Lacewings were collected from May to October 2013–2018 in the Republic of Armenia, the Eastern Caucasus by G. Karagyan, T. Ghrejyan, I. Stepanjan, and A. Dantchenko, in the Republic of Dagestan, the North-East Caucasus, Russia by G. Khabiev and E. Ilyina, and in the Astrakhan region, Russia by E. Ilyina. Adult males and females were fixed in 3:1 (ethanol: acetic acid) fixative and then stored at 4 °C until required. Collection sites, sampling dates, and the number of studied specimens are given in Table 1. Species identification was made by G. Khabiev and/or V. Krivokhatsky. Voucher specimens have been deposited in the insect collections of Zoological Institute RAS (St. Petersburg, Russia).

The current study is part of a larger research project on Neuroptera, their cytogenetics and evolution. For the last decade, we have prepared dozens of cytological preparations to study the karyotypes and male meiosis of lacewings represented by a quite wide taxonomic range (Kuznetsova et al. 2015, 2016). For those purposes, the insects were fixed in the Carnoy fixative (ethanol and glacial acetic acid, 3:1), the gonads were then dissected out of the abdomen in a drop of 45% acetic acid on a microscope slide and squashed. Before squashing for the subsequent chromosome analyses, the testicular follicles of the male and the ovarioles of the female were carefully separated from each other and counted under a stereomicroscope SZX7, Olympus. Those data became a basis for the present paper. While estimating the variation of the number of follicles per testis and ovarioles per ovary, we calculated the arithmetic mean and its standard deviation (SD) in a given set of data.

Palparinae

Six adult males of Palpares libelluloides (Linnaeus, 1764) were available for examination. Their bean-shaped testes were encapsulated each by the yellow scrotal membrane. In a sample of eleven testes examined (Table 1), the number of follicles varied in the range of 267–472 (mean 347.4, SD 54.18).

Acanthaclisinae

An adult male and an adult female of Acanthaclisis occitanica (Villers, 1789) were examined. In males, the bean-shaped testes were enclosed by the yellow sheath. We counted 51 follicles in one testis and 50 follicles in another testis of the male. In the female, the paired ovaries consisted each of the ten transparent ovarioles.

Nemoleontinae

Six species from five genera were studied, namely, Creoleon plumbeus (Olivier, 1811) (6♂, 1♀), C. griseus (Klug in Ehrenberg, 1834) (1♀), Distoleon tetragrammicus (Fabricius, 1798) (4♂, 1♀), Delfimeus irroratus (Olivier, 1811) (1♀), Neuroleon lukhtanovi Krivokhatsky, 1996 (1♂), and Macronemurus bilineatus Brauer, 1868 (3♂). In males, testes were spindle-shaped and yellow in color. In all studied species, regardless of the tribe they belong, the number of follicles per testis varied between six and four (Table 1). In the male of N. lukhtanovi each of the testes comprised five follicles. In males of C. plumbeus, D. tetragrammicus and M. bilineatus, a variation of the follicle number was observed both between conspecific males and/or between different testes of a male. Specifically, eleven testes checked in C. plumbeus consisted each of five or six follicles (mean 5.63, SD 0.5), eight testes examined in D. tetragrammicus consisted each of four or five follicles (mean 4.6, SD 0.52), and five testes examined in M. bilineatus consisted each of five or six follicles (mean 5.8, SD 0.45). Each of the four species studied in respect to the ovary structure, Creoleon plumbeus, C. griseus, D. tetragrammicus, and Delfimeus irrotatus, displayed ten transparent ovarioles per ovary.

Myrmecaelurinae

Three species from two genera of the tribe Myrmecaeluruni were studied, namely, Novoheus zigan H. Aspöck, U. Aspöck & Hölzel, 1980 (1♂, 1♀), Myrmecaelurus trigrammus (Pallas, 1771) (32♂, 5♀) , and M. solaris Krivokhatsky, 2002 (3♀). In males, testes were spindle-shaped and yellow to orange in color. The only studied male of N. zigan displayed nine follicles per testis. In two geographically distant populations of M. trigrammus (Russia: Astrakhan region and Dagestan), a variation in the follicle number was observed between different males and different testes of a male (Table 1). Overall, in a sample of 53 testes examined the number of follicles varied in the range of 6–11 (mean 9.42, SD 0.99). Each of the three species studied in respect to the ovary structure, N. zigan, M. trigrammus, and M. solaris, displayed ten transparent ovarioles per ovary.

Myrmeleontinae

Three species from two genera were studied, namely, Myrmeleon inconspicuus Rambur, 1842 (4♂), M. immanis Walker, 1853 (3♂), and Euroleon nostras (2♂). In males, testes were spindle-shaped and yellow to orange in color. Testes of M. immanis comprised each five follicles in each of the males explored. In M. inconspicuus, a variation was observed between both males and different testes of a male (Table 1). Specifically, of the eight testes dissected, four had six follicles, two had five follicles, and two had six and five follicles, respectively (mean 5.62, SD 0.52). In E. nostras, each of the males had three follicles per testis.

