honey production of a colony has yet to be proved.

The general tendencies of the statistical constants of measurements on bees from colonies 25, 18 and 21 reveal the following facts: (1) The measurement of dry weight varies greatly in certain cases and, in general, correlates poorly. Consequently, the author feels that better experimental control should be exercised in the measurement of this character than was used in the present work. (2) The measurement, length of right fore wing, is significantly correlated with all of the other characters. Of the four characters employed in a regression on the length of proboscis and its integral parts the length of the right fore wing is the only character which gives a significant standard regression coefficient and is dominant in those factors used in the estimation of length of proboscis in the regression equations. (3) The width of the right fore wing, while correlating significantly with the length of the right fore wing, shows no tendency to correlate significantly with the sum of the widths of the third and the fourth tergites and dry weight and its tendency to correlate significantly with length of proboscis is questionable. (4) The measurement of the sum of the widths of the third and the fourth tergites tends to correlate significantly with dry weight and length of right fore wing, but there is apparently no significant correlation with width of right fore wing and with length of proboscis. Consequently, the author feels that the sum of the lengths of the third and the fourth tergites is unrelated to length of proboscis and should be omitted in a study of those factors which are related or contribute to length of proboscis. (5) The variation of the character dry weight is greatest in the case of bees from size of cell "C" and least in the case of bees from size of cell "A". The variation of the width of the right fore wing is greatest in the case of bees from size of cell "A" and least in the case of bees from size of cell "C". The variation of the sum of the widths of the third and the fourth tergites is greatest in the case of bees from the size of cell "B" and least in the case of bees from size of cell "A". The variation of the measurements of length of right fore wing and length of proboscis is greatest in the case of bees from size of ell "A" and least in the case of bees from size of cell "B". Since length of proboscis and length of right fore wing have been shown to be significantly correlated, it is of interest to note that the variation in the three sizes of cells has the same trend.

Since Merrill (42) measured only the glossa in his study on the honey-storing ability of the bee and since the Russian method uses the entire length of the proboscis as measured in this paper, a study was made of length of glossa, length of mentum and sum of the lengths of the glossa and the mentum in relation to length of proboscis. While all three measurements correlate significantly with length of proboscis, it was found that length of mentum was not as good an estimation of length of proboscis as was length of glossa and the sum of the lengths of the glossa and mentum. This study substantiates the methods of both the above mentioned sources.

The inconsistencies arising in the data presented in this paper can be attributed only to a lack of proper technique in the treatment of the material and to the fact, that, while the bees were selected from a specific size of cell and were the progeny of the same mother, they were, nevertheless, selected at random from a large population. Consequently, even though it connot be determined to what extent the above factors operated, the author feels that many of the inconsistencies of the data can be attributed to these causes.

Due to these inconsistencies arising in nearly all of the statistical constants presented in this paper, the author is of the opinion that the use of samples containing a larger number of bees selected from each size of cell would strongly tend to give consistently significant results. This opinion is substantiated by Phillips (53), who discounted the work of Merrill (42) on the basis that the correlations presented in his data were based on small numbers of bees taken at intervals during the season. Merrill (49) calculated that it was necessary to examine only forty bees of a colony in order to determine which colony would produce the most honey. The data at hand show that this is not always the case.

Contributing to the peculiarities of the correlations, particularly in the case of the correlation coefficients of length of proboscis with the lengths of its integral parts, is the fact that in certain cases there are spurious correlations. For example, even though length of glossa is significantly correlated with length of proboscis and length of mentum is significantly correlated with length of proboscis, length of mentum is not significantly correlated with length of glossa.

