Diversification of Morphological Features of the Dark European Honey Bee of the ‘Augustow M’ Line

Simple Summary The necessity of protecting the Dark European honey bee (Apis m. mellifera) in Poland was realized in the 1970s. This was a consequence of the displacement of native bees caused mainly by the mass importation of queen bees of foreign species. Today, we have four programs for the conservation of the following lines of our Dark European honey bee: Augustow M, Kampinos M, Asta M and Northern M. These programs aim to keep the bee gene pool as rich as possible and to preserve the phenotypic and behavioral features characteristic of the Apis m. mellifera species. The aim of this study was to investigate the diversification of morphological features of the Dark European honey bee of the Augustow M line. The authors have demonstrated that the morphological features of the Augustow M line—crucial for identifying the species affiliation, determined as part of the authors’ research—are consistent with the corresponding features described by relevant Polish references used for the conservation of native bee lines. Abstract The aim of this study was to investigate the diversification of morphological features of the Dark European honey bee of the Augustow M line. The authors studied the proboscis length and cubital index, as features determining the affiliation to the species; the width of tergite 4 and the sum of widths of tergites 3 + 4, as indicators of the bee body size; and the length and width of the right forewing. They compared bees sampled from (1) the “lead apiary”, (2) “associate apiaries” and (3) “conservation area apiaries”—apiaries situated in the conservation area established by the national program for the conservation of genetic resources of this bee line. The conclusion was that it is possible to protect bees of the Augustow M line under the existing program, based on resources available to the lead, associate and conservation area apiaries. The bees studied have the essential features of the Dark European honey bee and the values of parameters tested are consistent with the morphological feature references valid for Apis m. mellifera. On the other hand, based on the authors’ research and on other studies described in literature of 1960s, there is a dwarfing trend in the Dark European honey bee of the Augustow M line.


Introduction
The work to describe European bee populations and their diversity more thoroughly was inspired by the research of Alpatov [1][2][3][4] and Goetze [5]. Morphometric studies of phenotypes and of the diversification of these bees made it possible to describe changes in the bee body dimensions depending on latitude [4,6]. Sophisticated statistical analyses of these morphological features carried out by further researchers in 1975 to 1986 made it possible to distinguish the Apis mellifera subspecies and to locate their habitats [7][8][9][10].
Most of the territory of Poland was inhabited by bees showing features typical for Apis m. mellifera, while a population more related to forest bees (Apis m. silvarum Scor.) lived in the former Bialystok province [11]. Morphometric studies have identified four distinct populations of Dark European honey bee living in Poland. the lead apiary, on the one hand, and the associate and conservation area apiaries, on the other hand. The lead apiary sources only such queens that meet the reference requirements for the line and distributes quality genetic material to the associate and conservation area apiaries [20,21].
Bees introduced to the conservation area are strictly selected based on their confirmed affiliation to the Augustow M line to minimize the risk of contamination of the protected populations with foreign genes (e.g., as a result of the intrusion of drones or migration of swarms). The main criteria for the selection include body color, glossa length, cubital index value and tergite 4 width or the sum of widths of tergites 3 + 4 [28].
Various methods, evolving over time, were used to study the morphological features of bees. The standard morphometry was employed by many researchers [11,29,30]. The development in computer-aided techniques made it possible to elaborate the geometric morphometry method that has been readily adopted by many researchers [31][32][33][34][35][36][37][38][39][40][41]. However, comparisons of the two methods have revealed that the geometric morphometry is just slightly more effective than the standard one in the discrimination between bee subspecies [40]. Then, Bustamante et al. [42] found that the application of the geometric morphometry for the comparison of two bee subspecies was less accurate (73.7%), even taking into account data after studying all wings in comparison to the analysis of data obtained using the standard morphometry (97%). DNA tests are often used to eliminate hybridized specimens from protected populations [40,43,44] but a selection based on DNA markers as the single criterion can lead to a loss of genes making the phenotype. This means that the DNA test (identifying diagnostic alleles for the Apis m. mellifera) and the phenotypic analysis (recognizing alleles decisive for the characteristic phenotypic features of the Dark European honey bee including the Augustow M line) should go hand in hand.
Even though we have a number of methods, the conventional morphological evaluation has not become any less important. Researchers still value the method and use it willingly [40,42,45]. Note that the standard morphological method makes it possible not only to reliably measure such features as the body size or proboscis length but also to track the diversification of such features. This is very important for programs for the conservation of genetic resources of the Dark European honey bee including the Augustow M line. Additionally, references and models generally adopted and valid for the Apis m. mellifera, developed in, and applied since, the 1960s, are still used for the determination of the species affiliation of the Augustow M line bees [18,36,46].
The purpose of the authors' study was to analyze the diversification of morphological features of the native Dark European honey bee of the Augustow M line. Furthermore, the study aimed to verify whether the program for the conservation of genetic resources of the Augustow M line, based on collaboration of the lead, associate and conservation area apiaries, can sustain the distinct features of the Dark European honey bee and whether the program follows the references valid for Apis m. mellifera.

