1. Introduction
The subfamily Conoderinae, Schoenherr, 1833, formerly long known under the commonly accepted junior name Zygopinae Lacordaire, 1866, has been quite recently applied, as the oldest available name, to a much larger group of weevils, including the former subfamilies Baridinae, Ceutorhynchinae, and Orobitidinae, and nowadays comprises over 7500 described species in 940 genera. The former four subfamilies were reduced in rank to supertribes, among which Conoderitae is the second largest, with approximately 2200 species in 212 genera worldwide [
1]. However, Conoderitae seems to be the least known taxonomically among the four conoderine supertribes, and the real number of extant species must be several times higher. They are abundant and diverse in samples from rain forest canopies. For example, at least 520 species were counted during an inventory at La Selva Biological Station in Costa Rica by H. Hespenheide [
2], the number updated to 559 species until 2009 (H. Hespenheide, personal communication).
While common and hyperdiverse in tropical and subtropical zones, the group is highly under-represented in the Western Palaearctic, in fact by just a single tribe—Coryssomerini—with two genera and four species [
3,
4]. The strictly Western Palaearctic genus
Coryssomerus consists of
C.
capucinus (Beck, 1817), widespread in Europe and the Caucasus, along with two much less well-known North African species. In the second genus,
Euryommatus Roger, only its type species
E.
mariae Roger, 1857 was originally described from Europe, while eight further species currently attributed to
Euryommatus live in Eastern Asia, Asia Minor or the Arabian Peninsula [
3,
4]. The only European species
E. mariae ultimately turned out to be a Euro-Siberian element, relictual in Europe, but widely distributed throughout Siberia to the Russian Far East, eastern China (the first record here), the Korean Peninsula, and Japan [
3,
4,
5,
6,
7].
In Europe,
Euryommatus mariae has always been a mysterious and indeed legendary weevil. Described by Roger [
8] from a locality at Rudy near Kuźnia Raciborska (Upper Silesia—the type locality, now in Poland), it was not found again in Silesia or anywhere else in the whole of Poland for over a century. Subsequent records from Europe have been very scarce: All come from the 19th century and only from Austria (Salzburg vic.) [
9], and Latvia [
10]. The occurrence of
E.
mariae in Poland was finally confirmed after 140 years by Stachowiak [
11], who collected a single specimen in the Białowieża Primeval Forest (NE Poland). In the 21st century, the species has been found at several new localities in central Europe. In Poland, further specimens were collected in the Świętokrzyskie Mts in central Poland, near the village of Cisów [
12]. Recently, the weevil was recorded in southern Bavaria in Germany [
13]. At the same time, it was rediscovered in Austria, in the East Tyrol exclave [
14].
Euryommatus mariae was also recorded in Slovenia, European part of Russia, and Cyprus in recent Palaearctic catalogs [
3,
4], but these country records are not supported with literature data. Moreover, the latter record is apparently a misidentification, since just an unidentified species of
Euryommatus is listed from Cyprus in the recent catalog of Curculionoidea of this island [
15]. The occurrence of
E. mariae on this Mediterranean island seems rather unlikely.
Unlike
Coryssomerus capucinus, which develops on several herbaceous species of the Asteraceae, little is known about the host associations of
Euryommatus mariae and its congeners. In Austria, it was beaten from ‘Pinus alba’ by Sartorius [
9], and probably on this basis, Gerhardt [
16,
17] presumed its association with fir
Abies alba Mill. In Siberia, it develops in dying branches of
Abies sibirica Ledb. [
18], but the adults were also collected from
Pinus sylvestris L. (Legalov, personal communication). In Germany, the weevil was also collected from pine branches, albeit of
Pinus mugo subsp.
uncinata (DC.) Domin. In turn, all 28 Polish specimens of
E. mariae, collected in the forest near Cisów, were obtained using spiral screen-trunk traps (‘Geolas’ trap) deployed on the trunks of spruce
Picea abies (L.) H. Karst. This may indicate an oligophagous type of host association within the family Pinaceae.
Being familiar with the former sampling area in Cisów, we undertook an investigation there to find immature stages of Euryommatus mariae in branches from a dead and fallen spruce Picea abies. Laboratory examination of these branches in March 2012 revealed a pupa, found by M. Bidas, who reared it to the adult. The site was revisited in the same month to obtain further material. Here, we give the first morphological description of the larva and pupa of E. mariae obtained from the above-mentioned spruce branches. The larval characters of Euryommatus are compared with available descriptions of some Nearctic species of Conoderitae, including the conifer-associated Cylindrocopturus furnissi Buchanan.
