In Morchella species, mature ascospores are typically elliptical (Clowez et al., 2014, 2015; Taşkin et al., 2015; Loizides et al., 2016; Du et al., 2019), however, various ascospore shapes were observed among the specimens of M. galilaea (Figure 3). Elliptical ascospores dominated and accounted for 96.1% (total 845 spores investigated from the five samples, detailed information shown in Table 1), but irregular shapes, including spherical, heart-shaped, crescent, and other forms were also observed in all investigated fruiting bodies (Table 1, Figure 3). In Figure 4, the diversity of ascospore shapes and sizes is illustrated using the same scale for each element. Diversity of ascospore shape was not only found in different asci from the five samples (Table 1), but also from the same ascus (Figures 3W–Z,A). Certain asci harbored spores that were almost all spherical (Figure 3Ä) but several different spore shapes also were evident within a single ascus (Figures 3W–Z). The size of typically elliptical ascospores was (11.7-)16.8–22.9(−25.5) × (8.0-)9.6–14.5(−16.1) μm (average Q = 1.66) based on data from 200 ascospores of the five specimens (40 from each), whereas differently shaped spores were (7.9-)10.0–24.2(−29.1) × (5.8-)6.8–13.4(−19.6) μm (average Q = 1.82) based on observation of 236 spores from five specimens (13/ FCNU1061, 13/FCNU1116, 13/FCNU1117, 183/FCNU1118 and 14/FCNU1119).

Ascospore shape has been suggested to correlate with ecological niche and was shown to have an influence on the efficiency of active spore discharge (Ingold, 1975; Raju and Burk, 2004; Campbell et al., 2006; Shenoy et al., 2006; Hirooka et al., 2012). Most explosively ejected spores in ascomycetes have a drag-minimizing ellipsoid shape close to the theoretical optimum, which maximizes the range of their active liberation (Roper et al., 2008). Successful liberation of spores from ascomata facilitates dispersal. The irregular ascospore morphologies observed, such as spherical and heart shapes, seem to present difficulties for ascospore discharge compared to elliptical shapes. It is unknown whether these variant ascospore shapes are aberrations or have some unidentified function. The formation of moon– and heart-shaped ascospores in M. galilaea appears to arise from incomplete spore delimitation between two spores. Whether these variant spore shapes have to do with the autumn fruiting of M. galilaea remains to be studied. In further study, we will focus on whether these variant ascospores germinate and try to discover whether there is a potential function associated with these shapes.

Species of Morchella have previously been shown to harbor eight ascospores per ascus (Clowez et al., 2014, 2015; Taşkin et al., 2015; Loizides et al., 2016; Du et al., 2019), but we found that the number of ascospores produced in asci of M. galilaea varied from one to 16 (Figure 5). The number of ascospores produced in asci studied are shown in Table 1. Mature ascus length varied from 215.6 μm (ascus with seven ascospores) to 364.4 μm (ascus with two ascospores), and the width ranged from 16.5 μm (ascus with two ascospores) to 23.87 μm (ascus with eight ascospores) (Supplementary Table S2). The ascus in this study ranged from 215.6–364.4 × 16.5–23.9 μm. This was greater than previously determined for M. galilaea, namely 165–220 × 15–22 μm (Taşkin et al., 2015). This ascus size difference may be attributed to the higher number of ascospores that were identified in this study (see below in 3.2.3). Asci and ascospores illustrated using the same scale make clear the variations of shape and size for the sixteen types observed (Figure 6).

We further counted the number of asci with different numbers of ascospores (Figure 7, Supplementary Table S3). Their distribution pattern showed that asci with eight ascospores were predominant and accounted for 540/654 (82.6%) of the asci, which is consistent with the typical number of ascospores for most ascomycetes. Seven-spored and nine-spored asci were the second most numerous and were observed equally in 17/654 (2.6%) of asci. Asci that harbored 16 ascospores occurred least frequently (1/654; 0.15%). The second lowest prevalence was asci having one, 14, or 15 ascospores which each accounted for 2/654 (0.3%) of the total. Asci with other numbers of ascospores comprised 11.2% of the total (Supplementary Table S3).

