Our study demonstrates that the palpal bulbus of an entelegyne spider possesses neural tissue. Moreover, we provide the first evidence of a multisensillar sensory organ. For several decades it has been accepted that male spider genitalia are sensorily blind due to the lack of neurons in the part of the copulatory organ that is involved in sperm transfer. This notion was built on numerous histological studies2,9,12,13,14. The apparent lack of sensory feedback during genital coupling has been considered to be compensated by the evolution of a number of sclerites that interlock with structures of the female copulatory organ as is typical for most spider species11. Recently, however, neural tissue was found in the simple bulbus of the austrochilid cave spider Hickmania troglodytes 27 which gave the starting point for revisiting the neurobiological properties of the bulbus in a representative of the majority of spiders, the Neocribellates. Interestingly, the arrangement of the nerve and neurons in H. troglodytes 27 and P. cespitum is nearly identical although the bulbi differ considerably between these two species. The neural tissue in the bulbus of P. cespitum also consists of a nerve proceeding from the cymbium and a somata cluster inside the bulbus. In both species, the neural tissue reaches the base of the embolus and seems to be tightly associated with glands as the neurites penetrate their tissue. The striking similarity in neural anatomy of two distantly related taxa may suggest that the presence and arrangement of the neural tissue represents a ground pattern for all Araneomorphae and possibly for all spiders.

The nerve that runs in the bulbus of P. cespitum is clearly distinguishable from other tissues when using the combination of methods applied in our study. This may explain why the studies mentioned above failed to find neural tissue in bulbi of other neocribellate spider species. Depending on the quality of the chemical fixation process and magnification, the presence of neurons may have been obscured. Especially in paraffin and semi-thin sectioning, which were applied in most of these studies, the neural tissue does not conspicuously stain with the common staining protocols. Also, in specimens only fixed in ethanol the neural tissue does not seem to be preserved well enough to be distinguishable, as was the case for the congener Philodromus dispar 9. Therefore, we predict that nerves may be found in the palpal bulbi in the whole order Araneae. Whether the nerves have a sensory or motor function or both needs to be assessed in follow up studies. Our results suggest that both directions are conceivable, since the bulbus nerve contains several neurite bundles (Fig. 2B’). The fact that we found a senory organ requires that the input perceived by this organ is transfered and processed to an integrating unit. Consequently, some of the neurite bundles are likely sensory. Since the neurites connect to at least two of the three glands found inside the bulbus, a further motor function of some of the bundles is plausible. These nerves may be involved in the release of secretions and consequently in the expulsion of sperm during mating, a process that likewise is not understood at present. In the following, we first discuss our data on the sensory and motor aspect of the neural equipment of the P. cespitum bulbus.

The sensory organ of P. cespitum is situated close to the base of the embolus, the intromitting structure of the bulbus. The organ is composed of aggregated and internalized sensilla. The cuticle of the embolus is penetrated by the outer dendritic segments of the sensilla, however, we could not trace their full course in the cuticle. We tentatively suggest that this sensory organ is chemo and mechanoreceptive. We base our hypothesis on ultrastructural details (see below), the connection with the various bulbus glands, and the position of the sense organ in the embolus. In H. trogylodytes, neurites were likewise found attached to the cuticle near to the embolus base. As a result of a stress and strain finite-element modelling analysis a proprioreceptive function was posultated27. However, our ultrastructural data reveal that the sensilla of the bulbus of P. cespitum resemble in some respects the internal set up of the tarsal organ, a chemoreceptor found on the tarsus of the walking legs as well as on the pedipalp of various spiders (e.g.34,35,36). Shared (ultra-)structures are: (1) the tight, palisade-like packing of the numerous, deeply internalized sensillar units, (2) the number of receptor cells and dendritic outer segments emitting from them (usually 3, rarely two), and (3) the elongated and narrowed sensillum lymph space, which is (4) accompanied by a thick dendritic sheath along its entire length. Due to difficulties in sectioning the most apical region of the sensilla, we do not know the full course of the dendrite outer segments. If the bulbus sense organ were homologous to the tarsal organ, the sensilla should terminate in small, cone-shaped, and tip-pored protuberances on the cuticle or in a depression that connects to the outside through a minute pore36,37. When scrutinizing the embolus wall under the SEM, we could not detect a porous structure. Thus our study tentatively suggest that the internalized sensilla in the bulbus function either as pheromone receptors37 or as combined thermo- and hygroreceptors38 as was previously suggested for the tarsal organ or, alternatively, as a proprioreceptor. Our study also documents two neurite bundles that make contact and branch within the glandular epithelia associated with the embolus and fundus. We failed to depict synapses, probably because the fixative does not infiltrate easily into the bulbus. However, there is a reason to assume that these two neurite bundles (see Figs 2D’ and 4A) include axons that innervate and thus regulate the activity of the embolus and fundus gland. Our arguments for assuming that these neurite bundles carry efferents and would therefore have to be interpreted as branches of the bulbus nerve are as follows. First, the bulbus nerve is made up of four distinct compartments of unequal size, so called fascicles31. In insects, such nerve compartments generally carry axons of strictly either sensory or motor neurons when distant from the ganglia38,39. Accordingly, we assume that the fascicles that are small in diameter contain efferents coming from interneurons located in the brain, each innervating a particular gland, whereas the largest fascicles of the bulbus nerve contains afferents that come from the sensory organ.

