During the spring of 2014-2018, from the beginning of March (occasionally, from February) to the end of June, field surveys were undertaken (see locations in Romero et al., 2020a) following nonlinear transects. During these surveys, besides the collection data, temperature, time of the day, clouds coverage and wind were recorded (the last two variables were assessed qualitatively). The surveys were undertaken at different times of the day, in the morning, during the central hours of the day and at sunset, to observe the general activity of R. sticticum in different moments of the day and the use of shells in particular. The presence of R. sticticum was assessed visually or by capturing individuals on the flight or inside snail shells, just counting the collected individuals. If the presence was visually assessed, no counting was done. Each collected shell was stored in an individual bag, to maintain the bees coming out from different shells separated.

Four main locations in the Iberian Peninsula were selected to assess feeding plants: Gabasa, Huesca (42.007153°, 0.416735°); Buendía, Guadalajara (40.394967°, -2.791320°); Énix, Almería (36.877929°, -2.609264°) and Nuévalos, Zaragoza (41.213594°, -1.791900°). Observation of interactions in these four locations was performed in May and June, in 2017 and 2018, during the flowering peak of most plant species. They took place during the diurnal period of flower visitor activity (from 10:00 to 17:00 h). Sampling followed nonlinear transects, covering all different species that were on bloom within the area. The surveys were done by direct observation. An interaction was considered when R. sticticum touched the flower reproductive parts, but not when it only landed on the flowers. To identify the plant species, samples of all the visited plants were collected. Data about plants visited by R. sticticum was also collected from bibliography (Bosch et al., 1993; Torres et al., 2001, 2002; Escudero et al., 2003; Aguib et al., 2010; Vargas et al., 2010, 2013; Carrió & Güemes, 2013; Fernández-Mazuecos et al., 2013; Blanco-Pastor et al., 2015; Schurr et al., 2019), in addition to personal communications (Jordi Bosch).

The behaviour of both, males and females of R. sticticum, was observed and recorded through photographs and videos during May of 2016 and 2017 in Buendía (Guadalajara), in patches of Antirrhinum microphyllum Rothm., to describe flight ability, territoriality, mating and general behaviour. No individuals were marked. All these traits were described by means of both direct observation and detailed observation of the recordings.

In relation to R. sticticum life cycle, 66 observations were accomplished in different locations, under different weather conditions, days or time of the day, between 2014 and 2019, the earliest that an individual of Rhodanthidium sticticum was found on February 8 in La Breña Natural Park (Cádiz) and the latest, the 17^th of June in Nuévalos (Zaragoza). Most of individuals were observed or collected between mid March and the end of May (85% of the records, Fig. 1). Almost 70% of the records were obtained during sunny days. Around 75% of the observations were made during days without wind. Concerning temperatures, 71% of the records were observed under temperatures between 20 and 30 degrees Celsius. In 66% of the observations, more than five different individuals were captured.

Of a total of 196 collected individuals, 63% were males and 37%, females. 56% were captured flying or visiting flowers and 44% were captured inside snail shells. 86 individuals were collected in 60 shells (Fig. 2). For the individuals captured or observed on the flight, the conditions were sunny and warm, and they were recorded during the central hours of the day. Most of the captures inside snail shells took place in cold, rainy and windy days, or at late afternoon.

Concerning the individuals found inside shells, 70% did not share it, in 13% of the shells there were two individuals, in 8% of the shells three individuals of Rhodanthidium sticticum were sheltering and in one shell we found six individuals (Table 1A). Two shells were shared by an individual of Rhodanthidium sticticum and another bee of the genus Osmia Panzer, 1806 and two other shells were occupied by an individual of Rhodanthidium sticticum and an individual of Rhodanthidium siculum (Table 1C). Of the 18 shells shared by two or more bees, only three were occupied by females, each of them shared by a female and a male of Rhodanthidium sticticum. All the other females collected in snail shells were alone in the shell (19 out of 22 individuals, 86% of all the females collected in shells) (Table 1B).

Our field surveys showed a high diversity of plants on which R. sticticum feeds (Appendix 1). In particular, 43 species and 13 families of flowering plants from NE, C and SE of the Iberian Peninsula were identified. Among them, Papilionaceae (8 spp.), Lamiaceae (12 spp.) and Plantaginaceae (5 spp., all of them Antirrhineae) were the most frequently visited (Appendix 1). From literature, 38 more species belonging to 14 families were retrieved (Appendix 2). Therefore, polylecty appears to be predominant in Rhodanthidium sticticum diet. Despite that, R sticticum shows a clear attraction for Antirrhinum species.