In this experiment, as has been previously stated, all three sizes of cells were placed in the same colony in order that the emerging bees would be the progeny of the same mother. While the worker bees extended the side walls of the cells of the combs in a normal manner and immediately proceeded to make use of the cells, at least for storage purposes, difficulties were experienced in getting the queens to oviposit worker eggs in the enlarged cells. This was especially true in the case of size of cell "C". The combs containing size of cell "C" were often found to contain patches of drone brood and little or no worker brood. These observations showed that, while the worker bees apparently recognized no difference in the three sizes of cells, the queen bees showed a tendency to prefer the smaller cells for ovipositing. This observation agrees with experiments conducted by Lovchinovskaya (39), who showed that, when nine combs containing normal cells and one comb containing enlarged cells were placed in a colony, the worker bees apparently recognized no difference in the size of the cells but that the queen bee apparently recognized this difference and did not oviposit in the enlarged cells. In the reverse experiment the queen bee oviposited in the enlarged cells.

From a single brood count made by the author during the summer of 1931 on colonies entirely supplied with combs containing size of cells "A", colonies entirely supplied with combs containing size of cells "B" and colonies entirely supplied with combs containing size of cells "C", it was shown that the reaction of the colonies to each size of cell was apparently the same. While this observation corroborates the latter experiment of Lovchinovskaya (39) and the experiences of Baudoux (8), the author feels that further experiments, concerning the brood activities of colonies supplied with combs containing enlarged cells, should be made throughout a period of two or more seasons.

While this paper was being prepared, a correlation of dry weight, length of right fore wing, width of right fore wing, sum of the widths of the third and the fourth tergites, length of proboscis and length of the second member of the right labial palpus was calculated on bees from the three sizes of cells from colony 14. Of special interest in this correlation is the fact that only in the case of the bees from size of cell "A" was the correlation coefficient of length of proboscis with the length of the second member of the right labial palpus significant. The correlation coefficients of these factors in the other two groups were insignificant.

Gotze (29), in a recent paper, did not follow the Russian technique for measuring the individual parts of the proboscis and proclaimed that the probability of error in this technique of measurement was so great that it was quite easy, with small differences of individual colonies, to obtain incorrect results. He states that with all the colonies that he "accurately investigated" it was found that the length of the second member of the labial palpus varies with the total length of the proboscis. In his studies he obtained a judgment of the second member of the labial palpus and prescribed a definite formula by which the length of proboscis was estimated from the value of the measurement of the second member of the labial palpus.

A further comparison of the standard deviations of length of proboscis and length of the second member of the right labial palpus shows that the variability in the three sizes of cell differs between the two measurements. The variability of the length of proboscis is greatest in the case of bees from size of cell "C" and least in the case of bees from size of cell "B", while the variability of the length of the second member of the right labial palpus is greatest in the case of bees from size of cell "B" and least in the case of bees from size of cell "C". From this data it is shown that the variability of these two parts as influenced by size of cell is not at all consistent. From the combined study of the correlation coefficients and the standard deviations, it is shown that the second member of the right labial palpus is not related to length of proboscis and does not vary with it. Consequently, the author cannot agree with the conclusions of Gotze (29) upon this subject.

The author takes this opportunity to strongly recommend the technique of measuring the individual parts of the bee as described under "Methods and Materials" in this paper. With a little practice, from three to four measurements could be taken per minute, a speed which would be very difficult to obtain through the use of an ocular micrometer in a binocular microscope. A comparison of the measurements taken by means of the projection system with those taken by means of an ocular micrometer in a binocular scope showed that the projection system is just as accurate, if not more so.

The data presented in this paper show conclusively that it is possible to obtain a larger bee through the use of brood combs constructed from artificial foundation having enlarged cell bases. The author, therefore, corroborates the contentions of Baudoux (8) and Pincot, according to Gillet-Croix (26), that the bees reared in enlarged cells are larger than bees reared in normal cells. Even though the data at hand show that an increase in the size of the cells in which the bees are reared is accompanied by an increase in the dry weight, length of right fore wing, width of right fore wing, sum of the length of the third and the fourth tergites and the length of proboscis, the author hesitates to support them in their contentions that the larger bee is a better bee. Since the crucial test is honey production, further experimental data are necessary.

Baudoux's contentions (8) that the bee reared in the enlarged cell could fly further, would have a greater carrying capacity and could acquire nectar from flowers having a deeper corolla are not substantiated by the data of this experiment. The author is of the opinion that an extensive investigation of these characteristics should be made. Since in this experiment, all three sizes of cells were placed in the same colony in order that the bees emerging from the cells would be the progeny of the same mother, no indication of the increase in the ability of the bee to gather nectar was observed.