Materials and Methods
The experimental material consisted of specimens of the Dark European honey bee (Apis m. mellifera) of the Augustow M line, covered by the program for the conservation of genetic resources of bees in Poland, sampled over ten successive years. The specimens were sourced from the Breeding Apiary in Parzniew (formerly owned by the Animal Breeding and Insemination Station in Bydgoszcz and currently by the National Board of Agricultural Chambers). The stationary apiaries were operated in an extensive manner. The apiaries performed standard treatments of the hive management method with a relatively poor forage. The health of the honeybee colony was regularly monitored and standard procedures were used to combat varroasis.
Bees were sampled from 265 bee colonies, the following nine lead, associate and conservation area apiaries breeding the Augustow M line as part of the conservation program:
Associate apiaries (localities: Jasionowo, Bryzgiel, Lipsk, Danowskie); The multistage clustering method (cluster draw) [47] was used for the sampling in the successive years of random bee specimens from formerly selected random families. A different number of specimens were collected in each year, so the "system" was unbalanced.

Morphological Analyzes
Specimens placed in the Fotie boxes were thrown into boiling water to straighten up glossa for measurement. Then, the specimens were drained, placed in glass containers in groups of 30 and preserved with 70% ethyl alcohol (supplier: Poch S.A., Warsaw, Poland). The right forewing (further referred to as the "wing"), abdominal tergites 3 + 4 and glossa were collected and prepared for each specimen as an input to seven measurements. Four measurements were taken for the wing (length, width and cubital index) shows images 1-3 ( Figure 1), two for the abdomen (widths of tergites 3 + 4) shows images 4-7 ( Figure 2) and one for the glossa shows images 8-10 Figure 3). In aggregate, 6927 worker bees were examined by 48,489 measurements.
The multistage clustering method (cluster draw) [47] was used for the sampling in the successive years of random bee specimens from formerly selected random families. A different number of specimens were collected in each year, so the "system" was unbalanced.

Morphological Analyzes
Specimens placed in the Fotie boxes were thrown into boiling water to straighten up glossa for measurement. Then, the specimens were drained, placed in glass containers in groups of 30 and preserved with 70% ethyl alcohol (supplier: Poch S.A., Warsaw, Poland). The right forewing (further referred to as the "wing"), abdominal tergites 3 + 4 and glossa were collected and prepared for each specimen as an input to seven measurements. Four measurements were taken for the wing (length, width and cubital index) shows images 1-3 ( Figure 1), two for the abdomen (widths of tergites 3 + 4) shows images 4-7 ( Figure 2) and one for the glossa shows images 8-10 Figure 3). In aggregate, 6927 worker bees were examined by 48489 measurements.     Prepared body parts of each specimen were placed on three slide frames and covered with microscopic cover slips glued to the frames: one for the wing, one for the 3 + 4 tergites and one for the glossa. The morphological measurements were performed on the frames loaded in the Apimetr instrument (custom-manufactured by Polskie Zakłady Optyczne/Polish Optical Company/; equipment dedicated to morphometric measurements; device data: electronic slide caliper-measurement accuracy 0.01 mm, slide projector with asymmetric lens eliminating distortion of the image of elements measured on the screen of the apimetr-magnification 20× optical zoom) as shown in image 11 ( Figure 4). Images of body parts displayed by this instrument in a 20× optical zoom were measured with a special perspex jaw caliper. The readouts were entered in the computer database as actual dimensions. Prepared body parts of each specimen were placed on three slide frames and covered with microscopic cover slips glued to the frames: one for the wing, one for the 3 + 4 tergites and one for the glossa. The morphological measurements were performed on the frames loaded in the Apimetr instrument (custom-manufactured by Polskie Zakłady Optyczne/Polish Optical Company/; equipment dedicated to morphometric measurements; device data: electronic slide caliper-measurement accuracy 0.01 mm, slide projector with asymmetric lens eliminating distortion of the image of elements measured on the screen of the apimetr-magnification 20 × optical zoom) as shown in image 11 ( Figure 4). Images of body parts displayed by this instrument in a 20 × optical zoom were measured with a special perspex jaw caliper. The readouts were entered in the computer database as actual dimensions.

Statistical Analyses
A unifactorial analysis by the mean least squares method, carried out by means of the PASW Statistics 23 (2020) software suite, was used for the statistical elaboration of the results. After initial statistical analyses, the description of the study took account of only statistically significant relations between various factors (p < 0.01 or p < 0.05). Pearson correlation coefficients were computed for the examination of relations between the seven morphological features of the Dark European honey bee of the Augustow M line.

