Mesoplia sapphirina, Rozen & Vinson & Coville & Frankie, 2011, Rozen & Vinson & Coville & Frankie, 2011

Rozen, Jerome G., Vinson, S. Bradleigh, Coville, Rollin & Frankie, Gordon, 2011, Biology of the Cleptoparasitic Bee Mesoplia sapphirina (Ericrocidini) and its Host Centris flavofasciata (Centridini) (Apidae: Apinae), American Museum Novitates 2011 (3723), pp. 1-36 : 10-18

BIOLOGY OF MESOPLIA SAPPHIRINA

In searching for nests of C. flavofasciata , female M. sapphirina fly swiftly over large areas, tightly circling a suspected spot of interest before dashing to the next one, making an unpredictable zigzag pattern across the landscape from one inspection point to the next. One suspects that they can quickly detect through sight or odor when a nest is present, because only rarely do they briefly land. We noted in 2010 when they found a nest where a female host bee was present, they showed great persistence in trying to scout the site and enter the nest. They repeatedly returned, each time either to be chased by the much larger host or to be blocked from the entrance where the host female stood guard. In 2011, we observed female M. sapphirina searching for nests as in the previous year, but they were fewer in number, and we were rarely present when they attempted to enter.

In 2010 female M. sapphirina entered nests that were still open as well as those that had been closed. Some visits were brief, perhaps less than 1 min, and almost certainly did not result in egg deposition, while others may have lasted for some minutes and thus were thought to result in parasitism. In one case a host female returned and entered while the cleptoparasite was still inside, resulting in the host rapidly chasing the cleptoparasite from the entrance. In 2011 we observed a female M. sapphirina take 8–10 min to enter a nest, the entrance to which we capped with a plastic tumbler. An hour later the female finally emerged. Thus it had taken her approximately 1 hr to penetrate the nest, successfully introduce her egg through the cell closure, seal the hole, and emerge. This observation casts doubts on the success of all presumed introductions of the previous year.

Our data on the number of nests (i.e., cells) containing recent instances of parasitism during six visits to the site over a 10-day period in February 2010 are limited because of the difficulty of finding nests. A total of 10 nests containing fresh inhabitants were discovered and studied. Cells from a number of older nests representing earlier generations were uncovered by chance. Some were cells from which adults had emerged, as shown by cocoon fragments, and others contained remnants of immatures that perished during development .

The 10 cells from the then current generation are listed in table 1 with features that varied among nests. As indicated (table 1: first row) nine cells had been parasitized by at least one female of M. sapphirina as evidenced by eggs or larvae found therein. Only one cell, discovered on the last day, had not (yet) been parasitized. This suggested an extremely high rate of parasitism, placing into jeopardy survival of the local population of C. flavofasciata if the rate was continuous through the entire adult breeding season (assuming that C. flavofasciata is univoltine). Not only were all but one cell attacked, but half of them (five of 10) had been visited by more than one cleptoparasite, and one was presumably visited by as many as six cleptoparasites (assuming that a cleptoparasite inserts only one egg per visit and visits a cell only once). The high rate of parasitism presented in these figures was supported by our casual observations of the rapid, thorough searching behavior of female M. sapphirina . The study by Vinson et al. (1987) reported a parasitism rate of 59% based on a sample of 22 cells. Although somewhat lower than the rate suggested by the 2010 study, it still demonstrated a very successful cleptoparasite.

By returning to this study in 2011, J.G.R. and S.B.V hoped to ascertain what effect the heavy parasitism rate the previous year had on the host population. We also wanted to better understand how Mesoplia sapphirina attacks its host and how C. flavofasciata defends itself from such attacks.

On returning both authors were surprised to discover (1) the host population active and roughly as abundant as in the previous year and (2) the parasite population substantially reduced in numbers and seemingly far less aggressive, in that both authors only rarely observed female Mesoplia digging into closed host nests in nine visits to the site. During these visits, we uncovered 36 active host nests, which were excavated a recorded number of days after they were discovered for most trips. The most interesting statistics are as follows: 11 cells were parasitized by Mesoplia giving a parasitism rate of 30% (contrasting with 90% for 2010); no cell contained more than a single Mesoplia egg or early instar (contrasting with half of all nests of previous year parasitized by 3–6 cleptoparasites); 20 cells had live host eggs and 7 others contained early larval host instars (contrasting with 4 cells with live host eggs in 2010); no cells found where all bees had been killed (contrasting with 30% of all bee inhabitants had failed a year before); no ants were found to have invaded cells (whereas ants were associated with at least 2 nests in 2010).