The data presented in this paper agree with data presented by Michailov (43 and 44), stating that an increase in the size of the cells in which worker bees are reared is accompanied by a corresponding increase in the weight, length of right fore wing, width of right fore wing, sum of the widths of the third and the fourth tergites and length of proboscis.

Since most writers agree that the principal point in the selection of bees and the acquisition of nectar is length of proboscis, it is of interest to note that in the case of colony 25 the length of proboscis was increased 2.07% through the use of enlarged cells. In the case of colony 21 the length of proboscis was increased 1.51% and in the case of colony 18 the length of proboscis was increased 1.40%. The average maximum increase in the length of the proboscis for the bees of all three colonies is 0.11 mm. Whether this increase in the length of proboscis is significantly related to an increase in the honey production of a colony has yet to be proved.

The general tendencies of the statistical constants of measurements on bees from colonies 25, 18 and 21 reveal the following facts: (1) The measurement of dry weight varies greatly in certain cases and, in general, correlates poorly. Consequently, the author feels that better experimental control should be exercised in the measurement of this character than was used in the present work. (2) The measurement, length of right fore wing, is significantly correlated with all of the other characters. Of the four characters employed in a regression on the length of proboscis and its integral parts the length of the right fore wing is the only character which gives a significant standard regression coefficient and is dominant in those factors used in the estimation of length of proboscis in the regression equations. (3) The width of the right fore wing, while correlating significantly with the length of the right fore wing, shows no tendency to correlate significantly with the sum of the widths of the third and the fourth tergites and dry weight and its tendency to correlate significantly with length of proboscis is questionable. (4) The measurement of the sum of the widths of the third and the fourth tergites tends to correlate significantly with dry weight and length of right fore wing, but there is apparently no significant correlation with width of right fore wing and with length of proboscis. Consequently, the author feels that the sum of the lengths of the third and the fourth tergites is unrelated to length of proboscis and should be omitted in a study of those factors which are related or contribute to length of proboscis. (5) The variation of the character dry weight is greatest in the case of bees from size of cell "C" and least in the case of bees from size of cell "A". The variation of the width of the right fore wing is greatest in the case of bees from size of cell "A" and least in the case of bees from size of cell "C". The variation of the sum of the widths of the third and the fourth tergites is greatest in the case of bees from the size of cell "B" and least in the case of bees from size of cell "A". The variation of the measurements of length of right fore wing and length of proboscis is greatest in the case of bees from size of ell "A" and least in the case of bees from size of cell "B". Since length of proboscis and length of right fore wing have been shown to be significantly correlated, it is of interest to note that the variation in the three sizes of cells has the same trend.

Since Merrill (42) measured only the glossa in his study on the honey-storing ability of the bee and since the Russian method uses the entire length of the proboscis as measured in this paper, a study was made of length of glossa, length of mentum and sum of the lengths of the glossa and the mentum in relation to length of proboscis. While all three measurements correlate significantly with length of proboscis, it was found that length of mentum was not as good an estimation of length of proboscis as was length of glossa and the sum of the lengths of the glossa and mentum. This study substantiates the methods of both the above mentioned sources.

The inconsistencies arising in the data presented in this paper can be attributed only to a lack of proper technique in the treatment of the material and to the fact, that, while the bees were selected from a specific size of cell and were the progeny of the same mother, they were, nevertheless, selected at random from a large population. Consequently, even though it connot be determined to what extent the above factors operated, the author feels that many of the inconsistencies of the data can be attributed to these causes.

Due to these inconsistencies arising in nearly all of the statistical constants presented in this paper, the author is of the opinion that the use of samples containing a larger number of bees selected from each size of cell would strongly tend to give consistently significant results. This opinion is substantiated by Phillips (53), who discounted the work of Merrill (42) on the basis that the correlations presented in his data were based on small numbers of bees taken at intervals during the season. Merrill (49) calculated that it was necessary to examine only forty bees of a colony in order to determine which colony would produce the most honey. The data at hand show that this is not always the case.