Results
The wing length, wing form factor and glossa length of the Dark European honey bee

Statistical Analyses
A unifactorial analysis by the mean least squares method, carried out by means of the PASW Statistics 23 (2020) software suite, was used for the statistical elaboration of the results. After initial statistical analyses, the description of the study took account of only statistically significant relations between various factors (p < 0.01 or p < 0.05). Pearson correlation coefficients were computed for the examination of relations between the seven morphological features of the Dark European honey bee of the Augustow M line.

Results
The wing length, wing form factor and glossa length of the Dark European honey bee of the Augustow M line had the smallest coefficient of variation among the features tested ( Table 1). The parameter of the wing length was 56% and 49% smaller than the variation coefficient for the wing width and for the width of tergite 4, respectively. The values of the coefficient of variation were similar for the wing width, for the width of tergite 4 and for the sum of tergite 3 + 4 widths-the variation ranged from 2% to 16%. The Augustow M line conservation program keeps the breeding material in the lead, associate and conservation area apiaries. This distribution has been addressed in the elaboration of the results ( Table 2). Based on Table 2, bees sampled from the associate apiaries had longer and wider wings than bees from the lead and conservation area apiaries: 0.6% and 0.55% longer and 2.3% and 0.93% wider, respectively. Likewise, bees from the associate apiaries had wider Animals 2021, 11, 1156 7 of 13 tergite 4 and a larger sum of the widths of tergites 3 + 4. The mean width of tergite 4 was 1.85% and 0.4% larger than those for the lead and conservation area apiaries, respectively. The same was the case for the sum of the widths of tergites 3 + 4 (1.4% and 0.5%) and for the cubital index (1.8% and 4.63%). Just one parameter, the glossa length, broke the pattern: specimens from the conservation area had a length 0.01% and 0.9% longer than those from the associate and lead apiaries, respectively.
Associate apiaries kept the breeding material by following the same methods as the lead apiary, so morphological features were compared for both the apiary types taken together (Table 3). There were statistically significant differences only between the wings of the specimens. The wings of bees from the associate apiaries were significantly longer (0.029 mm) and wider (0.02 mm) than of those from the lead apiary. There were no statistically significant differences for the remaining features. This legitimized the treating of these two apiary types as a single group and comparing the typical features of their bees to those of bees from the conservation area apiaries (critical for the Augustow M line conservation program) ( Table 4). Bees from the conservation area apiaries had significantly longer wings (0.03 mm) and wider tergite 4 (0.013 mm) than bees from the two other apiary types. The remaining differences turned out to be statistically insignificant (Table 4). Pearson correlation coefficients were computed for the examination of relations between the seven morphological features of the Dark European honey bee of the Augustow M line. The correlations were estimated for the same number of degrees of freedom (df = 5748) ( Table 5). Only in two cases did the correlations turn out to be statistically insignificant: those between the wing length and the cubital index and between the bee size (the sum of widths of tergites 3 + 4) and the cubital index. The values of the coefficient r for these relations was negative and close to null. The correlation between the wing width and the form factor was weak but statistically significant (r = 0.026 at p = 0.0447) as in the case of the negative correlation between the width of tergite 4 and the cubital index (r = −0.030 at p = 0.0219). The remaining relations were statistically highly significant. A negative and equally strong correlation existed between the wing form factor and the wing width. It was interesting to discover weak, but statistically highly significant, correlations between the glossa length and the bee size (both the width of tergite 4 and the sum of widths of tergites 3 + 4, for which r ranged from 0.191 to 0.202). It is also notable that there are weak but highly significant negative correlations Animals 2021, 11, 1156 8 of 13 between the wing form factor and the bee size (tergite 4 width) and between the wing form factor and the glossa length, for which the coefficient r was equal to −0.266 and −0.251, respectively (Table 5).