2. Materials and Methods
The immature specimens used in this study: A premature (probably fourth) instar larva (one ex.), a mature larva (2 exx.), and pupae (♀)(2 exx). All were collected from one site: Cisów, forest district Daleszyce, forest comp. 144, 50.74N/20.86E, 264 m alt., 22.03.2012, branches cut from a fallen spruce trunk, leg. M. Wanat and M. Bidas.
The larvae and pupae are deposited in the collections of the Department of Zoology and Nature Protection, Maria Curie-Skłodowska University (Lublin, Poland).
The ten adult specimens examined are all from the collection of M. Wanat, stored at the Museum of Natural History, University of Wrocław: Poland: Świętokrzyskie Mts: Cisów, o. 144, UTM: DB82, 7–BMw G.Sw. 2, 29–31 VIII 2006, screen traps on spruce, 2 ♂ 3 ♀, leg. J. Borowski; same locality, reared ex dead spruce branches, imagines emerged 6 II 2008 (1 ex.), 17 I 2012 (1 ex.), 1 III 2012, 1 ♂, leg. M. Bidas. China (East): Harbin, 1 ♀. Russia: W Siberia: Altai, Teletzkoe Lake, 12 VII 1975, 1 ♂ 1 ♀, leg. F. I. Opanassenko.
2.1. Habitat
The locality near Cisów is a part of large, forested areas within boundaries of the Cisowsko-Orłowiński Landscape Park, undergoing sustainable forest management. It is a natural mixed humid coniferous forest, with dominant pine and spruce, and a small admixture of deciduous trees, mainly birch, alder, and oak (
Figure 1A). The forest floor is mossy, and the undergrowth consists primarily of spruce saplings. The site is situated very close to the open peat bog area of the Białe Ługi Reserve. The fallen spruce sampled for
Euryommatus immatures was a dead tree of medium age, with its side branches still largely covered with loose bark (
Figure 1B); hence, it must have lain there for months rather than years, being nearly or quite dead when it fell. Several branches about 3–4 cm in diameter were sawn into pieces and taken to the laboratory in Wrocław. The branches were examined by carefully removing fragments of bark, under which a number of short and simple galleries could be seen. Four galleries contained single larvae, and in one, a pupa of
Euryommatus was detected in a terminal elongate-oval and only slightly deepened pupal chamber (
Figure 1C). The largest larva was left alive in its gallery, where it pupated a few days later. All the immature stages were preserved for morphological studies.
2.2. Methods
2.2.1. Morphological Studies—Immature Specimens
All the specimens described were fixed in 95% ethanol and examined under an optical stereomicroscope (Olympus SZ 60 and SZ11) with calibrated oculars. The graduation was scaled up to 1/10 mm. Measurements were made under a 30× magnification. The following measurements of larval instars were made: Body length (BL), body width (BW) (at the third abdominal segment), and width of the head capsule (HW). The pupal measurements included body length (BL), body width (BW) (at the level of the mid legs), head width (HW) (at the level of the eyes), length of rostrum (RL), and width of pronotum (PW). The drawings and outlines were made using a drawing tube (MNR–1) installed on a stereomicroscope (Amplival Pol-d, Carl Zeiss Jena, Germany). and processed by computer software (Corel Photo-Paint X7, Corel Draw X7 (Corel Inc. Austin, TX, USA).
Slide preparation basically follows May [
19]. The larvae selected for study under the microscope were cleared in 10% potassium hydroxide (KOH), then rinsed in distilled water and dissected. After clearing, the head, mouthparts, and body (thoracic and abdominal segments) were separated and mounted on permanent microscope slides in Faure–Berlese fluid (50 g gum arabic and 45 g chloral hydrate dissolved in 80 g of dissolved water and 60 cm
3 of glycerol) [
20].