Ascalaphinae

Two species from two genera were studied, namely, Libelloides macaronius kolyvanensis (Laxmann, 1770) (1♂, 3♀) and Bubopsis hamata (Klug in Ehrenberg, 1834) (2♂). Unfortunately, in the male of L. m. kolyvanensis we were unable to calculate exactly the number of follicles; however, there were no less than twelve follicles in each of the testes. Every female had ten transparent ovarioles per ovary. In B. hamata, one male had approximately 33 and 56 follicles in different testes, whereas the other one had approximately 44 and 50 follicles in different testes (mean 45,75, SD 9,8).

Nemopterinae

Two species from two genera were studied, namely, Lertha ledereri (Sélys-Longchamps, 1866) (2♂) and Nemoptera sinuata Olivier, 1811 (6♂, 2♀). In both species, males had six follicles in each of the paired testes. The testes were spindle-shaped and light yellow in color. The females of N. sinuata had ten transparent ovarioles per ovary.

As mentioned above, we did not study the internal reproductive organs of males and females in depth. In species under study, the male and female reproductive organs seem to correlate with those figured by Quartey and Kumar (1973) for Palpares sp., Neuroleon sp. and Ascalaphus sp. (= Ascalaphus Fabr. cf. africanus (McLachlan, 1871), by Walker et al. (1994) for Palmipenna spp. (Nemopteridae), and by Brauer (1854) for Ascalaphus macaronius/hungaricus (= Libelloides macaronius macaronius (Scopoli, 1763); see Krivokhatsky et al. 2018 for synonymy).

More two-thirds of all hitherto studied species (Table 1) were found to have relatively low number of follicles per testis lying in the range between three (Euroleon nostras) and eleven, the latter number having been revealed in only one of the 53 testes examined in the polymorphic for this trait species Myrmecaelurus trigrammus. Both species belong to the family Myrmeleontidae, in which 14 species from eleven genera and five subfamilies were studied. Low numbers (3–11) were found in the subfamilies Nemoleontinae, Myrmecaelurinae, and Myrmeleontinae, but not in Palparinae and Acanthaclisinae. In the Palparinae, males of Palpares sp. and P. libelluloides showed 40 and 347.4 (on average) follicles per testis, respectively (Table 1). In the only studied male of Acanthaclisis occitanica (Acanthaclisinae) 50 and 51 follicles were counted in different testes. In the family Nemopteridae, all the four studied species in three genera, Lertha Navás, 1910, Nemoptera Latreille, 1802, and Palmipenna Tjeder, 1967, were found to have low numbers, namely, six in the first two genera and five in the last one. The family Ascalaphidae seems to display a rather wide margin of the follicle numbers, although the existing data are scarce and partly controversial. We found that the two studied males of Bubopsis hamata have a relatively high number of follicles, 45,75 (in average) per testis. On the other hand, Libelloides macaronius macaronius and L. m. kolyvanensis showed a relatively low number of follicles although their actual number remained uncertain for both. According to Brauer (1854), males of L. m. macaronius possess approximately six follicles per testis; however, in the only studied male of L. m. kolyvanensis we were able to count at least twelve follicles in each of the testes. It is worth noting that many species showed a variation in the follicle number between both different males and even different testes of a male, this being equally true for species with high and low number of follicles. It is interesting that a variability of the follicles number was likewise reported in some species of Raphidioptera (snakeflies) and Megaloptera (alderflies and dobsonflies), the two other orders of the superorder Neuropterida (Neuropteroidea) (Stitz 1909).

In the majority of genera, the only species was unfortunately studied. The exceptions are the genera Palpares Rambur, 1842 and Palmipenna, each with two explored species, and Myrmeleon Linnaeus, 1767 with three studied species. An important point is that in every case, the congeneric species share similar trends, namely, Palpares spp. have high numbers (40 and higher) while the two other genera have low numbers (five and/or six).

Note that the species differing significantly in the number of follicles seem to have no evident ecological, physiological, and developmental differences; then, it is not quite clear what could be the cause of such a wide variety of testes. It may be that the variation in follicle number manifests adaptation to the male abdomen size and, then, one might expect that the larger the insect the more follicles it has. Indeed, males of Palpares libelluloides, Acanthaclisis occitanica, and Bubopsis hamata, the three species characterized by numerous follicles, have very voluminous (long and/or thick) abdomens. Moreover, P. libelluloides showing the highest number of follicles, 267–472 per testis (mean 347.4, SD 54.18), is known as the largest species in the antlion fauna of Russia: in males, the abdomen (together with ectoprocts) can reach 45 mm in length (Krivokhatsky 2011).

Working with aphids, Blackman (1987) introduced the possibility that the reduction of the number of testicular follicles is related to the transfer of spermatogenesis to early larval stages. Our observation is that in the neuropteran males, meiosis does generally occur in the very young individuals. Nevertheless, this is equally true for species with both high and low number of follicles.

