On the Natural History and Functional Morphology of the Clam Shrimp, Lynceus Brachyurus Müller (Branchiopoda: Laevicaudata)

Abstract

The small bivalved clam shrimp, Lynceus brachyurus Müller, 1776, is seasonally abundant in two long-duration vernal pools in central Ohio, USA. Previously unrecognized anatomical features include dorsal patches of spines or filaments on posterior trunk segments, diverse medial spines on the dorsal lobes of the leg exopods, two-segmented dorso-lateral appendages on the male eleventh segment, and brown coloration on portions of male legs. It can swim actively, but is basically benthic. Females are sometimes totally quiescent. Gut contents, observation, and experiments showed a diet of substrate scrapings, suspended detritus, and small zooplankton. Females ovulate following male stimulation and produce new egg batches without an intervening molt. Repeated matings are needed to produce the large egg masses often carried. Larvae were first collected in late February, and adults in late April. Lynceus was found in nine of the eleven years studied, being absent in the years that followed two years of high rainfall.

Introduction

For several months each year, the vernal pools in woodlands of the formerly glaciated regions of eastern North America harbor a diverse community of species, including several restricted to this habitat (Colburn, 2004). The discovery of populations of the clam shrimp, Lynceus brachyurus Müller, in two vernal pools in central Ohio provided an opportunity to investigate one of the lesser known specialist species.

The clam shrimp of the order Laevicaudata (Linder, 1945; Fryer, 1987) are small, up to |$7.0\;{\rm{mm}}$| carapace length (Timms, 2013) and can be completely enclosed within their bivalved carapace. They have a patchy distribution in temporary waters on all continents except Antarctica (Martin and Belk, 1988). Although long included with spinicaudatan clam shrimp in a Class Conchostraca, morphological and molecular studies now show that Laevicaudata is well separated from other bivalved branchiopods (Linder, 1945; Fryer, 1987; Olesen, 2009; Rieger et al., 2010). The anatomy of species of the principal genus, Lynceus, has been described and illustrated (Liévin, 1848; Grube, 1853; Sars, 1896; Martin et al., 1986; Martin and Belk, 1988; Fryer and Boxshall, 2009), while a number of papers deal with aspects of biology (Liévin, 1848; Grube, 1853; Sars, 1896; Botnariuc, 1947; Valtonen, 1966; Martin, 1989; Rabet et al., 2005; Fryer and Boxshall, 2009; Sigvardt and Olesen, 2014). Variation in functional features between species suggests adaptive radiation within the genus (Fryer and Boxshall, 2009).

Lynceus brachyurus occurs in seasonal bodies of water in Europe and North America (Botnariuc, 1947; Martin and Belk, 1988; Schell et al., 2001). It was studied over an eleven year period, with the goal of combining anatomical and behavioral observations with experimental and collection data to further understand the life of this species. Its ease of maintenance in the laboratory, translucent carapace, and habit of lying on one side on the substrate facilitated observation and experiment, while repeated collections provided information on the life cycle.

Materials and Methods

Collection Sites

The study site consisted of three adjacent ephemeral pools in depressions in second growth forest belonging to the Stratford Ecological Center, Delaware, OH, USA. The West Pool is a shallow, mostly shaded, depression of approximately |$0.1\;{\rm{ha}}$| maximum extent. It is the most short-lived of the three pools and does not have Buttonbush (Cephalanthus occidentalis) growing within it. The North Pool (approx. |$0.26\;{\rm{ha}}$|⁠) has flat margins that recede markedly with declining water level, a generally soft substrate, and contains Buttonbush. The South Pool (approx. |$0.47\;{\rm{ha}}$|⁠) has a smaller proportion of shallow marginal area and has Buttonbush and other vegetation growing from a firmer substrate. It retains water longest and was sampled most intensively. Each pool is about 100 meters from its nearest neighbor. They share a common drainage during seasonal overflow conditions, but direct transfer of water between them does not seem possible. Weather data are from Marysville, OH, USA, |$25\;{\rm{km}}$| to the west of the study site.

Collection Procedure

Collections from the three pools were made in most weeks from February to June 2004-2009 (Table 3). The pools were sampled by sweeps above the substrate with a |$1 - {\rm{mm}}$| mesh D-net. The sweeps followed a fairly standard path of about 15 meters in the same area within each pool, allowing rough comparisons to be made between collections. The proportion of the pool from which animals were removed was felt to be too small to affect subsequent sampling. Net contents were everted into small aquaria, leaf and other debris removed, and the water poured through a small sieve. Retained material was preserved in |$70\%$| ethanol. Animals were collected for laboratory observation and experiment.

Samples of South Pool zooplankton were collected at approximately weekly intervals during weeks 12-18 in 2008 and weeks 7-18 in 2009 and 2010. About thirty liters of pool water were poured through a |$30\;\mu {\rm{m}}$| plankton net and the contents preserved in formalin. In the 2007-2008 season, samples of water, leaves, and bottom sediment were collected weekly from the South Pool as conditions permitted, brought into the laboratory, and visually examined at roughly weekly intervals for the presence of Lynceus.

Laboratory Procedures

The weekly collections were examined and any L. brachyurus present counted. Up to 50 specimens from the 2008 North Pool weekly collections, and the entire collection of 11 June, were sexed and measured. Males were recognized by their modified first legs. Carapace lengths were determined to the nearest |$0.1\;{\rm{mm}}$| with an ocular micrometer, although tilting of the oval body sometimes made precision difficult. Plankton samples were examined under a dissecting microscope and any Lynceus stages removed. When nauplius larvae were present, the width of the dorsal shield was measured for up to 50 larvae per collection.

Animals to be dissected were stained with Lignin Pink and cleared in glycerine alcohol. Material for electron microscopy was fixed in |$3\%$| gluteraldehyde or |$70\%$| ethanol, sonicated for 10 seconds to remove attached debris, dehydrated in a graded alcohol series, critical point dried, gold coated, mounted on aluminum stubs, and examined under constant or variable pressure.

Lynceus brachyurus was maintained in the laboratory in dishes of pool water and dead leaves, either at room temperature, or in an environmental chamber at |$18 - 20^\circ {\rm{C}}$| and a 14/10 hour light/dark period. Behavior was observed with a stereo-microscope and gut movements with an inverted compound microscope. Some animals maintained in the laboratory had slower limb movement which facilitated observation. Photographs and videos were made of living animals, including some fixed by one valve to a wire with cyanoacrylate. Experiments were conducted in 11 or |$20\;{\rm{mm}}$| finger bowls, or small jars with fragments of decomposing leaf and approx. 150 or approx. |$600\;{\rm{ml}}$| of pool water. Plankton capture was investigated using |$11 - {\rm{cm}}$| finger bowls of pool water enriched with zooplankton and with or without 6 L. brachyurus. After 48 hours, formalin was added to the dishes, and the contained plankton concentrated and counted with a counting wheel. To study egg production and release, paired or single females were collected on 30 May 2008, maintained individually, and examined twice a day for four days. Mating and egg-bearing status were recorded, and any eggs that had been released counted using a grid. Other experiments are explained in the text. Comparative studies were assessed without knowledge of group membership.