The photographs were taken using an Olympus BX63 microscope and processed with Olympus cellSens Dimension software. The larvae selected for SEM imaging (scanning electron microscope) were first dried in absolute ethanol (99.8%), then rinsed in acetone, treated by CPD (Critical Point Drying), and finally gold-plated. TESCAN Vega 3 SEM was used to examine selected structures. The general terminology and chaetotaxy follow Anderson [
21], May [
19], Marvaldi [
22,
23,
24,
25], and Skuhrovec et al. [
26]; the antennae terminology follows Zacharuk [
27].
2.2.2. Morphological Abbreviations
Abd. I–X—abdominal segments 1–10, Th. I–III—thoracic segments 1–3, at—antenna, clss—clypeal sensorium, ds—digitiform sensillum, st—stemmata, Se—sensorium, sa—sensillum ampullaceum, sb—sensillum basiconicum, snp—sensillae pore, ss—sensillum styloconicae, tra—terminal receptive area, lr—labral rods, ur—urogomphus; setae: als—anterolateral, ams—anteromedial, as—alar (larva), as—apical (pupa), cls—clypeal, d—dorsal (pupal abdomen), des—dorsal (larval head), dms—dorsal malar, ds—discal (pupal prothorax), ds—dorsal (larval abdomen), eps—epipleural, es—epistomal, eus—eusternal, fs—frontal, les—lateral epicranial, ligs—ligular, lrs—labral, ls—lateral, lsts—laterosternal, mbs—malar basiventral, mds—mandibular, mes—median, mps—maxillary palp, pda—pedal, pds—postdorsal, pls—posterolateral, pes—postepicranial, pfs—palpiferal, pms—postlabial, prms—prelabial, prns—pronotal, prs—prodorsal, ps—pleural, sls—super lateral, sos—superorbital, ss—spiracular, stps—stipal, sts—sternal, ves—ventral, vms—ventral malar, vs—vertical. HW—head width, BL—body length, BW—body width, RL—rostrum length, PW—pronotum width.
2.2.3. Morphological Studies—Adult Specimens
Images of adult specimens were taken with a Leica M205C stereomicroscope and attached digital camera JVC KYF75. The images obtained were combined using the AutoMontage software of Syncroscopy (Cambridge, UK), and enhanced using Adobe Photoshop CS2 program. The aedeagus was photographed under transmitted light using the same equipment, while the details of the conduit of the ejaculatory duct were illustrated using a Nikon Eclipse Ni compound microscope with attached Nikon D7500 camera, and the stacks were combined using the Helicon Focus software (ver. 7.5.1 Pro, Helicon Soft Ltd., Kharkiv, Ukraine).
Genitalia preparations were made according to the standard method after a maceration of the separated abdomen for 5–10 min in hot KOH solution. Membranous structures were stained in a glycerol solution of chlorazol black. After rinsing in distilled water, further separation and examination of the terminal segments and genitalia were carried out in pure glycerol on a microscope slide, under both stereoscopic and compound microscopes. Only the spermatheca was first photographed in distilled water, to illustrate its natural shape, before the eventual collapse of its wall after transfer to glycerol. After the study, all the parts were stored in glycerol in a microvial pinned beneath the specimen, while the abdominal ventrites were glued onto a card adjacent to the specimen.
4. Discussion
The host associations and basic details of larval development have been identified in just one European species from each genus within the tribe Coryssomerini.
Coryssomerus capucinus lives on several herbaceous plants from the genera
Tripleurospermum Sch. Bip.,
Achillea L.,
Chrysanthemum L., and
Anthemis L. (Asteraceae); its larva develops in the root neck and pupates in the soil [
32,
33]. The biology of
Euryommatus mariae is quite different, however: It has saproxylic, subcortical larvae associated with gymnosperm trees. Such an association between conifers and saproxylic larvae is unknown in any other Palaearctic member of the supertribe Conoderitae. However, analogous host associations with coniferous trees and wood-boring larvae have been well documented in several Nearctic species of the genus
Cylindrocopturus, classified in the tribe Zygopini [
34], and distributed along Pacific coast of North America [
35]. According to Furniss and Carolin [
36], they attack the twigs and boles of various conifers, including pine (
Pinus L.), true fir (
Abies L.), Douglas fir (
Pseudotsuga Carrière), larch (
Larix Mill.), and hemlock (
Tsuga Carrière). The Douglas fir twig weevil
Cylindrocopturus furnissi Buchanan develops on
Pseudotsuga menziesii (Mirbel) Franco [
37], while
C.
eatoni Buchanan, known as the pine reproduction weevil, attacks primarily ponderosa and Jeffrey pines (
Pinus ponderosa Douglas ex C.Lawson,
P.
jeffreyi Balf.). Although associated with conifers in much the same way as
E.
mariae, the biology of these Nearctic
Cylindrocopturus species is significantly different. They both attack the branches and boles of saplings and young trees with green foliage and develop in living wood, their larvae often killing them by destroying the phloem and cambium, thus cutting off sap transport. The larvae of
C.
eatoni can be found in all woody parts of pine saplings, even in the rootstock some centimetres below ground level, but the top of the stem and the upper branches are infested much more abundantly [
38].