In ascomycetes, species with varying numbers of spores per ascus were previously reported (Yao and Spooner, 2000; Bonito et al., 2010; Kušan et al., 2018; Pfister and Healy, 2021). However, variation in spore number within asci from the same species and from the same fruiting body are more rarely reported. Variation in spore number in yeast asci was suggested to depend closely on the carbon-to-nitrogen ratio of the growth medium, with more spores produced in media with a high carbon-to-nitrogen ratio, and fewer spores with a low ratio (Okuda, 1961). However, the carbon-to-nitrogen ratio is unlikely to be a factor in this study since spore numbers varied not only among different fruiting bodies, but also within a single fruiting body of M. galilaea. Variation in ascospore number has been suggested to be related to the lack of coordination of meiotic divisions, spore wall formation and successful mitotic divisions. Varying propagule number may contribute to differences in dispersal range, colonization of new habitats, or the production of heterokaryotic ascospores for pseudohomothallic species (Davidow and Goetsch, 1978; Sherwood, 1981; Hanlin, 1994; Raju and Perkins, 1994; Réblová and Mostert, 2007; Quijada et al., 2022).

In Morchella species, heterothallism, pseudohomothallism and unisexual reproduction modes have been reported not only from different species but also from the same species, such as M. importuna (Chai et al., 2017, 2019; Du et al., 2017, 2020; Du and Yang, 2021). To date, only heterothallism has been found in M. galilaea (Chai et al., 2019; Du et al., 2020; Du and Yang, 2021), given the variation in spore number in asci and thus the potential for heterokaryotic ascospores formation (especially in those cases of reduced spore number), we presume pseudohomothallism may exist in this species. This awaits verification. As the only species of Morchella fruiting in autumn and its wide distribution in Asia, Europe, North American and Africa, the variation in ascospore number in M. galilaea might be beneficial in increasing the numbers and diversity of propagules and introduce the potential for the formation of heterokaryotic ascospores aiding in dispersal and long-distance colonization.

In the Pezizomycetes, the number of nuclei in mature ascospores is considered to have taxonomic value, and varies from 1, 2, 4 to many (Pfister and Healy, 2021). Using laser confocal microscopy we observed the nuclear condition of ascospores in the asci with the typical eight ascospores, and atypical one, two, three, four, five, seven, 11, 13, and 14 spores (Figures 8A–M). For the typical eight-spored asci, usually six to ten nuclei were observed in each ascospore (Figure 8J). In the one-spored ascus (Figure 8A), only four nuclei were observed, indicating the absence of mitosis and/or nuclear degeneration. Although, the possibility also existed of meiosis with cytokinesis defects producing one cell with four nuclei. Two two-spored asci, respectively, possessed six (Figure 8B, three in each spore) and eight nuclei (Figure 8C, six in the large ascospore, one in the small one and one in the cytoplasm). Four nuclei were observed in a three-spored ascus, with two nuclei in a large ascospore and one in each of the small spores (Figure 8D). Four nuclei total in a four-spored ascus were observed, with two nuclei in the large ascospore, two in two other spores, and one spore aborted without a nucleus (Figure 8E). In a five-spored ascus, four to eight nuclei were observed in each spore (Figure 8F). In seven-spored asci (Figures 8G–J), a minimum of four nuclei were observed in each spore, and one moon- and one heart-shaped spore (Figures 8H,I) had almost double the numbers of nuclei (around 16 to 20) compared to typical ascospores. A failure of spore delimitation and cytokinesis defects during mitosis may be responsible for the formation of moon- and heart-shaped spores that otherwise might have divided into pairs of ascospores.