Results

Body Plan

This section adds supporting figures and new observations to an outline of the anatomy of Lynceus relevant to the following sections. The sub-spherical bivalved carapace is tough but flexible, and may completely enclose the body. The valves have an outer rim that is most pronounced on the posterior margin and slopes more gradually than the rest of the carapace (Figs. 1A, 2A, C, 6E). In decaying animals, the rim can separate as a narrow arc, complete except for the region of the hinge. The valves are lined with a substantial membrane (“respiratory membrane” Liévin, 1848), the edges of which can meet to form a seal between opposing valves. (Figs. 1A, 2A). This membrane leaks blood when perforated and so must have a respiratory function, making it appropriate to call the space between the valves the branchial cavity. The valves are opened by the dorsal hinge when the adductor muscle relaxes (Fig. 11B, E). The body is attached to the carapace by an anterior transverse muscle and a slender dorsal connection near the front of the hinge (Fig. 1A). The head comprises nearly half the total body length and can turn downward under the body, “like the beak of a flamingo” (Liévin, 1848, p. 6), so that the ventral surface of the rostrum lies uppermost and nearly parallel to the hinge (Figs. 2A, 3A). Alternatively, the anterior head and eyes may protrude well beyond the valve edges (Figs. 5D, F, 12A). A ridge of exoskeleton (fornix) extends along each side of the head (Figs. 1B, C, 2A). It curves dorsally and medially to produce an indentation that accommodates the antennal peduncle and is distinctive enough to be called the antennal notch (Fig. 2A, C). The fornix continues forward to terminate in a lateral spine (Figs. 1B, 2B). A large labrum extends down from the head. The head appendages consist of small first antennae, prominent second antennae, large mandibles, and small maxillules. The second antennae extend outwards when beating, or lie back within the valves, either medial or lateral to the legs (Figs. 1B, 2A, C).

Fig. 1.

Male Lynceus brachyurus. A, lateral view, left valve removed; B, frontal view of head and anterior appendages; C, dorsal view of head and anterior appendages. |${\rm{A2}} = {2^{{\rm{nd}}}}$| antenna; am = adductor muscle; cm = carapace membrane; dc = dorsal connection between body and carapace; E2, 3 = endites 2 and 3 of T1; ep = epipod of T1; fo = fornix; he = head; ne = neck (cervical segment); lc = less chitinized region of epipod where it joins the leg; ls = lateral setae on dorsal lobe of exopod of T1; mn = mandible; ms = spines on medial margin of dorsal lobe of exopod of 1^st trunk leg; T1 = trunk leg 1; tr = trunk. Scale bar = |$200\;\mu {\rm{m}}$|⁠.

Fig. 2.

Female Lynceus brachyurus. A, lateral view with left valve and egg mass removed; B, dorsal view of head and 2nd antennae; C, ventral view showing position of head and appendages when valves are closed (left valve removed). an = antennal notch; ap = antennal peduncle; cm = carapace membrane; deT1, 2, 9, 10 = dorsal lobes of exopods of T1, 2, 9, 10; dl = dorsal lamella; ep = epipod of T1; fo = fornix; hi = region of hinge between carapace valves; lc = less chitinized region where epipod where joins the leg; mn = mandible; ov = opening of oviduct; sh = shelf on posterior edge of valve; T1 = trunk leg 1. Scale bar = |$200\;\mu {\rm{m}}$|⁠.

The trunk may bend down to approach the substrate when moving over it, or exit mid-posteriorly when lying on one valve, or swimming. The dorsal surface of the last trunk segments reveals variously developed rows or clusters, of soft or spine-like, posteriorly directed, projections (Figs. 8A-C, E-F, 9A, C, D). The female trunk bears unique, dorsal lamellae (Figs. 3A, 4A). Males have a distinct eleventh segment behind the one bearing the last pair of legs (Fig. 8A, C). In females the last segment aligned with the twelfth pair of legs (Fig. 9A), and there was never an obvious thirteenth segment, although a few animals had triangular wedges of cuticle protruding laterally behind the twelfth segment. Figure 9D shows the unusual trunk segments 8 and 9 of one female. The trunk ends in a complex anal segment with dorsal median protuberances containing a long telsonal seta and paired lateral lobes (Figs. 8A, C, 9A). Each lobe has a dorsal portion, and a ventral one containing a short, vaguely annulated, terminal projection (Fig. 9A). Figure 9E shows these structures developing in the first post-larval stage.

Fig. 3.

A, lateral view of right side of female with valves removed; B, plumose setae of 2^nd antennae; C, dorsal lobe of exopod of female T1; D, dorsal lobe of exopod of male T4. A2 = 2^nd antenna; Ap = probable Amoebidium parasiticum epibiont; as = anal segment; de 4, 9, 10 = dorsal lobes of exopods of T4, 9, 10; dl = dorsal lamella; ls = lateral seta of dorsal lobe of exopod; mn = mandible; ms = medial spine of dorsal lobe of exopod; T2, T4, T6, T8 = trunk legs 2, 4, 6, 8; veT4 = ventral lobe of exopod of T4. Scale bar = |$50\;\mu {\rm{m}}$|⁠.

Fig. 4.

A, medial view of right trunk of a bisected female; B, ventral view of posterior trunk of male. 2, 4, 6, 8 = trunk segments 2, 4, 6, 8; cs = comb setae; da = setae of dorso-lateral appendage; E1-6 = endites 1-6 of T1; ep = epipod of T1; fg = food groove; ll = lateral lobe of anal segment; ps = plumose setae on ventral lobes of the exopods; ss = posterior (screen) setae on E2 and E3; st = anterior stout setae on E2 and E3; T1, T2 = trunk legs 1, 2; to = toothed seta; ts = telsonal seta; ve = ventral lobes of exopods; vp = terminal projection of ventral portion of lateral lobe of anal segment. Scale bars: |${\rm{A}} = 100\;\mu {\rm{m}};\;{\rm{B}} = 50\;\mu {\rm{m}}$|⁠.

The trunk has ten pairs of homologous legs (thoracic or trunk appendages) in the male (T1-10), and twelve in the female (T1-12). The bases of opposing leg pairs are separated by a median food groove (Fig. 4A) that is widest anteriorly, and progressively narrows as the later leg pairs insert closer together. For the sake of simplicity, I follow the leg terminology of Martin and Belk (1988). Thus a leg may maximally consist of a lateral exopod with distinct dorsal and ventral lobes, a substantial, sac-like epipod, and, medially, an endopod with six inwardly-facing endites (labeled E1 to E6 from proximal to distal). Both the exopod and the endites contain muscles that enable their independent movement (Grube, 1853; Fryer and Boxshall, 2009). The first endite (E1, coxal lobe, gnathobase) insertsposteriorly to the other endites (Fig. 5A-E, I). The various legs differ markedly in the form, size, spination, and presence of their component parts (Sars, 1896; Martin et al., 1986; Fryer and Boxshall, 2009; Fig. 5). The male T1 forms an elaborate clasper (Figs. 1B, 10A, B), while the dorsal lobe of the exopod of T9 and T10 of females is a twisted rod (Fig. 9A). Diverse epibionts were seen sporadically on the leg setae of both sexes of L. brachyurus (Fig. 3C). Most common were sac-like structures with a stalked holdfast (Fig. 3C). The observation of small amebae emerging from sacs attached to the leg of a freshly killed Lynceus, exactly as described by Whistler (1965) for forms attached to cladocerans, indicate that these are Amoebidium parasiticum. This protozoan (Benny and O’Donnell, 2000) is a common epibiont on cladocerans and other aquatic arthropods (Whistler, 1965).

Fig. 5.

Trunk appendages. A, posterior view of female T3; B, anterior view of male T6; C, posterior view of male T7; D, posterior view of male T8; E, posterior view of male T9; F, anterior view of male T10 and opercular lamella; G, distal spines and setae on E5, E6, and ventral lobe of exopod of male T3; H, anterior view of female E2 showing stout anterior setae and basal regions of slender posterior setae on T2 and underlying T3; I, medial view of female T8-12; J, diverse spines on E3 of female T9. Ap = probable Amoebidium parasiticum; de = dorsal lobe of exopod; ds = dorsal setae on dorsal lobe of exopod; E1-6 = endites 1-6; ep = epipod; lc = less chitinized region of epipod where it joins the leg; ll = lateral lobe of anal segment; ol = opercular lamella; T8, 9, 11 = trunk legs 8, 9, 11; ve = ventral lobe of exopod. Scale bars: |${\rm{A-F}} = 50\;\mu {\rm{m}},\;{\rm{G-J}} = 20\;\mu {\rm{m}}$|⁠.