Although we recognize that the small size of the 2010 sample is less than a reliable measure of the population for that year, the samples of both years show striking differences. We suspect that the activities of 2010 may have resulted from an overly large population of Mesoplia (for reasons unknown) relative to the host population resulting in multiple egg depositions in host cells and an extremely high rate of parasitism. Nests in which all bee inhabitants had died may have resulted from Solenopsis invasions. As table 1 indicates, most cells in 2010 exhibited small irregular holes (ca. 0.5–0.6 mm in diameter) in the cell closure through which M. sapphirina eggs are inserted (fig. 22), as also reported by Vinson et al. (1987). One cell also exhibited a small hole in the upper cell wall. As indicated in the 2011 study, most holes are probably egg insertion holes of M. sapphirina . However, in 2010 there was lack of congruence between the number of such holes and the number of cleptoparasites in a cell: generally fewer holes than parasites. Some evidence suggested that a female M. sapphirina sealed the oviposition hole afterward, as was affirmed in 2011. Some oviposition holes may have been reopened and even enlarged by ants (e.g., table 1, cell 5), as is suggested by the 2011 study, which revealed no ant infestations and, in general, showed a one-hole-to-one-cleptoparasite ratio.

The 2011 study helped resolve questions stemming from the previous year. Are Mesoplia eggs introduced into the brood cells only after cell closure? Some observations in 2010 indicated that M. sapphirina females enter host nests before nest closure, thereby suggesting that the brood cells may also be open and available to the parasite. Not only did we not observe any cleptoparasites entering open nests in 2011, but of nine parasitized cells only one did not reveal any parasite egg insertion holes, which are hard to detect because of their small size and sealing. We conclude that M. sapphirina probably attacks only closed host cells. Further, we reason that if parasite eggs were deposited before cell closure, any egg as large as that of M. sapphirina would be easily detected and eliminated by a returning female C. flavofasciata , as has been discussed for other cleptoparasitic bees (Rozen, et al., 2006: 24, 25).

In 2010, it was uncertain how many eggs a single female cleptoparasite deposits in a cell per visit. Limited data favored only one, but the high rate of parasitism in some cells, particularly when there was only a single hole in the closure, could indicate more. This matter was resolved from data gathered in 2011: 8 of 9 parasitized cells (10th cell destroyed in excavation) each exhibiting a hole, usually stoppered with sand grains, and each containing only a single M. sapphirina . The lack of numerous fully formed oocytes in dissected females also strongly points to only one egg being deposited per visit.

Partway through our examination of the collected cells in 2010, we examined provisions of cell 8 (table 1) submerged in water and discovered a large number of extremely small ants identified as Solenopsis sp. that had been totally overlooked in provisions not examined in water. Although it was impossible to review cells examined earlier, this discovery may well explain why in some cells, all M. sapphirina were dead, e.g., table 1, cell 3 (although why all parasite eggs were attacked, whereas the host egg was not, remains unexplained). Data from 2011 revealed that ants had not attacked a single cell among the 36 examined, possibly because most nests had been excavated by us within three days after closure.

In the 2010 study, evidence was unclear why cleptoparasite eggs in some cases hung by one end from the cell closure as was the situation in cell 3 where two eggs hung by one end next to the hole through which each had been deposited (before presumably being attacked by Solenopsis ). An egg attached by one end to a closure had been reported for Aglaomelissa duckei (Friese) from Trinidad ( Rozen, 1991: 32). In all other cases, eggs in the 2010 investigations were found on the provisions. With the 2011 study three eggs were attached by their anterior ends about 1–2 mm from the hole through which they were inserted. Since all five first instar Mesoplia were accompanied by holes through the cell closure, there is little doubt that M. sapphirina eggs are normally attached presumably always by their anterior ends to the cell closure. How the female attaches her egg to the inner surface of the closure 1–2 mm from the hole remains unknown, although possibly some Mesoplia egg detachments in 2010 resulted from numerous visitations by conspecific cleptoparasites that dislodged attached eggs while attempting to oviposit or by ants.

