For bisexual reproduction it is essential that the diploid chromosome complement is reduced to haploid during gametogenesis. This is achieved by a specialized cell cycle, meiosis, where the chromosome number of germ line cells is reduced during two rounds of cell divisions after just one round of chromosome replication. In the first, reductional division homologous chromosomes segregate from each other while in the other, equational division individual chromosomes divide. During early prophase of the first meiotic division, homologous chromosomes pair, undergo tight synapsis and form a bivalent. Bivalent integrity is retained throughout prophase with either of two ways. Most often a crossing-over occurs during synapsis creating a physical link or chiasma between the homologs, which are held together until anaphase I. Alternatively, in achiasmate meiosis, pairing, alignment and synapsis are normal, but no chiasmata are formed. In these cases, the paired condition of homologs is retained till the onset of anaphase I (for references see Nokkala and Nokkala 1983, Nokkala 1987, Grozeva et al. 2010).

In natural populations, the meiotic mechanism is under strong selection; all mutations causing disturbances in meiosis are quickly removed from the population. The only exceptions are thelytokous parthenogenetic populations, where males are occasionally found (Lynch 1984, Palmer and Norton 1990). In these rare males defects affecting normal meiotic pattern and making males nonfunctional in sexual reproduction have been described. In two oribatid mite species Trhrypochthonius tectorum (Berlese, 1896) and Platynothrus peltifer (CL Koch, 1839) (Acari: Oribatida: Desmonomata) male meiosis was normal, but spermiogenesis was aberrant; males produced few spermatophores only and females left them without attention (Taberly 1988). In the thelytokous false spider mite Brevipalpus obovatus Donnadieu, 1875 (Phytoptipalpidae) males had low sperm production and did not inseminate females despite copulation (Pinjacker et al. 1981). Spermatogonia are developed into sperm without meiosis in the spider mite Tetranychus urticae Koch, 1836 (Tetranychidae) (Pijnacker and Drenth-Diephuis 1972).

The object of our study is the Holarctic psyllid species Cacopsylla myrtilli (Wagner, 1947), which has long been considered as parthenogenetic (Linnavuori 1951, Lauterer 1963, Ossiannilsson 1975, 1992), since collections have been made from all-female populations (Ossiannilsson 1992). Recently, we established that parthenogenetic females in Cacopsylla myrtilli were triploids and displayed parthenogenesis of an apomictic type (Nokkala et al. 2008). However, also males have been found in some populations in northwest Russia (Ossianilsson 1975, Labina et al. 2009), Norway (Hodkinson 1983), northern Sweden (Hodkinson and Bird 2006) and northern Finland (Labina et al. 2009). Based on the presence of males, either facultative parthenogenesis (Hodkinson 1983, Hodkinson and Bird 2006) or even bisexual reproduction (Labina et al. 2009) has been suggested.

For the present study we have collected Cacopsylla myrtilli from many populations from Norway including recollection of Hodkinson (1983), Sweden including recollection of Hodkinson and Bird (2006), and Finland. For northwest Russia, we have utilized the collection of Labina et al. (2009). We have analyzed populations quantitatively for the presence of males and when present studied male meiosis in detail to reveal possible aberrations and their significance for the meiotic mechanism in general. In addition we wanted to find firm evidence for the origin of males in populations.

Adult Cacopsylla myrtilli specimens were collected in Norway, Sweden, Finland and northwest Russia. For the present study, we selected only those populations, where males were found (Table 1). The Rindhovda population in Norway has been earlier studied by Hodkinson (1983) and the Abisko population in Sweden by Hodkinson and Bird (2006).

Specimens were fixed immediately after collection in the field in freshly made 3:1 Carnoy and stored in fixative in the laboratory at + 6°C until slides were made according to the method by Nokkala et al. (2008). For males collected in alcohol in the field, a novel approach was used. The abdomen of a male was immersed in fresh 3:1 Carnoy fixative for two hours, testes were dissected in a drop of 45% acetic acid and squashed. This procedure allows both chromosomal and molecular analysis of the same individual.

For staining of chromosomes the Feulgen-Giemsa method by Grozeva and Nokkala (1996) was employed with extending staining in Schiff’s reagent to 25 min and subsequent Giemsa staining in Sørensen’s buffer, pH 6.8, to 30 min.