The trunk legs can be placed in three anatomical/functional groups. Legs of the first group (T2-6 of males and T1-6 of females) are the largest, and usually beat actively. While their position changes with leg beat, in the “resting position” each forms an elongate, shallow, rear-facing scoop that partially nests within the leg that precedes it. The dorsal lobe of the exopod bends posteriorly at its origin to form the lateral margin of the scoop. The large, flat E2 and E3 lie back medially, while the ventral lobe of the exopod and E4-6 turn backwards to lie nearly parallel to the trunk and form the floor of the scoop (Figs. 1A, 2A, 3A, 4A). The dorsal lobe has a lateral margin that bears prominent setae with fine setules and in-curving tips and has a curve which matches that of the adjacent valve (Figs. 3A, C, D, 5A). Its medial margin bears varied stiff spines (Figs. 1A, 3C, D), and has an inner curve that accommodates the adjacent epipod (Fig. 2A). Dorsally, there a few long setae (Figs. 2A, 5A). The ventral lobe of the exopod is cylindrical, has a scaly or roughly segmented surface (Fig. 3A), and bears long, flexible, plumose setae along its entire length (Figs. 4B, 5A-D, G, 10D). The epipod stretches dorsally. It has a distinct, less chitinized, and somewhat narrowed, attachment to the limb (Figs. 1A, 2A, 5C). In preserved material, it can be bent outward at this “hinge.” The E1 extends medio-dorsally and has anterior brush setae and stout apical spines with forward-facing teeth (Fig. 4A). The second and third endites are wide and prominent. They have a row of stout, toothed, anterior setae that angles slightly forward in front of a row of long, flexible posterior setae (screen setae of Fryer and Boxshall, 2009). The latter setae have regularly spaced, lateral setules with backward-bending tips (Figs. 4A, 5A, H, 10D). The fourth and, especially, the fifth and sixth endites are long and slender. Comb setae with a lateral row of perpendicular teeth, which variously appear blunt or sharp, are characteristic of E6 (Fig. 5G). The four distal projections of the anterior legs (E4-6 and ventral lobe of the exopod) often move together. They may curve posteriorly to extend nearly to the back of the animal (Figs. 1A, 2A, 3A, 4A, 5A), or reach out ventrally with their setae extending well beyond the valve margins (Fig. 11E).

Legs of the second group (T7-8) resemble the preceding legs, but are smaller (Fig. 3A), and beat with less amplitude. Their first endites face more medially (Fig. 4A), while E2 and E3 are more separated from each other, and have reduced screen setae (Fig. 5C, D). E4-6 turn backwards more sharply than those of T1-6 (Fig. 6A-C). T7 has an epipod which is absent on T8 (Fig. 5C, D).

Fig. 6.

A, Lynceus brachyurus swimming with ventral surface forward; B, whole mount of female previously placed with mixture of carmine and yeast particles; valves were removed, and animal somewhat straightened to show labrum; C, fecal pellet; D, string of pseudo-feces containing leaf debris; E, pseudo-feces particles released from inactive animals. (C and D stained with Alcian Blue, E stained with Toluidine Blue.) es = esophagus; fg = food groove; la = labrum; mn = mandible; mg = midgut; ov = ovary; pf = pseudo-feces. Scale bars: |${\rm{A}} = 3\;{\rm{mm}};\;{\rm{B}} = 1\;{\rm{mm}};\;{\rm{C-E}} = 500\;\mu {\rm{m}}$|⁠.

Legs of the last group (T9-10 in males and T9-12 in females) are small and quite paw-shaped (Fig. 5E, F) and move with a lateral-medial motion. E2-3 lack screen setae (Fig. 5I). E1-5 have stout, tooth-bearing spines, and are now quite similar (Fig. 5I, J). E6 is relatively large with long, terminal, spines that bend inward (Figs. 7A, 8A) and is inserted somewhat posteriorly on the leg (Fig. 9A) and is thus not visible in anterior view (Fig. 5E, F). As noted, the dorsal lobes of the exopods of T9-10 are modified into egg-holders in females (Fig. 9A). Dorsal lobes are absent from T11-12 of females and T9-10 of males (Fig. 5E, F). The ventral lobes of the exopods are short, with a few stout spines, and a scaly appearance (Figs. 9A-C, 10A).

Fig. 7.

A, ventral view of mandible pair, maxillules, and underlying head; B, ventral surface of head showing toothed strips; C, medial view of mandible; D, molar surface of a mandible (left is posterior) showing material between tooth ridges (arrows); E, fine structure of ridges on a mandible that was separated from soon to be molted exoskeleton; F, posterior view of labrum showing setal fields and median ridge of short, flexible spines; G, posterior view of labrum showing laterally compressed distal region and junction with the head; H, postero-lateral view of distal labrum, possibly showing precipitated mucus released from lateral groove. A1 = 1^st antenna; mu = mucus; mx = maxillules; oc = oral comb; ts = toothed strip on ventral surface of head and posterior surface of labrum. Scale bars: A, B, |${\rm{E-G}} = 50\;\mu {\rm{m}}$|⁠; C, |${\rm{D}} = 10\;\mu {\rm{m}}$|⁠.

Fig. 8.

Male Lynceus brachyurus. A, posterior view of trunk and anal segment of dorso-ventrally flattened specimen; B, dorso-lateral view of right side of posterior trunk; C, trunk and anal somite viewed from left; D, lateral view of posterior trunk (left trunk legs removed); E, F, posterior views of dorsal teeth on last trunk segments; G, lateral view of juvenile male. 9, 10, 11 = trunk segments 9, 10, 11; as = anal segment; da = dorso-lateral appendage; E6 = endite 6; ll = lateral lobe of anal somite; ol = opercular lamella; pc = peritrich ciliate; T3, T9, T10 = trunk legs 3, 9, 10; ts = telsonal setae; ve = ventral lobe of exopod. Scale bar = |$50\;\mu {\rm{m}}$|⁠.

Fig. 9.

Female and juvenile Lynceus brachyurus. A, dorsal view of posterior trunk and appendages of female (greatly flattened); B, enlargement from Fig. 2A showing opening of oviduct (arrow) between T10 and T11; C, D, posterior views of last trunk segments; E, ventro-lateral view of probable 1^st post-larval stage. as = anal segment; de 8, 9, 10 = dorsal lobe of exopods 8, 9, 10; dl = dorsal lamella; E6 = endite 6; ll = lateral lobe of anal segment; ol = opercular lamella; ov = opening of oviduct; T2, 4, 6, 8, 10, 12 = trunk legs 2, 4, 6, 8, 10, 12; ts = telsonal seta; ve = ventral lobe of exopod; vp = ventral projection of lateral lobe of anal segment. Scale bars: A, B, D, |${\rm{E}} = 50\;\mu {\rm{m}};\;{\rm{C}} = 20\;\mu {\rm{m}}$|⁠.

Fig. 10.

Male appendages. A, posterior view of left T1, showing ridges on e2 (arrows) that extend into notches on e3; B, anterior view of E3-6 of T1; C, possible duct openings within the enlarged setae on the distal margin of E3 of T1; D, anterior view of a portion of left T2; E, anterior view of portions of right T10 showing coating on some of the spines (arrows). de = dorsal lobe of exopod; E1-6 = endites 1-6; ep = epipod; ol = opercular lamella; ve = ventral lobe of exopod. Scale bars: A, B, |${\rm{D}} = 100\;\mu {\rm{m}}$|⁠; C, |${\rm{E}} = 10\;\mu {\rm{m}}$|⁠.

An opercular lamella, not found in other branchiopods (Sars, 1896), extends downwards behind the last appendages (Figs. 5F, 9A, D, 10A). Linder (1945) reasonably proposed that it represents the fused appendages of a terminal trunk segment. Males, but not females, have a previously unrecognized pair of two-segmented dorso-lateral appendages that extend backwards from the eleventh trunk segment onto the anal segment (Fig. 8A-D, G). Although they certainly lack homology with previous legs, these dorso-lateral appendages do seem to be modified trunk appendages, which along with the opercular lamella, emerge from the male eleventh segment.