C. furnissi also attacks small living trees of Douglas fir, and in older ones it evidently prefers the previous four years’ growth for oviposition, frequently killing scattered small branches [
36]. In both
Cylindrocopturus species, the eggs are laid individually in holes chewed by the female in living bark—globules of resin produced by the tree reveal their locations. Their larvae bore galleries in cortical tissue, phloem or even pith, and often pupate there in thin twigs, while in thicker branches and the trunk, mature larvae usually work their way into the outer layer of wood for pupation. The different habits of
Euryommatus mariae can be inferred from the data collected at the Cisów site. There, the weevil seems to be confined to dead or dying spruce branches, presumably weakened and dying lower laterals, subsequently appearing on older trees that retain their bark intact for quite a long time after death. Unlike the American conifer beetles
Cylindrocopturus,
E.
mariae is a saproxylic species utilizing the outermost layer of dead or dying wood.
Apart from this basic difference in diet, many other aspects of the biology of
E.
mariae appear to resemble that of the above-mentioned
Cylindrocopturus species. Their development takes a year with a winter break [
36,
37,
38], an aspect that also appears to hold true for
E.
mariae, where the larva is the diapausing stage. In its phenology, too,
E.
mariae seems to follow
C.
furnissi in particular. This latter species starts to make feeding holes in the bark of small Douglas fir branches, mating, and laying eggs after mid-June, with adults leaving the pupal chambers by the beginning of August [
37]. The emergence of the adults of
C.
eatoni begins about a month earlier, already in late May, and peaks in mid-June, when the weevils can be found feeding abundantly on pine needles [
38]. Although the larvae collected in Cisów on 22 March pupated in the same month, and one adult emerged on 1 March from material preserved early in winter, they were all reared under laboratory conditions at room temperature, which could have speeded up metamorphosis considerably. The numerous adults of
E.
mariae obtained by Rutkiewicz [
12] from the screen-trunk traps deployed in Cisów are all dated to the end of August. Moreover, the recent collections of this species took place in mid-August in Germany [
13], and in mid-July in Austria [
14].
The morphology of immatures of species from the tribe Coryssomerini remains poorly studied. Only the paper by Urban [
39] contains some basic, unillustrated information about the larva and pupa of
Coryssomerus capucinus (Beck, 1817), subsequently published by Scherf [
32]. Urban [
39] highlighted the larval cuticle densely covered with hook-like asperities, six variously long setae on each of the pedal lobes, the two-segmented labial palpi, and bifid mandibles. Unfortunately, other characters are either very common or were described inaccurately. But it is noteworthy that the larval cuticle of
Euryommatus mariae exhibits a very similar structure to that of the larva of
C. capucinus. Urban’s [
39] description of the pupa of
C. capucinus did not contribute any significant information.
Hinz and Müller-Schärer [
40] reported five larval instars of
C. capucinus based on the width of the head capsule. According to these authors, the head width of the premature larva is 0.70 mm, while in the mature larva it is 0.93 mm [
40]. Both values resemble the head measurements of
E. mariae (0.70 mm and 0.95 mm). From this, one may presume that the larval development of
E. mariae also involves five instars. Despite the highly precise measurements, the paper by Hinz and Müller-Schärer [
40] does not contain any other information about the larval morphology. Thus, the descriptions of the larva and pupa of
E. mariae given here are the first complete, illustrated information on immatures of the tribe Coryssomerini.
It seems that the immatures of most genera belonging to the subfamily Conoderinae Schoenherr have yet to be described. The only exceptions are a few species from the tribe
Zygopini (regarded as pests of silviculture): The genus
Cylindrocopturus—C. crassus van Dyke, Keifer [
41],
C. quercus (Say) [
42], and
C. furnissi Buchanan [
43], and a single species from the genus
Eulechriopus—E. gossypii Barber Böving [
44]. Analysis of these papers is sometimes difficult, however, because of the different nomenclature used in them and the lack of descriptions of certain structures: For instance, the abdominal setae of
E. gossypii, according to Böving [
44] are extremely small and impossible to count accurately.