Discussion
The study investigated the morphological diversification of the Dark European honey bee of the Augustow M line covered by the conservation program. The cubital index is one of the features recognized by researchers as enabling the verification of the affiliation of a given bee population to a specific subspecies. The mean values of the cubital index for the Danish population of Dark European honey bees range from 1.580 to 1.880 [16]. The mean value of this parameter determined by the authors for the Polish Augustow M line-first without the split into the apiary types (1.664) ( Table 1) and, then, taking into account the split into the lead apiary (1.665), associate apiaries (1.695) and conservation area apiaries (1.620) ( Table 2)-fits within this range. The foregoing results are also similar to the figure established by Goetze [6] for bees in Germany (1.690). On the other hand, bees from the Polish conservation area and associate apiaries are 5.2% and 10.1%, respectively, larger than bees in Russia (cubital index 1.540) [2]. At the same time, the foregoing figures coming from the author's research are 6.4% smaller than the largest mean value of the index for the Augustow M line (associate apiaries) and smaller than the bottom value determined for Lithuanian bees [48] ( Table 2). The values of the cubital index for bees in French Brittany [7,49] and in Cavennes [50] ranged from 1.760 to 1.780 and from 1.70 to 1.88, respectively. According to Ruttner [16], the cubital index value was 1.721 for bees in Austria and 1.840 for western-European bees known as "black bees". By comparing the values of the cubital index for France, Austria and Germany, the authors found that bees from these regions had values 4.9 to 11% larger than bees of the Augustow M line (Tables 1 and 2).
The mean values of the glossa length for all bees of the Augustow M line are as follows: without the split into the apiary types-6.104 mm; with the split into the lead, associate and conservation area apiaries-6.082 mm, 6.137 mm and 6.138 mm, respectively. The results of the authors' research are consistent with the values determined by Ruttner [10], 5.8 to 6.4 mm; and by Ruttner et al. [16], 5.950 to 6.190 mm (Tables 1 and 2).  (Table 1) and with the split (Table 2). Similarly, the authors' results for the size (the sum of widths of tergites 3 + 4) of bees in Poland coincides with the findings of Gromisz [51], 4.730 to 4.90 mm; and Prabucki and Mickiewicz [52], 4.78 mm. According to earlier studies, Dark European honey bees in Poland used to be larger in the 1960s and 1970s. The mean sum of tergites 3 + 4 ranged from 4.790 mm to 4.990 mm for bees from Szepietow and Konskowola [53] and was equal to 4.850 mm for bees living in northern Poland [54]. Then, in the 1980s, the size of Apis m. mellifera in Poland, measured by the width of abdominal tergite 4, decreased to 2.356 mm [55] and, later on, for the Augustow M line, according to the authors' research, to 2.284 mm ( Table 1). This shows a clear dwarfing trend.
The mean values of the cubital index for bees studied between 1971 and 1999 are consistent with the results of the authors' research but were higher by 0.186 in 2009 (Table 6). Likewise, the comparison of the mean glossa length determined, contemporaneously with the cubital index, by other researches-Bornus [53], 6.238 mm; Gromisz [51], 6.120 to 6.230 mm; Gromisz [55], 6.115 mm; Gromisz and Bornus [54], 6.149 mm; Gromisz and Platek [56], 6.151-to the results of the authors' research supports the claim that this parameter has not changed since 1960s.  [36]. **-Index values determined by the Goetz method, converted into the Alpatov method using Rostecki's equations [36].
The implementation of the breeding programs in Poland is based, among others, on the application of the mathematical-morphological models for individual bee species (Table 7). These models, developed for the Dark European honey bee, serve as a reference for the evaluation of consistency of morphometric parameters of the lines covered by the conservation programs including the one for the Augustow M line. According to Table 8, bees of the Augustow M line from the lead apiary are sufficiently consistent with the model, although values z are smaller than 2.1. However, the values are negative for the width of tergite 4 for all apiary types, which reveals the dwarfing trend. The analysis of mean values of the index of similarity, y (0 < y < 3), for three features-the width of tergite 4, the glossa length and the cubital index-shows that bees of the Augustow M line from all the three apiary types very closely resemble the reference population, the evidence of which is the value of the y < 1 indicator.   Table 9). The mean values determined by the authors were within the reference range and the mean values of the cubital index and of the glossa length were 2% and 3% larger, respectively. Only the mean width of tergite 4 was 1.6% smaller than in the reference for the Augustow M line. Based on the comparison of the mean values of morphological features of bees sampled from the lead, associate and conservation area apiaries, and for the overall population of the Augustow M line, to the morphological reference applicable to this line, the results were within the range of the reference. The mean cubital index value closest to the reference was found in bees from the conservation area apiaries; the difference was just 0.01. For the lead and associate apiaries, the value was larger by 0.04 and by 0.07, respectively. Bees from all the apiary types had the fourth tergite width smaller than the reference one, though the difference was smallest for bees from the associate apiaries (0.01 mm). The differences for the two remaining apiary types ranged from 0.018 mm to 0.05 mm. The mean values of the glossa length in bees from all the apiary types were smaller than the mean value of this parameter given in the reference and the differences were as follows: 0.182 mm for the lead apiary, 0.237 mm for the associate apiaries and 0.238 mm for the conservation area apiaries.
The authors compared the values of the coefficient of correlation between the bee size (defined by the sum of tergites 3 + 4) and the remaining parameters (wing length and width, cubital index value and glossa length) determined by the authors and by Bornus [53]. According to the authors, the coefficient of correlation between the wing width and the