Appendage Movement: Locomotion, and the Water Current

Freshly collected L. brachyurus swim about rapidly with the antennae and anterior legs extended and beating vigorously. Swimming direction changes abruptly and frequently. They may swim, hinge outward, near the bottom or the sides of their container, but when away from a substrate generally swim horizontally with the trunk within a few degrees of vertical (Botnariuc, 1947; Fig. 6A; 51 of 65 photographic images examined). The carapace valves are open about |$20^\circ$|⁠. The head is somewhat straightened so that the eyes are above the valve margins and the antennal bases reach outward through the antennal notch (Figs. 1B, 2B). The antennae extend antero-laterally (Fig. 11E), and move backwards about |$90^\circ$| in an elliptical stroke, bending at the grooves in the antennal peduncle, and against the antennal notch. Swimming on the back is not uncommon, in which case the antennae sometimes stop beating and the animal continues to move forward propelled by the legs alone. Initiation of swimming by a sedentary animal always involved the antennae. Experiments did not reveal any difference in swimming frequency between the sexes, although when swimming, males were more active and changed direction more often.

Fig. 11.

A, male manipulating edge of closed valves of female; B, anterior view of male attached at right angle to female; C, lateral view of attached pair showing brown “coating” on distal parts of male T2-9; D, subsequent image, showing depressed male trunk and the insertion of legs between female valves; E, recently dead male, showing position of antennae and brown markings on legs; F, ventral view of posterior trunk of flattened female. de = dorsal lobes of exopods; em = egg mass; mg = midgut; mx = maxillary gland; ov = ovary; p = peritrich ciliates; T7, 8 = trunk legs 7 and 8. |${\rm{scale}}\;{\rm{bar}} = 1\;{\rm{mm}}$|⁠.

Animals commonly move slowly over or under leaf fragments or along algal growth on the side walls or bottom of their container. The antennae extend outward, but do not beat, while the distal portions of the anterior legs (Fig. 4B) move against the substrate during their metachronal beat. Lynceus sometimes moves rapidly with both the antennae and the trunk appendages beating actively. Although there is clearly contact between the legs and the substrate, this seems more like swimming than walking.

A strange, twirling behavior (more than one revolution/second up to a minute in duration), with one valve on the substrate and the antennae beating vigorously, was occasionally seen. Animals sometimes attached by their posterior legs to a leaf, stick, or even the water surface. After extended periods in the laboratory, females were sometimes totally quiescent, with the valves tightly shut, and no movement of the appendages within. Repeated vigorous uptakes and expulsions of the animal from a pipette were required to produce leg movement.

After a time in the laboratory, most shrimp are lying on one valve on the substrate and are easily observed. The valves are slightly opened, the eyes and the edge of the head extend just beyond the valve margins, and the antennae lie back within them. The ventral lobe of the exopod and E4-6 of the anterior legs bend posteriorly to lie nearly parallel to the dorsal hinge. The anterior legs beat in a slow metachronal rhythm, the dorsal lobes of their exopods move back and forth between the valves, and a gentle water current, often no more than one-fifth the carapace length, flows inward mid-ventrally between them. The water current usually passes dorsally on either side of the trunk and exits dorso-posteriorly through the triangular space between the back of the hinge, the sides of the valves, and the top of the variably flexed trunk. It can, however, leave posteriorly beneath the trunk as seen by Fryer and Boxshall (2009) in L. gracilicornis. When the legs beat faster, the animal is still stationary, but current velocity increases. The valves open wider (about ten degrees), the antennae extend outward but do not beat, the distal endites of the legs are less bent and their tips extend just beyond the valve edge. When the legs beat still more vigorously, the trunk is straighter, and the carapace is open further and well separated from the trunk.

Food Sources

The gut contents and fecal pellets of field-collected L. brachyurus consist mainly of leaf debris, with small amounts of algae and zooplankton. In the laboratory, animals regularly move over and under leaf fragments, and along the bottom and sides of their container and a track is often visible where underlying material has been removed. Animals lived for weeks in dishes with a few pieces of decomposing leaf. To see if Lynceus might accelerate leaf breakdown, 12 containers were set up with roughly equal amounts of leaf material (not including the slow decaying oak leaves), and two liters of pool water. Twenty Lynceus were added to half of them and all were kept at constant temperature. After five weeks the leaves and surviving animals were removed, and the containers scored as to particle content (3 = much, 2 = moderate, 1 = little), revealing no difference between those with or without clam shrimp (2.0 vs. 1.8, |$t = 0.31,\;n = 12,\;p \gt 0.05$|⁠).

In sedentary animals, suspended material of diverse sizes can be seen to flow inwards with the water current from at least two carapace lengths away. Experiments showed that this suspended matter can be the food of choice. When animals that had been isolated for a few days in water that had passed through a |$30\;\mu {\rm{m}}$| plankton net were added to fresh pool water containing a leaf fragment, they did not immediately begin leaf scraping, but instead swam around, or settled on the bottom with the legs beating vigorously. Similarly, a swimming clam shrimp usually stopped swimming when a particle suspension (yeast, etc.) was pipetted nearby. Generally there was then a marked increase in the beat of the anterior legs, and a resulting increase in the strength and velocity of the in-flowing current. The bouts of twirling described above were sometimes seen. When fasted animals were placed in a suspension of carmine and yeast, these particles soon appeared in the gut (Fig. 6B). There were no scrape marks on material settled on the substrate, so the gut contents had come directly from the water.

When Lynceus that were added to water enriched with added zooplankton, they moved about actively, but never showed directed movement toward individual animals. Larger cladocerans and copepods bounced off an advancing shrimp. Copepod nauplii often changed course and swam sharply away when they were within one or two carapace lengths of a stationary Lynceus, presumably upon detection of the in-flowing water current. Some small zooplankton were drawn between the valves and retained, while others passed on through. An experiment showed that Lynceus can efficiently remove small crustaceans from water enriched with added zooplankton (Table 1). The dishes with Lynceus contained only a third of the zooplankton found in those without them. Comparing the two groups showed that in two days, six Lynceus had removed an average of 600 crustacean zooplankters. Copepods and cladocerans were removed at equal rates. At natural densities, plankton did not affect short-term survival. Three |$20\;{\rm{mm}}$| finger bowls with approx. 20 L. brachyurus and leaf fragments that were transferred every third day to freshly collected pool water had a similar 30 day survival rate |$(63\%)$| to three that were transferred to water that had passed through a plankton net |$(54\%;\;t = 0.54,\;n = 6,\;p \lt 0.05)$|⁠.

Table 1.

Capture of crustacean zooplankton. Zooplankton present after 48 hours in finger bowls with or without six Lynceus brachyurus.

	With 5 Lynceus			Without Lynceus			Significance
	N	Mean	SD	N	Mean	SD
Copepods (including nauplii)	8	233	64.6	8	811	155	|$t = 9.72,\;p \lt 0.001$|
Cladocera	8	15.8	4.65	8	51.4	28.6	|$t = 3.47,\;p \lt 0.05$|

Food Collection and Processing

As an animal moves across a leaf, the trunk bends downwards, pressing the various spines and setae on the legs (Fig. 4B) against the substrate as the legs beat. In addition to scraping, there is considerable manipulation of larger items. The posterior legs with their stout, medially facing spines on E1-5 (Fig. 5E-F, I, J) and the distal “basket” of spines on the larger and more posterior E6 (Figs. 8B, 9A) hold bits of leaf, stem, and mucus-entangled material. These are probed by the tips of the anterior legs. Small particles are dislodged and carried inward with the water current, while larger objects are released posteriorly, or rotated by the water current and probed again before eventual release. When a particle suspension was pipetted near a sedentary animal, this usually caused the speed and scope of leg beat to increase and the material to be collected. Sooner or later, and often sooner, it was brushed posteriorly by the legs, or caused the water current to reverse, or the animal to swim away. In other cases, high particle concentrations entered, but the collecting mechanisms were by-passed, and most moved out with the exhalent current.