Males spend most of the time patrolling their territories (estimated 75% of the time), looking for females (Fig. 3) or resting on nearby rocks and branches, usually promontories where they can warm up on the sun while they watch their area. From time to time, they visit the flowers in their territory to feed on nectar, but they do this activity more rarely. Rhodanthidium sticticum males also perform exploratory flights when a new object enters their territory (a bag or a camera, for example), during which they remain in static flight in front of the object, observing it directly. Females spend their time looking for suitable shells for nesting, building the nest, feeding on nectar and collecting pollen for their offspring (Fig. 4).

Males protect territories with abundant flower patches, showing preference for Antirrhinum, Labiatae and Papilionaceae shrubs. They show a ferocious defence of the territory, chasing, attacking and striking other male intruders, but also other bees of similar or bigger size (including much bigger bumblebees or Xylocopa violacea (Linnaeus, 1758) individuals) and other flying insects that happen to enter their surveillance area, like flies or butterflies. We did not observe Rhodanthidium males attacking small flying insects in their territories.

A number of mating attempts end up in failure. Females are reluctant to the copula, so males try to catch them unaware. When a female detects the proximity of a male while feeding or collecting nectar, quickly take flight to avoid it (Fig. 5). Contact between a male and a female does not guarantee mating, either. Frequently, females wriggle out before males are able to copulate. If other male candidates are around, they usually interrupt the copula by trying to force the first male out and get the female. Even if the male has been able to start intercourse with the female, sometimes they become too heavy for the stems they are leaning on and they fall, allowing the female to escape (Fig. 6). During a successful intercourse (Fig. 7), a very characteristic sound, a continuous, rhythmic, dry clickety-clack noise can be heard, caused by the hitting of the cuticles.

Rhodanthidium sticticum is a very active bee during the flowering period of most Spring flowers, from March until June, but in warmer areas it is usual to find it as early as February. Most of records were taken between March and May, enlarging until mid June, matching the data from other areas of the species (Morocco, Algeria and Sicily: Kasparek & Lhomme, 2019; Algeria: Aguib et al., 2010). There are no records in literature or in collections about the presence of the species later than June, despite Kasparek (2019) stating, without providing any evidence (“an early spring and late autumn species with possibly two generations per year”) and that it was found in Spain in October/ November. It seems to be, however, an univoltine species, like R. siculum or other Anthidiini bees (Anthidium florentinum, Fortunato et al., 2013). Flying period for this species starts early and is relatively long, similar to R. septemdentatum (Kasparek, 2019) but longer than R. siculum (Ortiz, 1990; Aguib et al. 2010; Erbar & Leins, 2017), probably due to the fact that R. siculum is associated to warmer climate and its optimal period is shorter (Romero et al., 2020a).

Our results show that R. sticticum is active in sunny, not windy days, as most of the records on the flight were obtained under those conditions, corroborating previous observations by other studies (Torres et al., 2001; Romero et al., 2020a). However during cloudy, rainy and windy days, or at the late afternoon, R. sticticum bees were mostly found sheltering inside snail shells. Apparently, Afranthidium hamaticauda Pasteels, 1984, A. odonturum (Cockerell, 1932) (Gess & Gess, 1999, 2008, 2014) and Hoplitis conchophila Kuhlmann, 2011 (Kuhlmann et al., 2011) also use snail shells for sheltering, but this behaviour has rarely been described. In those cases, the authors did not provide any information about the weather conditions or the time of the day they did their records at, but considering the location (the Namibian desert) the use for sheltering is, most likely, against the heat of the central hours of the day or to sleep at night. Despite the existence of detailed descriptions of the nesting behaviour of Rhodanthidium bees inside snail shells (Pasteels, 1977; Erbar & Leins, 2017), sheltering has never been described before on this species.

Of all the bees found in shells, 30% were sharing and most of them were males. This behaviour is in contrast with the territoriality usually showed when they are flying and there are no previous records of other bees in their adult phase sharing shells. The reasons could be diverse. Individuals sharing the same shells they were born from (phylopatry), high density of individuals and few shells disposable, which would force them to share, or the urge to rapidly find a shelter when the weather suddenly changes. The fact that most of the sharing bees were males is probably due to the higher proportion of males. The male-female proportion (3 males per female) is unlike other Anthidiini bees (Anthidium florentinum, Fortunato et al., 2013).