Spore abortion has been suggested to be one of the reasons that fewer than eight spores are found in some asci (Booth, 1959; Rossman, 1983; Glawe and Jacobs, 1988). Spores without nuclei usually fail to reach full maturity or abort after they have been delimited (Dodge, 1928). Based on nuclear observations of two-, three-, and four-spored asci above, larger ascospores usually possess more nuclei than smaller ones (Figures 8C–E). Sherwood (1981) reported that larger spores potentially are multinucleate and may have a heterokaryotic advantage. This may aid in colonization and enhance propagule survival. In pseudohomothallic species, fewer spores and larger ascospores are produced (Raju and Perkins, 1994). The production of multi-nucleate heterokaryotic ascospores gives an advantage in establishment and completion of a species’ life cycle (Uhm and Fujii, 1983; Raju and Perkins, 1994). The pseudohomothallic condition has been demonstrated in Morchella species, such as in M. sextelata, M. importuna, and Morchella sp. Mes-15 (Liu et al., 2021), but not in M. galilaea yet.

For asci with more than eight spores, visualization with laser confocal microcopy requires observations at different angles in order to see all the nuclei. We cannot show a single image in which all nuclei are visible. Therefore, we state how many nuclei were observed in the spores of 11-, 13-, and 14-spored asci, but the corresponding images in Figure 8 may not clearly show the exact number of nuclei. In the 11-spored ascus, one large ascospore harbored eight nuclei, two contained three nuclei, one had two nuclei, six harbored only one nucleus in each and one aborted without nuclei (Figure 8K). For the 13-spored ascus, eight ascospores contained four to six nuclei and the other five spores harbored at least one nucleus in each (Figure 8L). In the 14-spored ascus (Figure 8M), two spores contained six nuclei, two contained four nuclei, one had two nuclei, and one harbored one nucleus, but the other seven aborted spores lacked nuclei even at different observational angles.

In contrast to asci with more or fewer than eight spores, eight-spored asci in M. galilaea have more nuclei per spore, usually six to ten and even more. Ascospores with no nuclei (aborted) or fewer than six nuclei were commonly found. In ascomycetes, the fusion nucleus in the ascus regularly divides in succession by meiosis and post-meiotic mitosis to produce eight nuclei before spore delimitation for typical eight-spored asci (Dodge, 1928; Pöggeler et al., 2006); for asci with more than eight spores, the eight post-mitotic nuclei undergo further successive mitotic divisions so that 16, 32, or more nuclei are formed before spore delimitation (Dodge, 1928). Spore fragmentation to form part spores and budding of ascospores within asci may be also responsible for more than eight spores being found in asci in some fungi (Read and Beckett, 1996). When there are fewer than eight ascospores per ascus, nuclei degeneration followed by spore abortion, an absence of mitotic divisions, or multiple nuclei encompassed by a single wall to form a giant, multinucleate spore may be the reasons (Esser and Stahl, 1976; Leonard, 1988; Read and Beckett, 1996), as we observed in one-, two–, three– and four-spored asci in M. galilaea. Among all the different asci observed in M. galilaea, a minimum of four nuclei in developing asci was observed, except the diploid state with a single nucleus present, and nuclear number in spores can vary from zero to around 20. The four nucleate states observed in some asci suggest that meiosis occurs, but whether post-meiotic mitosis follows needs to be verified. Further studies of meiotic and mitotic events correlated with nuclear status in asci and spore delimitation are needed in the future.

A negative correlation exists between ascospore size and the number of ascospores per ascus in most ascomycetes, i.e., production of fewer than eight ascospores per ascus is correlated with a maximum spore size whereas having more than eight ascospores is correlated with a reduced spore size (Sherwood, 1981). Ascus size was also shown to correlate with ascospore size in the Nectriaceae (Hirooka et al., 2012). To probe the relationships among ascospore number, ascospore dimension and ascus size, we measured the length and width of ascospores in one- to 16-spored asci observed from M. galilaea, as well as the length and width of the asci (Supplementary Table S2).