Adhesive secretions are involved in the collection and transfer of incoming material. The dense mass behind the mandible (Fig. 7B) shows adhesion of carmine/yeast particles, small masses of leaf material often stick briefly to the posterior legs while being manipulated, and the pseudofeces discussed below are cohesive. The labrum contains large labral glands (Martin and Belk, 1988) arranged in two pairs. In its normal position, the distal portion of the labrum lies within the food groove (forced downwards in Fig. 6B), so its secretions could drift backwards among the legs. Figure 7H may show the precipitation of mucus leaving the posterior (dorsal) surface of the labrum.

Whether removed from the substrate or already suspended, much of the material brought inward with the water current, including particles as small as |$20\;\mu {\rm{m}}$| in diameter, is absent from the exhalent current and so has been retained. Several structures are involved in its capture. Long posterior setae (screen setae of Fryer and Boxshall, 2009) on the large E2 and E3 of the anterior legs characterize two thirds of the medial surface of the trunk (Fig. 4A). When the anterior leg pairs of preserved animals were moved backwards, their E2 and E3 turned inwards and the anterior setae of opposing legs intermeshed. These setae have setules with inward-facing, back-turned tips that often contained debris (Figs. 5A, H, 10D). Animals placed with a carmine/yeast mixture accumulated particles in this region (Fig. 6B). Additionally, the plumose setae of the ventral lobes of the exopods of the anterior legs have a tangle of setules along their length (Fig. 5A, G) that often contains debris (Fig. 4B). Due to its attachment more proximally on the leg and its long setae (Fig. 5A, B) the ventral lobe is longer than any of the endites. It is also more flexible, perhaps due to the creases in its exoskeleton (Fig. 5A, D). In a stationary animal with actively beating legs, these setae were observed extending outwards beyond the valve margins to form a rear-facing fan that was pulled through the water on the back stroke. On the recovery stroke, the ventral lobe bent backwards and the fan collapsed, and was dragged through the comb setae on the E6 of the following leg (Fig. 5G). On two occasions, a copepod nauplius was seen to be impaled on one of the small spines on the medial edge of the dorsal lobe of an exopod of an anterior leg (Fig. 3C, D).

Most material seems to enter the food groove toward the back of the animal, also observed by Liévin (1848) and Fryer and Boxshall (2009). The food groove of trunk segments 5-7 seems to be a “zone of decision.” Material is either moved forward to the mouth, or is expelled posteriorly without entering the gut (Fig. 6B). By analogy with the mucus-entangled material released by bivalve mollusks without being ingested, the mildly processed adherent material discharged from the branchial cavity of Lynceus is called pseudo-feces. It can be formed and released within minutes. The nature of this pseudo-feces varies greatly. When an animal is scraping, it consists of plant debris (Fig. 6D). When the shrimp is placed with a carmine/yeast mixture, the pseudofeces appears identical to the material being moved into the gut (Fig. 6B). Animals kept for a week with a few leaf fragments were generally inactive, and had presumably spent long periods lying on one side with the carapace valves ajar and the legs beating gently. The sediment in their dish contained fecal pellets, but was dominated by flocculent bits very different from leaf debris (Fig. 6E) that must have been pseudo-feces.

Food is moved forward along the food groove by the spines and brushes of the E1s of the anterior legs. The E1s attach behind the other endites of their leg (Fig. 5A-E, I), have a fibrous endoskeleton (Fryer and Boxshall, 2009), and move forward when the “power stroke” drives the rest of the leg backward (Botnariuc, 1947; Fryer and Boxshall, 2009).

A large, soft, and deformable labrum extends down from the head (Fig. 6B). Due to the curvature of the body, its proximal region is pressed against the overlying mandibles, while the somewhat laterally compressed distal portion (Fig. 7F, G) reaches back to the second pair of legs to form a floor beneath the anterior food groove (Grube, 1853; Fig. 7B). The E1s of T1-3 are elongate and extend medially and dorsally into the food groove (Figs. 4A, 5A, 10A; Fryer and Boxshall, 2009). When viewed posteriorly, the E1 of both male (Fig. 10A) and female T1 are seen to consist of a large, and presumably muscle-containing, base and a slender medial portion. This extension is flattened dorso-ventrally (Fig. 10A), and can be forced between the head and the labrum to carry food forward over the fine hairs of the distal labrum (Fig. 7F, G). While not obvious in the male (Fig. 10A), the E2 of the female T1 has an indented medial area that lacks the usual slender setae (Fig. 4A) and must facilitate the lateral passage of the leg past the labrum. Material is passed to the brush setae of the maxillules, which move horizontally above the labrum and place items between the mandibles (Fig. 7A; Fryer and Boxshall, 2009).

The mandibles are large with pointed tips that articulate with the exoskeleton at the junction of the head and neck (Fig. 2A; Grube, 1853; Sars, 1898; Martin, 1989). The mandibles extend distally, bulge outwards (Figs. 1B, 7A), and then bend sharply inwards and compress to form elongate molar surfaces (Fig. 7A, C, D). The flattened region just proximal to the molar surface (Fig. 7A) indicates the depth of the medial penetration of the mandible between the proximal labrum and the overlying head. The molar surface contains twelve parallel tooth ridges, and two small and one large posterior spines (Fig. 7C). A soon-to-be-molted mandible (Fig. 7E) has more prominent denticles on its tooth ridges than are seen later (Fig. 7A, C, D). Other specimens viewed from other angles show lateral teeth on the tooth ridges than are not apparent in Fig. 7A, C-E. Material is carried forward between the tooth ridges of the mandible (Fig. 7D), and both above and beneath them (Fig. 6B).

The mandibles move together, but somewhat irregularly, at about a |$45^\circ$| angle to the long axis (Figs. 1A, 2A). They rotate against a roughly X-shaped region with curved limbs on the ventral surface of the head (Fig. 7A, B; figured by Grube, 1853) and a roughly similar structure on the posterior surface of the proximal labrum (Fig. 7F). Both have slender teeth laterally, and blunt, cornified, teeth medially where the component arcs of the structure come together, and the mandibles meet. The large transverse muscle that attaches within the hollowed areas of the mandibles (Fig. 7C; Martin, 1989) must cause the opposing tooth ridges of the mandibles to contact each other, and in all probability interdigitate, as they rotate past the cornified median areas. The flattened mouth (Fig. 7G) and the forward-facing teeth of the “oral comb” (Martin et al., 1986; Fig. 7G) appear to contact the mandibles and so will be forced to open further as the mandible’s widening molar surfaces (Fig. 7C) rotate forward and apart. On the return stroke, material lying between the mandibular ridges (Fig. 7D) will be scraped off, and presumably carried inward by peristaltic contraction within the short esophagus.

A small gizzard at the junction of the esophagus and foregut, described by Fryer and Boxshall (2009) in L. simiaefaciesHarding, 1941 was apparent in dissected L. brachyurus, but not in living animals or whole mount slides. “Capacious ducts” (Sars, 1898) from the lobes of the hepatopancreas in the rostrum and proximal labrum enter laterally into the clear, thin-walled, fore-gut. Small particles and flocculent material were seen moving back and forth between the left and the right ducts, sometimes extending well into them for a few seconds before reversing direction. About ten reversals occurred within a 30 second period. The ducts did not change shape during particle movement, so material is probably propelled by cilia as has been shown for spinicaudate clam shrimp (Schlecht, 1979). In the Lynceus exposed to a suspension of carmine/yeast, there was usually little transfer of particles to the lobes of the hepatopancreas (Fig. 6B), although in some individuals it was quite substantial. The midgut contents are compacted from the middle onwards. The relative positions of items within the gut can change as seen by the carmine particles that have by-passed pre-existing detrital material (Fig. 6B). The gut narrows sharply a few segments before the anal segment (Figs. 6B, 11F). Defecation is a response to material coming through the gut. Animals held in filtered water for over a day had full guts and had not released any fecal pellets. When food is being taken in, compact feces are regularly expelled by peristaltic contractions within the posterior gut. Individuals that are actively scraping can release well over 10 gut lengths of feces per day. The feces usually break apart quickly, but in a resting animal may extend for a body length, supported by the lateral lobes of the anal segment and the opercular lamella. In some animals exposed to the carmine/yeast suspension, loose collections were expelled within a sac-like peritrophic membrane. For most fecal pellets a peritrophic membrane was not obvious (Fig. 6C).

