MORPHOMETRY OF BEE WINGS ON EIGHT SIGNS INDICES: Ci, Dbi, Disc.sh, Pci, Ri, Ci.2, Ci.2.1, Ci.3

Abstract

Introduction: The use of the latter is available, usually on a paid basis. In practice, five wing indices are most often used to discriminate wings: Ci, Disc.sh., Dbi, Ri, Pci. Due to the peculiarity of the structure of the wings, the indices Ri and especially Disc.sh. are determined with significant errors (Std.Err.), which reduces the reliability of the classification as a whole. The goal of the work: The authors set out to find a small number of wing identification indices that could be used to improve the classification of bee wings using computer programs that are publicly available. Material and methods of research: 1545 wings of worker bees were used for research. Veining measurements were performed on a computer using wing photographs and the TpsDig2 wing digitization program according to the accepted DAWINO methodology. Calculation of values ​​of signs (indices) of discrimination of the wing, performed by their own program. Statistical processing of the results of the calculation of these indices was performed using StatSoft's STATISTICA 8.0 software package. Results of research and discussion: The coefficients of classification functions for all possible variants of classification into four clusters using the 5th, 6th, 6th, 7th and 8th indices were found. The criteria of weight of each Partial – Lambda index (λр) are given, which show the individual contribution of each to the data classification, and the reliability of the classification (Percent Correct) for each cluster separately and the overall reliability as a whole (Total).  Analysis of the results of the classification shows that the weight the 8 indices used to classify the wings are quite variable, and depend on the base array of wings as well as the number of features used in a particular variant. In summary, it can be considered fair to orient beekeepers in assessing the distinction of bees in the first approximation to the value of the cubital index. However, with more precise discrimination, significant errors are possible in each case, if we do not take into account the influence of a set of other indices and use only the values ​​of the cubital index (Si). Conclusions and prospects for further research: The use of discriminant analysis allows to obtain a reliable tool for the classification of worker bees and bee colonies. The use of 8 versus 5, eliminates the ambiguity of classification on the basis of Ci for two clusters and a priori increases confidence in the quality of classification and gives grounds with high reliability to use it for the initial rapid assessment of individual bees for compliance with certain groups of clusters Key words: Classification of bees, wing indices, discriminant analysis