Compared with the larvae of other Conoderinae species, the larva of E. mariae reveals some important, original features, different from Cylindrocopturus, and in some cases, Eulechriopus, such as: (1) Single as (vs. two as); Abd. VIII with three pds (vs. two ds); Abd. IX with three ds and three ps (vs. two ds and two ps); (3) head rounded (vs. narrowed bilaterally); (4) abdominal segments dorsally with transverse rows of elongate plate-like asperities, each parallel to long body axis and with anterior and posterior denticle (vs. only the simple, thorn-like asperities); (5) erect setae on thorax and abdominal segments VIII and IX very long, evidently longer than on the remaining abdominal segments (vs. setae on all body segments subequally short).
The best visible common characters of the genera Cylindrocopturus and Euryommatus include the extended head; the single pair of ocelli placed close to the antenna; the frons with five setae; the mala with seven dms, and five vms; the clypeus with very long setae; the structure of the spiracles; the dorsal folds of the abdominal segments I–VII divided into three lobes; the well-developed abdominal setae; the prothorax with 10 (11 on C. furnissi) prms; each pedal lobe with six pda; the meso- and metathorax each with one prs and four pds; each of abdominal segments I–VII with one prs, five (four on C. quercus) pds, two ss, two eps, and two ps.
On the other hand, the larva of
E. gossypii displays many original characters, different from both
Cylindrocopturus and
Euryommatus, above all the strongly retracted head and the structure of the last abdominal segments closely resembling the type “B” described by van Emden [
45], found, for example, on
Tanymecus,
Strophosoma and
Philopedon [
46]. Moreover, the larva of
Eulechriopus has the dorsal folds of abdominal segments I–VII divided into two lobes, two pairs of ocelli, and the clypeus without setae [
44].
The larval characters common to all Conoderinae genera are (1) endocarina present; (2) conical, moderately elongated antennal sensorium; (3) epipharynx with three als, three ams, and two mes; (4) labral rods very elongated, slightly converging posteriorly; (5) bifid mandible; (6) two-segmented labial palpi; (7) dorsal part of the body densely covered with asperities; (8) all spiracles bicameral.
Based on head width measurements of
C. quercus, Piper [
42] reported three larval instars: The first 0.18–0.19 mm, the second 0.23–0.24 mm, and the third 0.50–0.72. Hence, GF takes a value of 1.35 between the first and second instars (1.32 on
C. capucinus), but 2.60 between the second and third instars (1.41 on
C. capuccinus). Moreover, GF measured between the smallest and largest larvae, assessed by Piper [
42] as third instars, takes a value of 1.44, which is similar to the GF estimated between the fourth and fifth instars of
C. capucinus [
40]. Hence, the measurements performed with the GF proposed by Dyar [
47] indicates (most probably) the existence of five larval instars in
C. quercus.
According to Böving [
44], the pupae of
E. gossypii and
C. crassus are visibly similar in shape; this may also apply to the pupae of
C. furnissi [
43],
C. quercus [
42], and in some ways to
E. mariae, mainly because of the great morphological similarity of the adult stages of those species. However, the lack of accurate drawings and precise nomenclature of the chaetotaxy does not permit any far-reaching inferences to be drawn. Anderson [
43] described the urogomphi of
C. furnissi as being placed laterally, “slightly sclerotized at the tip and terminating in two or three minute projections”. It is hard to say whether the structure described by Anderson [
43] is something that corresponds to the urogomphi of
E. mariae. The pupa of
E. mariae has quite elongated urogomphi with anchor-like apices, placed dorsomedially, which is rather uncommon in weevil pupae and well visible. Thus, it does not seem possible that such a crucial feature has been overlooked in the descriptions of
Cylindrocopturus and
Eulechriopus. Ultimately, however, it is impossible to distinguish any features of taxonomic importance for Conoderitae, based on existing descriptions of these pupae.