Reproduction

Lynceus brachyurus has numerous sexual dimorphisms. The most obvious is the transformation of the male first legs (T1) into sturdy claspers that grasp a female valve during mating (Figs. 1, 10A, B). The claw (E6) is long, rigid, and strongly curved (Figs. 1B, 10B), with a basal space that accommodates the rim of the female valve as it fits around and under it. E4 and E5 (smaller palp and larger palp respectively in Sigvardt and Olesen, 2014) are soft with long, terminal, sensory setae (Fig. 10B). The former remains outside the female valve, the latter goes beneath it beside the claw. The E3, which elsewhere is similar to, and barely separate from, the E2 (Fig. 5A) is now the enlarged palm of the chela (Fig. 10A, B). The sac-like spines on its anterior margin (Fig. 10B) deform when poked, and might be expected to flatten against the female valve. They appear to have a terminal opening (Fig. 10C) that might release adhesive material. However the videos of Sigvardt and Olesen (2014) and my observations show the palm not tightly pressed against the carapace and the claws changing position along it. The distal margin of the E2 has two projections that fit into notches in the palm and may allow the leg to resist torsion when an attached female pulls away (Fig. 10A). Muscles in the enlarged base of the E1 (Fig. 10A) must aid attachment.

Although less obvious than the claws, there are other features of the male appendages that relate to mating. The small, two-segmented, first antenna has short papillae with a presumably sensory function, and flicks back and forth actively. Sars (1896) reports it as being “perhaps rather larger” and more papillose in males, while Botarniuc (1947) states that the first antenna of the sexes are of similar length, but twice as thin in females. While there was overlap in the data sets, measurements from prepared slides showed the outermost segment of the first antenna to be both longer (0.38 vs. |$0.31\;{\rm{mm}},\;t = 5.06,\;n = 22,\;p \lt 0.001$|⁠) and wider (0.13 vs. |$0.11\;{\rm{mm}},\;t = 3.75,\;n = 22,\;p \lt 0.01$|⁠) in males than in females. Sars (1896) pointed out that the E5 of the male T2 has a row of peculiar, short, medial spines (13 or 14 in the specimens I examined). They point slightly forward, and become larger and change from 3-pointed to 2-pointed hooks as they extend distally (Fig. 10D). Additionally, there are a few strong, distal, spines on E4 and E5, while E6 has sharp spines, rather than the usual comb setae (Figs. 5C, 10D). The atypical setae on T2 could catch the narrow edges of the closed valves of the female when the male is turning an inactive female during the early stage of the mating sequence (Fig. 11A).

Since the male inserts the end of the trunk between the female valves during mating, sexual dimorphism in the posterior legs was expected. The spination of these legs is prominent and diverse (Figs. 4A, 5C-F, I, J, 11E), but greater development in males was not recognized. The dorsal spines on the posterior trunk segments were dramatically stouter in some males (Fig. 8F), and could contact the female legs during the deepest trunk insertions. However, the development of these spines varied considerably and they were often similar in the two sexes (Figs. 8A-C, E, 9A, C). The brush of hairs extending posteriorly from the unique dorso-lateral appendages on the male eleventh trunk segment (Figs. 4B, 9B-D) could be stimulatory.

The male rostrum has a flat, or slightly concave, anterior margin (Fig. 1B) and a carina (keel) that ends abruptly to rise almost perpendicularly from the front (Fig. 1B, C), producing a rigid T-shaped contact when pressed against a female valve (Fig. 11B). The female rostral carina slopes gently backward (Figs. 2A, 3A). It always has a pointed tip (Figs. 2B, 3A), although individuals varied in the degree of its extension beyond the lateral spines. Developing males have a medial point on the rostrum. At the molt of maturity, the flat rostral edge appears and the male is able to grasp a female (Valtonen, 1966). In Week 18 of 2007 North Pool collection (Table 3), all of the males recognized had a pointed rostral tip, and none of the females bore eggs. A week later seven males had a flat rostrum and 12 females were ovigerous.

In females the hinge attachment between the carapace valves is about half the valve length (Fig. 2A). It is less than half in males which allows greater extension of the head and trunk than would be possible with a broader attachment, and may facilitate grasping the female with the T1s, or insertion of the trunk between the female valves during mating. Males are typically tan or light brown, while females are usually redder (Fig. 11B), presumably due to greater concentrations of hemoglobin (Martin, 1992). Some females maintained for a month or longer in the laboratory lost the reddish-brown color, while others retained it. An unexpected observation was that mature males commonly, but not always, have brown coloring on the distal margins of their endites (Fig. 11C, E). This appears to be a coating and gives the endites a “heavy” appearance. It is quite distinct from the color of hemoglobin, was never noticed in females, and is visible enough through the carapace to allow the sexes to be distinguished. It is probably the material seen on parts of a male T10 (Fig. 10E). Brown coloration is visible on the legs of male L. brachyurus in the videos of Sigvardt and Olesen (2014).

The paired ovaries extend close to the end of the trunk. A modest number of fairly large eggs (about 40 in Fig. 11F) are visible posteriorly. The opening of the oviduct was seen in one specimen, seemingly between the bases of T10 and T11 (Figs. 9B, 2A). The openings of the male ducts were not located. If they are on the eleventh segment, perhaps in association with the dorso-lateral appendages (Fig. 9B-D), this would agree with Grube (1853) and Sars (1896) who suggest they are in the same place as those of the female. However, Linder (1945) illustrates histological sections showing the passage of vasa deferentia through the lateral lobes of the anal segment.

Females have unique, paired, dorsal lamellae that attach on trunk segments 9-12 and extend back along part of the anal segment and laterally around the opening of the oviduct and the bases of the dorsal lobes of the exopods of T9 and T10 (Liévin, 1848; Figs. 2A, 3A). Post-maturity females generally bear two dorso-lateral masses of adherent eggs under the carapace (Figs. 2A, 11D). The number of eggs carried varied greatly, but seemed fairly equally distributed between sides. Each egg mass attaches to terminal tufts of filaments (Fig. 9A) on the modified dorsal lobes of T9 and T10 (Figs. 2A, 3A, 9A), and in some cases, to the underlying tip of the anterior-most lobe of the dorsal lamella.

A container of L. brachyurus virtually always shows some sexual activity, with males attempting attachments and sedentary, or rarely, swimming pairs. Often several males gather around a female, with the first legs extended and the claws (E6) turned forward. Males sometimes swam past a female, turned abruptly, and attempted attachment. However unlike the anomopod branchiopod Chydorus sphaericus (van Damme and Dumont, 2006), the addition of water from containers of females did not produce a detectable change in the swimming behavior of isolated Lynceus males. Males regularly attempted attachments with each other (also Sigvardt and Olesen, 2014). If one or both claws hooked under a valve of another male, this usually elicited vigorous swimming, and escape by the captured male. In older collections with less active males, extended periods of holding, as well as trunk insertions between the valves of the inactive male, were observed.

In the laboratory, a reproductive sequence typically began with a male with claspers open and pointing forward, approaching an egg-bearing female that was lying on one side with valves firmly closed (Fig. 11A). The male moved over and around the female with the antennae and thoracic appendages beating actively and brushing her carapace, often turning her over repeatedly. This “courtship” had variable results: the female valves remained closed during this manipulation, the female swam away without being grasped, or the valves opened slightly allowing insertion of the male claws.

In a successful attachment, the male is at a right angle to the female with the head pointing toward the hinge. The rostrum and claws produce a stable connection that can withstand considerable buffeting by other males or the experimenter. The male usually attached somewhat behind the female mid-line, and fortunately for the observer, most often to the lower valve, so that the head pointed down and the trunk was easily seen as it extended upward (Fig. 11C and D, but not B). Attached pairs usually lie on the substrate, although in both field and lab, they were occasionally seen swimming, usually in a spiral manner. The female seemed to be carrying the attached male, as her antennae beat most actively.