Our results indicate low specificity of R. sticticum for particular plant species, in concordance with previous studies (Bosch et al., 1993; Aguib et al., 2010; Torné-Noguera et al., 2014). Specificity of bees to special feeding plants has been described as a rare phenomenon (Cane & Sipes, 2006; González-Varo et al., 2016). Indeed, R. sticticum appears to be a polylectic bee (Bosch et al., 1993; Müller, 1996), although it probably has a preference for bee-specialized (melittophilous) plants, such as Antirrhinum, Linaria and Papillionaceae (Blanco-Pastor et al., 2015; Vargas et al., 2017), and for deep-corolla flowers, like the Labiatae (Appendices 1–2). The flowers of melittophilous plants show fusion of petals (sympetaly) and a bipartite perianth that hinders the entrance of insects other than bees and other Hymenoptera (Vargas et al., 2010; Blanco-Pastor et al., 2015). This flower structure protects the floral reward, thus ensuring that individuals reaching the bottom of the flower will most certainly get it, a reason why it could be so appreciated by R. sticticum females. It is particularly noteworthy the interest showed by R. sticticum in Antirrhinum species all through its range (see Appendices 1–2), proving that it is not a local preference, but a general one.

Flight ability by R. sticticum is identical to that described for R. septemdentatum (Nachtigall, 1997a). It is fast and precise and could be hypothesized that its good sight is, in part, responsible for it. Very few and recent studies concerning sight and perception have been performed on solitary bees (Loukola et al., 2020), but multiple discoveries in bumblebees confirmed learning skills, precise colour and shape discrimination (Giurfa et al., 1995; Spaethe et al., 2001; Dyer & Chittka, 2004; Solvi et al., 2020). The exploratory behaviour observed in R. sticticum males (and probably females, though it was not detected) is very similar to that described by Loukola et al. (2020) in Osmia bees when searching for suitable nesting sites.

Rhodanthidium sticticum males, like most Anthidiini bees, are usually larger than females (García-González & Ornosa, 1999; Michener, 2007; Erbar & Leins, 2017), a probable consequence of territoriality, which helps them to get and defend better territories (Severinghaus et al., 1981; Villalobos & Shelly, 1991). Defence of territory is widely spread among Anthidiini (Michener, 2007). Selection of territories where deep-corolla flowers are predominant has been previously reported for Anthidium maculosum (Alcock et al., 1977), A. septemspinosum (Sugiura, 1991), Anthidiellum notatum and A. perplexum (Turell, 1976). Deeper corollas difficult females from seeing nearby males and escaping them. In the closely related species R. siculum, however, the defended territory is not rich in feeding flowers, but in empty shells for nesting (Erbar & Leins, 2017). In this case, R. siculum males take advantage of females when they are inside the shells. Despite the abundance of deep-corolla flowers, many attempts at copulation were unsuccessful, due to the lack of receptiveness and the ready flight of the females, just like in other Anthidiini bees (Wainwright, 1978; García-González & Ornosa, 1999; Erbar & Leins, 2017).

In the case of R. sticticum, the fierce defence is oriented not only against other males, but also against individuals from other species. This behaviour has also been described for A. florentinum, against much bigger Xylocopa violacea bees (García-González & Ornosa, 1999), for A. maculosum against hawkmoths (Alcock et al., 1977) or for A. palmarum against Anthophora bees (Wainwright, 1978). A different behaviour is showed by Anthidiellum notatum and A. perplexum, smaller Anthidiini that do not strike the intruders, but just chase them until they leave the territory (Turell, 1976). On the other extreme are Anthidium manicatum, whose males bear long spines at the end of the abdomen which they use to hurt or even kill intruders (Wirtz et al., 1988), and R. septemdentatum (Nachtigall, 1997b), that attacks the wings of the intruders. That extremely aggressive behaviour was not observed in R. sticticum, but it could not be discarded. Interspecific territoriality as shown by R. sticticum and the other species has been described as a way of keeping competitors of females (for nectar and pollen) away and maintaining the attractiveness of the territory (Severinghaus et al., 1981; Wirtz et al., 1988).

The fact that some males enter the alleged territories of other males may indicate several things: territories overlap, new or young males looking for a territory, or opportunistic wandering males, without a territory. The presence of wandering males was described in A. florentinum (García-González & Ornosa, 1999), A. maculosum (Alcock et al., 1977), A. manicatum (Severinghaus et al., 1981), A. septemspinosum (Sugiura, 1991) and R. siculum (Erbar & Leins, 2017), fighting because of territories overlapping was described for A. banningense (Jaycox, 1967) and territories changing ownership was described for A. manicatum (Severinghaus et al., 1981). The three causes seem to be plausible for R. sticticum, but further observations including individuals measuring and marking should be done to confirm it.

Unlike most bees, Anthidiini females are polyandrous, whereas males exhibit resource defence polygyny (Michener, 2007). In this context, it has been demonstrated that, in A. manicatum, the last copulating males have an above-average chance of fertilizing the female’s egg (Lampert et al., 2014), and the same has been theorized for A. florentinum (García-González & Ornosa, 1999). The polyandrous behaviour of R. sticticum females, together with the polygyny by males, could be driven, as it happens in A. manicatum and most certainly in A. florentinum, by spermatic competence.