The neural equipment of the spider bulbus does not only impact on our perception of the possibilities and constraints that occur during the mating process in spiders. The finding that neurites enter the fundus and embolus glands may considerably help to advance our knowlegde on the mechanism by which sperm is released during mating from the interim sperm storage organ, the spermophor. It has been assumed that sperm expulsion is achieved through increasing internal hemolymph pressure applied on the bulbus by more proximal muscles or by means of a secretion from a gland attached to the spermophor12,13,40. Particularly, the spermophor gland was suspected to discharge its product through the porous wall into the lumen of the spermophor during mating2,13,41. Our histological examinations of P. cespitum bulbus reveal that the reservoir of the fundus gland is filled by a material identical to the one present in the lumen of the spermophor. In virgin males, large amounts of the material are deposited in the glandular reservoir and a small portion can be found in the fundus region of the spermophor. After copulation, however, the gland appears shrunken and the material fills more than half of the spermophor lumen. Our data strongly suggest that the fundus gland plays a crucial role in sperm extrusion and that the process is triggered by a neural stimulus. Since the mating process in spiders generally entails changes in pedipalp and bulbus conformation by internal pressure changes, the pressure pump and the neural stimulus may be coordinated for successful sperm transfer. Lamoral13 already reflected on why sperm is only released during mating after embolus insertion despite the fact that the pressure pump is active much earlier in the mating sequence, e.g. for external alignment with the female genitalia. He suspected that the release of a secretion for sperm transfer from the gland is triggered by neurohormones produced in more proximal parts of the male pedipalp, once the pedipalp achieved the right position. However, since sperm transfer often occurs very rapidly, a neurohormonal activation is unlikely14. Our study provides a tentative answer to this question since the fundus gland seems to be directly innervated by neurites projecting through the bulbus nerve. The sperm expulsion may be triggered when the sensory organ sends information about the correct positioning of the pedipalp during mating. This afferent transmission may cause the fundus gland to release the substance from its reservoir into the spermophor lumen to flush out the seminal fluid stored therein.

Before sperm release during mating, spider males have to charge their pedipalps. They transfer sperm from the production site (testes in opisthosoma) onto a sperm web and from there into the spermophor of their copulatory organs. Sperm uptake has to be repeated when the spermophor is emptied through copulation. It has been proposed that males take up sperm into the spermophor through resorption of material that fills the spermophor lumen before the bulbi are charged. Thereby the sperm mass is sucked into the spermophor10,13,42. We presume that the material we found in the fundus gland and spermophor fundus in virgin males of P. cespitum fills the entire spermophor before sperm induction and that it is resorbed by the fundus gland during the process of sperm uptake. Apart from the fundus gland, there are two more glands in the palpal bulbus of P. cespitum, the spermophor gland and the embolus gland. The spermophor gland seems to release a secretion into the spermophor before mating in many spider species (Uhl, unpublished) which may explain why there were no obvious differences between the spermophor glands of virgin and mated males in our study. Consequently, it may be involved in sperm uptake. The embolus gland may push the sperm mass out of the spermophor, and into the copulatory duct of the female which might be essential if the copulatory duct is long and the embolus cannot penetrate all the way to the spermathecae inside the female. Additionally, since P. cespitum males produce a mating plug by which males can hinder females from remating with rival males (Sentenská, Pekár & Uhl, unpublished) as was shown for other spiders42,43,44, one (or both) of these glands might be involved in producing the mating plug material.

In conclusion, our finding of a neural tissue and sensory organ in the pedipalp of an entelegyne spider together with the previous finding on nerves in the pedipalp of a basally branching Austrolichilid spider requires not only revising the common notion that genitalia of spider males are numb structures. The finding also calls for a reanalysis of the origin of the spider bulbus as a derivative of the tarsal claw2,45. This explanation has already been challenged by the observation that traces of a claw as well as of a bulbus occur simultaneously during early development of male pedipalps11,45. Overall, our findings offer a new perspective on genitalic sensory feedback and mate assessment in spiders.