After initial attachment, the female valves close against the male claws, and her thoracic legs beat feebly, or not at all. The male legs beat vigorously. In this, the longest lasting reproductive behavior, only the male rostrum, antennae, and claspers actually contact the female valves (Fig. 11B, C). The male trunk is bent ventrally and the dorsal lobes of the exopods of the anterior legs drive water against the female. Eventually the female valves open further, her appendages beat faster, and the trunk bends towards the male. Slow motion video showed the terminal endites of the female E3 and E4 extending between the male valves and being massaged by scissors-like movements of the endites of the male E2 (Fig. 10D). There are brief insertions of the male trunk (Fig. 11D). These last for a few seconds and involve vigorous brushing of the female appendages by the male legs. During most insertions, the male valves are open about |$60^\circ$| and the anal segment and opercular lamella rest outside the edge of the female valve. The trunk may straighten somewhat, prying the female valves further open. Eventually there is a brief insertion of the entire trunk, bringing the anal somite beneath the down-turned female trunk. The male trunk flexes a few times, presumably releasing sperm, and then withdraws. There are then several vigorous dorso-ventral movements of the female trunk and a few eggs appear laterally. The openings of the male ducts were not located. If they are on the eleventh segment, perhaps in association with the setose dorso-lateral appendages (Fig. 8B-D), this would agree with Grube (1853) and Sars (1896) who suggested that they are in the same place as those of the female. However, Linder (1945) illustrates histological sections showing the passage of vasa deferentia through the lateral lobes of the anal segment. On a single occasion, a coherent stream of small particles was seen to leave in the dorsal exhalent current of a paired female, turn sharply, and re-enter ventrally with the inflowing current. The particles were unlike anything seen previously, and must have been sperm. Two factors help retain sperm within the female carapace. The female leg beat, and thus the speed of the departing water current, slows, and the trunk bends sharply downwards so that the scoop-like dorsal lamellae (Figs. 2A, 3A) are almost vertical and keep sperm close to the underlying openings of the oviducts. Pairs could remain attached with periodic insertion and removal of the male trunk, and a gradually accumulating egg number, or could alternate periods of separation and reattachment. Eggs are guided forward by the hood-shaped dorsal lamella to attach to the setae on the dorsal lobes of the exopods of T9 and 10. Subsequent eggs move dorsally and push the existing ones forward to enlarge the egg disc. Eventual separation of the pairs resulted from male disinterest, not female struggle.

Ovulation only occurs after male stimulation. None of the ovigerous females held without a male carried eggs two days later, while all of those with a male did (Table 2, Day 2). When a new male was added to members of the former group, egg masses were produced (Table 2: Day 12). When females were isolated until existing eggs were shed and then placed with a male, the mating process described above was greatly compressed and male attachment and trunk insertions, followed by female trunk flexions and egg release could occur rapidly. Once, a mass of about ten eggs was seen within 4 minutes of male contact, and in several cases eggs appeared within 15 minutes.

Table 2.

Eggs deposited, egg bearing status, and survival of Lynceus brachyurus. The experiment was started on 11 May 2006 with replicate sets of a freshly collected male and ovigerous female or a lone ovigerous female. Every second day the egg mass born by each female was estimated and coded: 0 for none and 1, 2 or 3 for a small, medium or large egg mass. The eggs in each jar were counted and the water changed. On Day 8 a freshly collected male was added to four of the female-only jars.

Day	Male and female |$(N = 7)$|				Female alone |$(N = 8)$|
	Eggs deposited (mean (range))	Egg mass (mean (range))	Survivors		Eggs deposited (mean (range))	Egg mass (mean (range))	Survivors
				F			M	F	M
2	240.9 (158-325)	2.4 (2-3)	7	7	139.9 (36-247)	0	8
4	183.0 (22-331)	1.3 (0-3)	7	7	6.0 (0-32)	0	8
6	39.9 (3-160)	1.2 (0-2)	7	4	4.6 (0-30)	0	8
8	15.0 (0-55)	0.3 (0-1)	7	0	0.4 (0-1)	0	8
10	4.1 (0-21)	0.2 (0-1)	7	0	1.5 (0-4)	3	4	4
12	1.9 (0-13)	0	7	0	330 (137-481)	1.5 (0-3)	4	4
14					9.3 (0-18)	0.5 (0-1)	4	2
Eggs released per female over 4 days (mean (range))
D0-D4		423.9 (347-534)
D10-D14						331.3 (155-483)

Table 3.

Numbers of Lynceus brachyurus in weekly collections from the North (N) and South (S) Pools. 0, none found; –, collection not made; L, water too low for normal collection; E, first observation of an egg-bearing female.

	Pool	Week														Precipitation (mm)
		14	15	16	17	18	19	20	21	22	23	24	25	26	27	Jan-Feb	Year
Initial date		2/4	9/4	16/4	23/4	30/4	7/5	14/5	21/5	28/5	4/6	11/6	18/6	25/6	2/7
2002																76	1044
2003		–	–	–	–	–		Lynceus present in both pools through week 24.					0	–	0	141	1216
2004	N	0	0	–	0	0	0	0	–	0	0	0	0	0	L	138	1055
	S	0	0	–	0	0	0	0	–	0	0	0	0	0	0
2005	N	0	0	6	0	63	90	67E	41	127	1	22	10	1	0	272	1183
	S	0	0	0	1	22	42	40E	11	64	approx. 20	36	2	0	0
2006	N	0	–	0	0	2	0	2E	1	–	–	–	–	0	0	98	1032
	S	0	0	3	0	29E	12	16	22	–	–	–	–	0	0
2007	N	0	0	0	39	13	41E	65	69	57	19	5L	0L	0L	0L	150	1072
	S	0	0	0	15	26	158E	89	132	approx. 170	approx. 100	116	approx. 25	0L	0L
2008	N	0	0	0	0	105	302	68E	208	approx. 80	approx. 250	91	0	0L	0L	148	1041
	S	0	0	0	0	48	approx. 330	12E	approx. 140	approx. 170	143	141	0	0	0L
2009	N	0	–	0	0	6	15E	2	1	1	9	1	0	–	0L	87	846
	S	0	–	0	9	164	approx. 750E	approx. 160	approx. 225	approx. 100	77	20	19	0	0
2010								Present								94	936
2011								Present								153	1408
2012								Absent								119	917
2013									Present							108	–

The number of eggs held by a female varied considerably. Six females collected individually on 30 May contained 0, 0, 0, 123, 161 and 223 eggs. Eight egg-bearing females isolated upon collection eventually released between 36 and 282 eggs (Table 2). In the crowded populations in laboratory containers, virtually all were ovigerous. Sometimes a few eggs were shed when a female was disturbed, and on rare occasions, females were seen to rise to the water surface and twirl about, releasing eggs.

Ten male-female pairs that were examined twice a day for four days provided information on mating behavior and egg retention and release. (One male was dead on Day 4.) Males were attached to females on 31 of the 78 observation periods, including six times in the observation prior to the release of a large egg mass. There was no difference in the frequency of attachment between the first and last two days of the experiment (⁠|$45\%$| vs. |$34\%,\;{\chi ^2} = 0.95,\;p \gt 0.05$|⁠). Females carried an egg mass on 74 of the 78 observations. The four females without one were in containers with many eggs (143-168) that had been released since the last observation. Only one had an attached male, and all four bore eggs subsequently. Over the four day period, released eggs were found in 33 of 79 observations, and ranged from one to six releases and 34 to 520 total eggs per female. Egg release was highly clumped with no releases between 39 and 87 eggs and |$92\%$| of the eggs found in 17 releases between 88 and 234 eggs. The large releases were clustered around 160 eggs (eleven were from 143 to 193). Two females produced three large releases during the four day period, four had two, three one, and one none. In each multiple release, there were one or more intervening observations where the male and female were not attached. The above eight releases of a second or third large egg mass allowed egg minimum retention times to be estimated. Two females held some eggs for at least 13 hours, two for at least 25 hours and four for at least 37 hours. Maximum egg retention may be about 48 hours. While none of the four non-ovigerous females placed with a male had released eggs within 48 hours, all eight egg-bearing females kept without a male had released them (Table 2: Day 10, Day 2).