Negative linear correlations were apparent in a scatter diagram for length and width of ascospores within asci and ascospore number (r = −0.517 for ascospore length, r = −0.153 for ascospore width) (Figure 9A); this agrees with previous observations (Sherwood, 1981). Coefficient of determination (R²) values indicated that 21.6% of the variability in ascospore length may be explained by ascospore number which also was responsible for 8.5% of the observed variability in ascospore width (Figure 9A). In contrast, positive linear correlations were indicated for the length and width of asci and ascospore number within the corresponding asci (r = 0.111 for asci length, r = 0.242 for asci width) in which the variable ascospore number influenced ascus width more than length (Figure 9B). Coefficient of determination (R²) values indicated that only 0.02% of the variability in ascus length was explained by ascospore number, which was responsible for 13.9% of the observed variability in ascus width (Figure 9B).

Apparent positive linear correlations were shown for ascus length and ascospore size (r = 3.463 for ascospore length, r = 3.148 for ascospore width, Figure 9, C1 and C2). Coefficient of determination (R²) values indicated that 29.4% of the variability in ascus length may be explained by ascospore length, and 5.8% may be explained by ascospore width. Weak positive linear correlations were shown for ascus width and ascospore size (r = 0.015 for ascospore length, r = 0.139 for ascospore width, Figure 9, C3 and C4). Coefficient of determination (R²) values indicated that 0.09% of the variability in ascus length was due to ascospore length, and 2% was related to ascospore width.

In M. galilaea, ascus size is influenced more by ascospore length than ascospore width, which is probably in accordance with the similar shapes of asci (cylindric) and ascospores (elliptical). The production of fewer than eight spores in asci increases maximum potential spore size. Large, compartmentalized spores have clear advantages in a harsh environment and also increase impaction efficiency and the amount of mycelium which could be produced (Sherwood, 1981). Large spores in fewer than eight-spored asci have more chance for survival, but it may impede active discharged from asci. Their generation size also decreases. The large number of spores produced in polysporous asci decrease spore size, but increase their generation size and also survival possibility of offspring even if some spores fail to survive. Both conditions have some adaptive values, but the production of eight-spored asci probably has the optimal value suitable for ascus discharge mechanism and the ascospore size needed to harbor nuclei and other organelles to survive.

Two types of ascospore germination were observed in M. galilaea. Ascospores germinate via a single, long germ tube or by the direct production of conidia from ascospores, or by both simultaneously. Ascospores may germinate within and outside asci. Production of conidia directly from ascospores without the typical intervening mycelial phase was investigated here from one freshly collected and four dried specimens, results are shown in Table 1. Among the five specimens (Table 1), conidial formation was observed not only in ascospores within living asci from the freshly collected specimen (Figure 10A), but also from ascospores within dead asci or after discharge from the three dried specimens (Figure 10B), however, no conidia were found arising directly from ascospores from the specimen from Kenya.

Ascospores in M. galilaea can germinate terminally at one or both spore poles, or laterally at multi-sites, and by formation of hypha and/or conidia singly or in sequence. The types of ascospore germination are shown in Figure 10. Mature ascospores may be ejected from asci with conidia attached (Figures 10B3 and 11C4). Once conidia mature, they may gradually detach from ascospores within (Figure 10A2,A9,B2,B4) or outside of the asci (Figure 10B14–B16).

Conidia were produced from about 0.01% of ascospores based on observations from three dried specimens (1/1016, 1/1029, 1/1015). The hyaline conidia in M. galilaea were globose or occasionally elliptical, ranging in diameter from approximately 2–10 μm (Figure 10). The ontogeny of conidial emergence from ascospores to the detachment of single and sequential conidia is illustrated schematically based on our microscopy observations (Figure 12). More morphological types of conidia formation are presented in Figure 4B2, C1-C7.