The egg incubation period for M. sapphirina clearly seemed brief, for an open nest was identified two days before a first instar was recovered from the cell in 2010, and in 2011, five first instars were removed from cells that had been closed within the last two or three days. Finally on February 19, 2011, an adult M. sapphirina descended into a host nest after flicking away sand at the surface for ca. 8 min. After it had disappeared from sight, we placed a plastic tumbler over the entrance and thereafter monitored the site while working the area. One hour later (10:45 AM) it had emerged and died in the heated tumbler. That day the brood cell was excavated, and a Mesoplia egg was found attached to the cell closure. The first instar hatched 29.5 hrs later. Unfortunately, we missed observing the hatching process, although the ragged connection of the chorion with the rear of the abdomen was well documented (fig. 28). We placed the larva on the surface of the provisions of another cell containing a host egg. It made no obvious efforts to locate the host egg, not surprising considering lack of sight, but reacted with much body twisting and turning when touched with a forceps, perhaps suggesting a defense reaction to other cleptoparasites. When placed close to the host egg, it snagged the host chorion with one of its mandibles, but did not proceed to attack with both mandibles as if trying to eliminate the host. The larva seemed unable to crawl forward or backward but twisted rapidly and with agility, presumably defense tactics. However, when observed two hours later, the Mesoplia larva had completely destroyed the host egg. A subsequent observation of another first instar revealed that it could indeed slowly crawl forward and backward when not disturbed, e.g., by being touched with forceps.

Some of the assassinated eggs in the 2010 study showed a swelling at the anterior end that is slightly wider than the diameter of the rest of the egg (figs. 23, 29). The ventral chorion immediately behind the front end was substantially smoother than elsewhere on the egg (compare fig. 32 with figs. 30, 31, 33–35). In one egg that had been killed we noticed that the length of this anterior part was approximately the same as the length of the mandible of the developing embryo within, whereas most of the head of the embryo was posterior to the swelling. Furthermore, on some other dead eggs the anterior pole containing the micropyle was invaginated (figs. 23, 29, 36). We wonder whether these apparent deformities are real: the swelling of the anterior end allows the mandible to open and close and the micropylar area invaginates (through some unknown mechanism), permitting the sharp apices of the mandibles (fig. 47) to reach and rupture the anterior end of the chorion. Verification of the above speculation should be easily forthcoming by observing an egg that is hatching. The role played by the pronounced paired lateral lobes of the prothorax (fig. 27) of the first instar is unknown but might be involved with eclosion.

During both years virtually all live first instars of M. sapphirina and most dead ones had their seemingly still-inflated egg chorions (figs. 24–26) attached to the terminal abdominal segments, indicating that this is a normal behavior pattern for the species, the possible adaptive function of which is discussed below. Although the entire head, thorax, and first seven abdominal segments extend freely from the chorion, chorion covers abdominal segments 8–10, and the front end of the chorion is shredded (figs. 28, 39). As mentioned in the description of the egg, the chorionic surface texture a short distance behind the anterior pole is highly modified on the ventral and lateral surfaces. This region of the egg appears to adhere to abdominal segments 8–10 of the first instar. It is unclear how this attachment is maintained through the entire first stadium. Might the somewhat expanded form of abdominal segments 9 and 10 when viewed dorsally (fig. 43) and the declivity between 8 and 9 when viewed from the side (as in Rozen, 1991: fig. 55) play some role? Two of three first instars that were kept alive in 2011 lost their chorion several hours before molting to the second stage; the third kept the attachment until molting.

It is obvious that laterally expanded abdominal segment 10 in this species (and probably in all other Ericrocidini ) has nothing to do with a pygopodlike function as suggested by Rozen (1991), since we now know that that segment 10 is encased in the chorion and cannot assist larval crawling.

First instars of M. sapphirina are extremely agile and feisty. They are capable of bending their entire body from where it is attached to the egg chorion sideways or overhead, so that the head is pointed completely backward in the same direction of the tail of the body. By reaching backward it can quickly defend itself against any adversary that might be attacking its attached empty chorion. This agility is demonstrated in figure 38, a sequence taken in a 2 min period. Second instars, though no longer carrying chorions, also display considerable agility and combativeness, features that fade in successive instars.

Data suggest that all larval stadia are brief, for one larva collected in 2010 as a second instar had reached the fifth instar within a period of four to five days, and started defecating two days later. Soon thereafter it started spinning a cocoon, but when preserved, it was removed from its cocoon while still feeding on a large mass of food and with its digestive tract still containing food, perhaps a developmental artifact resulting from being reared under artificial conditions. In 2011, an egg of M. sapphirina that hatched at 10:45 AM on February 20 was preserved as a third instar at 4:49 AM on February 25. The same year a first instar collected on February 21 was observed as a second instar at 4:30 PM on February 22, and as a third instar at 6:45 AM on February 24. Clearly, developmental behavior and timing need further detailed investigation .