The study of the adult postabdomen has yielded several morphologically and evolutionarily noteworthy discoveries. The unusually long ejaculatory duct protected by a fibrous sheath is a unique character of
E.
mariae, not found in its congeners available for this study, or in
Coryssomerus from the same tribe. However, an unidentified
Euryommatus sp. from Fethiye in southern Turkey (coll. M. Wanat) possesses in its endophallus a flagellum-like pipe of similar fibrous structure, though incomparably shorter, not exceeding the tips of the penile apodemes. On the other hand, the ejaculatory duct in another
Euryommatus sp. from the Myohyang Mts. in North Korea (coll. Museum and Institute of Zoology, Polish Academy of Sciences, Warsaw) is completely different, broad and tape-like, and unprotected by any additional sheath. The functionalism of such a bizarre ejaculatory duct and the mechanics of copulation in
E.
mariae remain unknown and are difficult to imagine. The only observed peculiarity in the structure of the female postabdomen, eventually in response to the male modification, is the extraordinarily long and spiral spermathecal duct. This might suggest the penetration and spreading of this duct by the peculiar rigid male conduit. This hypothesis is supported by the fact that, as shown in (
Figure 15D), the spermathecal duct thickens progressively towards the bursa and is visibly broader near its opening than at its spermathecal end. If confirmed by direct observation, possibly based on the immediate freezing of mating individuals, this would be a striking example of evolutionary competition between sexes.
5. Conclusions
The larvae and pupae of Euryommatus mariae (Coryssomerini) were reared from dead spruce Picea abies branches collected in central Poland. Its association with coniferous trees from the family Pinaceae, highlighted in the literature, is thus confirmed, even though spruce has never actually been mentioned as a potential host of this weevil. The species turned out to be saproxylic, in contrast to several Nearctic species of Cylindrocopturus (Zygopini) that develop in the living tissues of various American conifers and have a similar life cycle and phenology.
The very limited knowledge of the morphology of the immature stages of Conoderitae hinders comparisons and estimates of differences. The larva of
E. mariae shares the cuticle densely covered by thorn-like asperities with both
Coryssomerus and the members of
Cylindrocopturus, and probably also with
Eulechriopus (both Zygopini), although in the last-mentioned genus, the description of this character in
E. gossypii in Böving [
44] was inaccurate. This peculiar character, thus, appears to be common to a wider group of genera of Conoderitae. The unique larval characters of
E. mariae, not recorded in the other genera being compared here, are the single alar seta (
as) (vs. two
as), the abdominal segment VIII (Abd. VIII) with three postdorsal setae (
pds) (vs. two dorsal setae (
ds)), and the rounded head (vs. narrowed bilaterally), and elongate, bicornuate plate-like asperities arranged in transverse rows on the dorsum of abdominal segments, the character not known in other weevil larvae. The most easily observed diagnostic character of
E. mariae is the presence of numerous long, erect setae on thoracic segments I–III and abdominal segments VIII–IX. These were not reported in any of the species compared, in which all the dorsal setae are short to poorly discernible. This may be a consequence of saproxylic mode of larval life of
E. mariae, in the galleries bored under loose bark of dead tree branch, where the larva has to control more space than the larvae of other compared species boring a narrow channel in tight living tissues.
Euryommatus mariae is widely distributed in the taiga zone of Eastern Palaearctic [
3,
4,
5,
6,
7], whilst in Europe it seems to be very local and relict species found in just a few spots of natural coniferous forests in both the mountains (the type locality, Austria, Germany), and the lowlands (Latvia, Poland). Its disjunct distribution summarized in
Figure 16, the apparent confinement to large and natural or semi-natural forests of boreal type, and the evidenced association with deadwood, all support well the selection of
E. mariae for the group of umbrella species in forest conservation [
48]. The species is sporadically collected, and despite a very large area of distribution, the number of documented records is very low, either in Europe or Asia. It makes uncertain the gap in its range between Siberia and Europe, resulting from the hitherto literature records.
The unusually long male ejaculatory duct protected with a fibrous sheath is a unique character of E. mariae, probably unknown in other weevils (Curculionoidea). This contrasts with the extremely long female spermathecal duct, which, moreover, is spiral along its whole length. These two striking characters seem to be correlated and suggest the possible deep penetration of the spermathecal duct by the male with his half-rigid ejaculatory duct sheath; but this will have to be confirmed by direct observation.