The nuclear status of ascospores at germination either by long germ-tubes (Figure 11A) or by production of conidia within or outside asci (Figures 11B,C) was further investigated. Migration of nuclei from ascospores through the bud necks into conidia was observed (Figure 11C2,C4–C6), but whether further nuclear mitotic divisions occur in conidia before detachment is unverified. One to six nuclei were observed from detached (Figure 11B1–B6) and attached conidia (Figure 11B8–C8), which was in sharp contrast to 14 nuclei noted in the ascospore (Figure 11B7). Nuclear migration through the bud neck from the primary conidium into successive conidia also was noted (Figure 11B14,B15).

In species of the genus Morchella, conidia are produced on upright, aerial conidiophores after a period of vegetative growth during their hyphal development (this state was previously identified as Costantinella). These conidia are spherical and 2.5–8 μm in diameter, and harbor one to three nuclei (Carris et al., 2015; Yuan et al., 2021; Pfister et al., 2022). The size, nuclear status and formation of conidia produced directly from ascospores are similar to conidia that are produced from mycelium. The similar morphology of conidia produced directly from spores and from aerial hyphae was also described in Nectria inaurata and was found in other fungi (Brefeld, 1891). These common features suggest that development of both modes of conidial formation differ primarily in the timing of conidial initiation in the life cycle, and thus further indicate the two kinds of conidia may have similar functions, such as serving as colonial propagules or as spermatia.

Among the five samples of M. galilaea used in this study, four collections from China exhibited conidia produced from ascospores, whereas the sample from Kenya contained mature ascospores, but their ascospores did not produce conidia. Phylogenetic analyses supported all the five samples as M. galilaea with high bootstrap (100%) (Figure 2). The current results suggested that conidial production from ascospores seems not to be a useful taxonomic feature in M. galilaea, but new markers and additional samples in further research on this species will likely provide more definitive information. That the sample from Kenya lacks conidia produced from ascospores may suggest emerging morphological difference that are due to geographic isolation (Figure 2) but caution is warranted since only a single sample from Kenya was used in this study.

Conidia produced directly from ascospores have rarely been reported in the Pezizomycetes, but have been noted in P. subisabellina (Van Vooren, 2020) and members of the Sarcoscyphaceae (Paden, 1975). Morchella galilaea is the first species reported in the Morchellaceae to produce conidia directly from ascospores. Given that seemingly no special conditions are required during conidial production directly from ascospores, it is likely that this phenomenon is widely distributed but overlooked in fungi. Our findings call for more research into investigating conidial production directly from meiospores.

Production of ascoconidia and conidia directly from ascospores often correlate with ecological niche (Ingold, 1975; Campbell et al., 2006; Shenoy et al., 2006). Such development has been suggested to be associated with drought-tolerance (Sherwood, 1981); colonization of new nutrient sources for coprophilous fungi (Quijada et al., 2022); infection of new hosts in harsh environments (Juzwik and Hinds, 1984; Hanlin, 1994) and performing the role of spermatia (Hanlin, 1994). The broad distribution of M. galilaea throughout Asia, Europe, North America, and Africa, and the autumn fruiting of the species (Du et al., 2015; Taşkin et al., 2015), suggests strong selective pressures from environmental and habitat factors that may favor rapid dispersal and colonization by means of diverse and numerous propagules. These propagules may serve as spermatia and be involved in sexual reproduction during the special fruiting timeline of this fungus. It is important to note that ascomata of species of Morchella are primarily produced in the spring, but their anamorphic spore mats are mainly produced in the fall. These spore mats were not only reported from the Northern Hemisphere (Carris et al., 2015), but also in the Southern Hemisphere (Pfister et al., 2022). So far, field-collected anamorphs have not been reported for M. galilaea. The function of conidia produced from ascospores in M. galilaea still requires further investigation.