Thus, our study resulted in the discovery of many novel values of the follicle number in Myrmeleontiformia. Our results also indicate that the limits of the variation in the follicle number are very broad, much more than previously known both for this group and the order Neuroptera as a whole. However, the mechanisms underlying the diversity of the follicle number in the group must be explored further.

In all hitherto studied species of the families Myrmeleontidae (ten species, eight genera), Ascalaphidae (two species, two genera), and Nemopteridae (one species), ovaries were found to consist of ten ovarioles each (Brauer 1854; Stitz 1909; Quartey and Kumar 1973; Garbiec and Kubrakiewicz 2012; present paper). The same ovarian structure was considered the most characteristic of the Neuroptera as a whole (Garbiec and Kubrakiewicz 2012). It should be noted, however, that the number of ovarioles per ovary is known to vary significantly (8–20) within the green lacewing family Chrysopidae (the suborder Hemerobiiformia) and even in the species of the genus Chrysopa Leach, 1815 (Vacacela et al. 2017).

The number of ovarioles in ovaries does not show a correlation with the female abdomen size, at least in the families studied here, considering that the same ovary structure was observed in all species, including relatively large Acanthaclisis occitanica. However, relevant data on other large species, such as Palpares libelluloides, are absent.

Large female insects were postulated to tend having comparatively higher potential fecundity (Berger et al. 2008). Moreover, the number of ovarioles per ovary is suggested to play a direct role in the number of eggs produced by a female and largely determines potential fecundity of an insect (Sarikaya et al. 2012). Even so, we can infer that in the neuropteran females the number of ovarioles is not the key factor responsible for their fecundity. An interesting avenue for future research would be to test whether there is a correlation between the size of ovarioles and the number of oocytes produced by these ovarioles, on the one hand, and the fecundity, on the other hand, in neuropteran females. Hodin (2009) suggested that plasticity in the number of ovarioles might be more prevalent for organisms living in fluctuating environments. In Myrmeleontidae and Ascalaphidae, the larvae are known to be underground inhabitants, mostly psammophilous (Mansell 1999). It is conceivable, that there is a certain relationship between this lifestyle and the stability of the ovary structure. In this context, the above-mentioned variation of the number of ovarioles in chrysopids (Vacacela et al. 2017) may be related to the open lifestyle of their larvae.

We have shown that lacewings, at least in the families Myrmeleontidae, Ascalaphidae, and Nemopteridae, are highly conservative in the number of ovarioles in females, and, conversely, highly diverse in the number of testicular follicles in males. It is assumed that testes and ovaries of insects were initially arranged by segments, and they fell into a common seminal duct or oviduct, respectively, on each side of the abdomen. The evolutionarily initial number of follicles per testis as well as ovarioles per ovary is therefore suggested to be seven as the number of the pregenital segments in adult males and females (Sharov 1966; Emeljanov et al. 2001). Since all so far studied species of the Myrmeleontiformia have ten ovarioles per ovary, it is likely that this pattern is a characteristic feature of the suborder as a whole. This character gives thus additional support for the monophyly of Myrmeleontiformia substantiated both by morphological and genomic data (Aspöck et al. 2001, 2003; Badano et al. 2016; Song et al. 2019). Taken seven ovarioles per ovary as an ancestral state of this character, the occurrence of ovaries with ten ovarioles might represent a result of a simple polymerization.

Despite a great variability of the follicle number among Myrmeleontiformia, at least in Myrmeleontidae and Ascalaphidae, the majority of studied species have relatively low numbers, three to eleven, with an apparent mode of 6–5 follicles per testis. Moreover, only these two numbers were discovered in the family Nemopteridae considered a sister-group of the [Myrmeleontidae + Ascalaphidae] clade (see for review Song et al. 2019). Taking into account that no testes with seven follicles are presently known in Myrmeleontiformia (likewise in Neuroptera as a whole), either six or five follicles per testis could be considered a plesiomorphic condition in Myrmeleontiformia. At present, it is tempting to believe that the testis with six follicles represents the ancestral character state, since it is the only pattern encountered in each of the three families explored. Both lower and higher numbers are, thus, the derived traits in Myrmeleontiformia. In the evolution of the families Myrmeleontidae and Ascalaphidae, the processes leading to increasing the follicle number were intensive. In these families, the number of follicles has increased up to 40–56 in some species (Palpares sp., Acanthaclisis occitanica, Bubopsis hamata) and further still has dramatically increased up to several hundred in Palpares libelluloides. It is of interest that the species of the genus Palpares have significantly different number of testicular follicles even though it is high in both cases. We can speculate, based on the currently available data, that the increasing of the follicle number in Ascalaphidae and Myrmeleontidae is the result of homoplasy, and the polymerization of the follicles has thus independently occurred in these two families (Fig. 1).

Figure 1. 

The proposed ancestral numbers of follicles per testis (white circles) and ovarioles per ovary (black circles) and limits of their variation across the families Myrmeleontidae, Ascalaphidae, and Nemopteridae (Myrmeleontiformia)