Contributing to the peculiarities of the correlations, particularly in the case of the correlation coefficients of length of proboscis with the lengths of its integral parts, is the fact that in certain cases there are spurious correlations. For example, even though length of glossa is significantly correlated with length of proboscis and length of mentum is significantly correlated with length of proboscis, length of mentum is not significantly correlated with length of glossa.

In this experiment, as has been previously stated, all three sizes of cells were placed in the same colony in order that the emerging bees would be the progeny of the same mother. While the worker bees extended the side walls of the cells of the combs in a normal manner and immediately proceeded to make use of the cells, at least for storage purposes, difficulties were experienced in getting the queens to oviposit worker eggs in the enlarged cells. This was especially true in the case of size of cell "C". The combs containing size of cell "C" were often found to contain patches of drone brood and little or no worker brood. These observations showed that, while the worker bees apparently recognized no difference in the three sizes of cells, the queen bees showed a tendency to prefer the smaller cells for ovipositing. This observation agrees with experiments conducted by Lovchinovskaya (39), who showed that, when nine combs containing normal cells and one comb containing enlarged cells were placed in a colony, the worker bees apparently recognized no difference in the size of the cells but that the queen bee apparently recognized this difference and did not oviposit in the enlarged cells. In the reverse experiment the queen bee oviposited in the enlarged cells.

From a single brood count made by the author during the summer of 1931 on colonies entirely supplied with combs containing size of cells "A", colonies entirely supplied with combs containing size of cells "B" and colonies entirely supplied with combs containing size of cells "C", it was shown that the reaction of the colonies to each size of cell was apparently the same. While this observation corroborates the latter experiment of Lovchinovskaya (39) and the experiences of Baudoux (8), the author feels that further experiments, concerning the brood activities of colonies supplied with combs containing enlarged cells, should be made throughout a period of two or more seasons.

While this paper was being prepared, a correlation of dry weight, length of right fore wing, width of right fore wing, sum of the widths of the third and the fourth tergites, length of proboscis and length of the second member of the right labial palpus was calculated on bees from the three sizes of cells from colony 14. Of special interest in this correlation is the fact that only in the case of the bees from size of cell "A" was the correlation coefficient of length of proboscis with the length of the second member of the right labial palpus significant. The correlation coefficients of these factors in the other two groups were insignificant.

Gotze (29), in a recent paper, did not follow the Russian technique for measuring the individual parts of the proboscis and proclaimed that the probability of error in this technique of measurement was so great that it was quite easy, with small differences of individual colonies, to obtain incorrect results. He states that with all the colonies that he "accurately investigated" it was found that the length of the second member of the labial palpus varies with the total length of the proboscis. In his studies he obtained a judgment of the second member of the labial palpus and prescribed a definite formula by which the length of proboscis was estimated from the value of the measurement of the second member of the labial palpus.

A further comparison of the standard deviations of length of proboscis and length of the second member of the right labial palpus shows that the variability in the three sizes of cell differs between the two measurements. The variability of the length of proboscis is greatest in the case of bees from size of cell "C" and least in the case of bees from size of cell "B", while the variability of the length of the second member of the right labial palpus is greatest in the case of bees from size of cell "B" and least in the case of bees from size of cell "C". From this data it is shown that the variability of these two parts as influenced by size of cell is not at all consistent. From the combined study of the correlation coefficients and the standard deviations, it is shown that the second member of the right labial palpus is not related to length of proboscis and does not vary with it. Consequently, the author cannot agree with the conclusions of Gotze (29) upon this subject.

The author takes this opportunity to strongly recommend the technique of measuring the individual parts of the bee as described under "Methods and Materials" in this paper. With a little practice, from three to four measurements could be taken per minute, a speed which would be very difficult to obtain through the use of an ocular micrometer in a binocular microscope. A comparison of the measurements taken by means of the projection system with those taken by means of an ocular micrometer in a binocular scope showed that the projection system is just as accurate, if not more so.