Chromosomes were photographed with BU4-500C CCD camera (BestScope International Limited, Beijing, China) mounted on Olympus BX51 microscope (Olympus, Japan) using ISCapture Sofware version 2.6 (Xintu Photonics Co LTD, Xintu, China). Photographs were processed with Corel Photo-Paint X5 software.

We found males in ten (Table 1) out of a total of 47 populations collected. All populations studied were highly female-biased, male frequency varying from 0.1 % to 9.1 %. The highest proportion of males was found from Rinhovda population in Norway collected above tree line. This same population was earlier studied by Hodkinson (1983). All other populations showed a considerably lower proportion of males, less than 3 %. The lowest proportion of males or 0.1 % was found in Finnmark, Šuoššjávri in Norway and Utsjoki in Finland.

The testes in male psyllids are organized into lobed structures, the number of which varies in a species specific manner (Głowacka et al. 1995). In Cacopsylla myrtilli, males have two lobed testes. The earliest meiotic stage found in an adult male is pachytene, while earlier meiotic stages are rare. The number of cells at pachytene in one primary spermatocyte cyst is 64.

Males collected in Finnmark, Šuoššjávri in Norway, in Paltamo and Utsjoki in Finland and the White Sea area in Russia (Labina et al. 2009), on one hand, and in Sjoa in Norway and in Abisko in northern Sweden, on the other hand, showed strikingly different course of meiosis.

Males from northern Norway, Finland and Russia showed meiotic stages typical for psyllids known to possess holocentric chromosomes. The most common stage in testes was the diffuse stage during which chromatin had a diffuse appearance covering the whole nucleus in a cell. During this stage the size of cells is increased considerably. When the chromosomes condensed out from the diffuse stage, diplotene stage with 12 autosomal bivalents with one chiasma in each and a univalent X chromosome were seen (Fig. 1a). At metaphase I, the bivalents showed axial orientation with homologous telomeres oriented to opposite poles and the univalent X chromosome lying at the equatorial plane (Fig. 1b). Segregation of bivalents at anaphase I resulted in half-bivalents moving towards the poles. The bi-oriented X chromosome was seen as a laggard at this stage. At telophase I, the X chromosome moved to one of the poles (Fig. 1c), resulting in two kinds of secondary spermatocytes, those with the X chromosome and those without it.

However, males collected near Sjoa, Norway (Rindhovda, Rudihøe, Kvernbrusætrin and Stålane) and northern Sweden (Abisko, Lapporten) showed 25 (24 + X) univalent chromosomes at diplotene and diakinesis stages, indicating a complete failure in chiasma formation (Fig. 2a). Univalent chromosomes oriented with sister chromatids to opposite poles at metaphase I (Fig. 2b). These bi-oriented chromosomes divided at anaphase I. No laggard was seen at anaphase I, indicating that the X chromosome also divided in the first meiotic division (Fig. 2c). Anaphase I resulted in secondary spermatocytes with 25 daughter chromosomes, oriented to both poles at metaphase II (Fig. 2d, e). As daughter chromosomes are unable to divide, anaphase II started but could not be completed (Fig. 2f, g). Consequently, diploid spermatids were produced. Meiosis resulted in a cyst of 128 developing spermatids instead of the 256 in normal chiasmate meiosis.

To find out reasons for the absence of chiasmata, early stages of meiosis were analysed in more detail. In chiasmate male meiosis pachytene stage (Fig. 3a) was easily found in all males studied. However, in males without chiasmata no pachytene cells were found in 61 males studied, but leptotene-like nuclei were abundant. When seen in side-view chromosomes showed distinct bouquet orientation (Fig. 3b), indicating that presynaptic alignment was normal, but synapsis did not occur. It seems apparent then that the failure in chiasma formation in these males is due to a mutation resulting in asynapsis of homologous chromosomes.

In this study we have established that males occurred in ten out of 47 Cacopsylla myrtilli populations sampled in northern Europe. All populations were highly female-biased, the proportion of males varying from 0.1 % to 9.1 %. Thus, all populations are in fact parthenogenetic and males are rare or so-called spanandric males. In northern Norway, Finland and Russia male meiosis was normal whereas in Norway and Sweden males displayed asynaptic meiosis leading to the formation of diploid spermatids. These males are nonfunctional in reproduction and appear in every generation as reversals from apomictic parthenogenesis.