The females survived well in the small containers used in the egg production experiments, while the males did not. Six days after the pairs were established, seven of the males in the above ten pairs and five of the paired males in Table 2 were dead. All 21 females in these pairs were alive and active.

Life Cycle

The nauplius larva of Lynceus has a distinctive large dorsal shield (Grube, 1853) that is easily recognized in plankton samples. Larvae retained by a |$30\;\mu {\rm{m}}$| plankton net were first seen in the 27 February 2009 and 15 March 2010 collections (Fig. 12). The combined collections show four larval stages with median dorsal shield widths of roughly 0.28, 0.35, 0.44 and |$0.55\;{\rm{mm}}$| (Fig. 12), comparable to the widths of 0.26, 0.36, 0.48 and |$0.60\;{\rm{mm}}$| reported from Russia by Monakov and Dobrynina (1977). As expected, development varied with temperature (Fig. 12). In 2009, larvae did not molt during the seven days between weeks nine and ten when the mean daily maximum air temperature was |$4.3^\circ {\rm{C}}$|⁠, molted twice in the twelve days between the week 10 and 11 collections when it was |$16.0^\circ {\rm{C}}$|⁠, and only once in the seven days between weeks 11 and 12 when it was |$13.8^\circ {\rm{C}}$|⁠. In 2010, the larvae molted twice in six days when it was |$17.0^\circ {\rm{C}}$|⁠. In each of three years studied, there was a progression in larval size over time. Larvae were found in four of the 12 weeks sampled in 2009 and two of 12 in 2010. None were found after 2 April (Fig. 12).

Fig. 12.

Occurrence and dorsal shield width of Lynceus brachyurus larvae from the South Pool, 2008-10. Up to 50 individuals were measured from each collection.

In the 2007-2008 season, samples of water and sediment were collected at intervals, maintained in the laboratory, and examined periodically for the presence of shrimp. No Lynceus appeared in ten samples collected between 28 December and 29 January. By contrast, four of the eight samples obtained between 3 March and 2 April, when the pool contained larval (Fig. 12), but not post-larval stages (Table 3), eventually produced 11 sub-adult Lynceus (28 to 36 days after sample collection). Laboratory conditions were thus inadequate for hatching of the eggs (cysts) that must have been present, but did support the growth and metamorphosis of existing larvae, and the growth of juveniles.

The smallest bivalved Lynceus noticed was found in the 11 April 2008 plankton sample (Fig. 9E). It appears to have seven pairs of trunk appendages, the number found by Grube (1853) in the earliest post-larval stage. The setae on the antennae and ventral lobes of exopods have fewer setules and the rostrum is less formed and relatively smaller than the adult. The lateral setae on the dorsal lobes of the exopods are less numerous (only 13 on T3 in Fig. 9E), more flexible, and relatively much longer (four or more times the width of the exopod versus less than twice) than those of the adult (Figs. 1A, 3A). The valves of animals at this stage were covered with small adherent particles. A later stage (Fig. 8G) was found in an early net collection (26 April, 2007; Table 3). It is larger than the above individual, with a more developed rostrum. There are considerably more than thirteen setae on the dorsal lobe of an anterior exopod, but they are still relatively longer and more flexible than those in the adult. It is similar to the animal suggested by Olesen (2005) to be a possible first juvenile stage.

Adults, as indicated by ovigerous females, appeared in early May (Table 3). In years of abundance, L. brachyurus comprised the bulk of the biomass in both North and South Pools by mid-May, but never approached the density of 90 animals per liter of pool water indicated for a pool in Wisconsin by Schneider and Frost (1996). Lynceus was present in both North and South Pools each year, except for 2004 and 2012 (Table 3). It was never found in the nearby, shorter duration, West Pool.

Lynceus brachyurus continues to molt after maturity. Egg-bearing females were measured with carapace widths of 2.5 and |$3.4\;{\rm{mm}}$|⁠, and flat rostrum (= mature) males at 2.5 and |$3.2\;{\rm{mm}}$|⁠. However, size cohorts are not apparent in the combined data, so the number of post-maturity molts may be limited. Sequential egg releases definitely occur without an intervening molt as molted valves were not seen in the repeated examinations of containers with females that produced multiple egg masses. While most molted exoskeleton seems to disappear rapidly, and in one case was seen being eaten by a newly molted animal, the valves are long lasting, and would have been noticed if present.

The weekly samples showed females to be more numerous than males in May (Fig. 13; other data), while males were more abundant later. The three large collections from Week 24 (Table 3) were 55, 91 and |$84\%$| male for a total of 265 males and 83 females.

Fig. 13.

Carapace lengths of post-larval Lynceus brachyurus collected from the North Pool in 2008. The entire 11 June collection was measured and a sample of about 50 individuals for the other dates. Ovigerous females were first found in Week 20.

Shrimp were gone by late June, in a few cases with declining water levels, but often before the pools were much reduced (Table 3). 2011 provided a dramatic example of the effect of temperature. Lynceus was abundant in the South Pool on 31 May, much reduced on 3 June, and absent, despite extensive sweeping, on 9 June, a date when they were common in most previous years (Table 3). The mean daily high air temperature for the total ten-day period was |$31.2^\circ {\rm{C}}$| (range |$26.7 - 34.4^\circ {\rm{C}}$|⁠), compared with the long term average high of |$26.2^\circ {\rm{C}}$|⁠. The pool remained quite collectable through July.

When maintained for 48 hours in dishes that forced contact with a predator, (backswimmer, dragonfly larva, beetle larvae or adult, or larval salamander), about half the Lynceus with dytiscid beetle larvae were killed, while the great majority of those with other predators were still alive. The above mentioned Amebidium parasiticum and other filamentous epibionts had a scattered attachment to the setae (Fig. 3C), but never seemed dense enough to cause mortality. Most field-collected L. brachyurus had attached peritrich ciliates. These were particularly abundant in the tail region, around the rostrum, and on the posterior valve margins. Toward the end of the season, they often covered the entire carapace (Fig. 11B), and would seem to interfere with swimming. Throughout the season some animals were noticed with rather evenly spaced, irregular, white, blotches on the respiratory membrane. These sometimes distorted the carapace, and, in the laboratory, soon led to death. However this condition was uncommon and no more frequent later in the year. A condition that was more common late in the season, but never predominant, was for the coils of the maxillary gland (Martin, 1992) under one of the carapace valves to be red and opaque rather than light and translucent (Fig. 11B). This may be a sign of senescence, or a cause of it and a contributor to mortality.

Discussion

Habitat, Water Current, and Feeding

Lynceus brachyurus can be an active swimmer (Fig. 6A), but in the laboratory at least, is usually found lying on the substrate or moving over it. Although all study animals were obtained by pushing a net through the water above the bottom, the collections often contained benthic isopods and mollusks, so an unknown portion of the Lynceus may have been stirred up by the netting procedure. The abundant debris in the gut contents, laboratory behavior, and three month survival in the laboratory with a few leaf fragments, support the conclusion of Fryer and Boxshall (2009) and others that Lynceus are benthic animals that scrape or brush material from the substrate.

Suspended material is drawn into sedentary animals with the water current. This is largely retained and is a more important food source than is generally recognized. When a suspension of carmine and yeast was added to their container, animals that were swimming immediately settled on the substrate, and increased their leg beat and the speed of the incoming current, indicating that this was a preferred food, best collected while sedentary. The current brings in particles from at least two carapace lengths away and at times appears strong enough to re-suspend material loosely settled on the substrate.

Experiments showed efficient capture of copepod nauplii and other crustacean zooplankton. This provides support for the suggestion made without further comment by Sars (1896) that, based on the structure of the mandibles, small crustaceans form a large part of the diet of L. brachyurus. Leaf detritus is on the bottom, suspended particles would be expected to be more abundant there, and, in the laboratory at least, mating is benthic. So it is tempting to speculate that a major function of the energy expended in episodic swimming may be to increase sampling of the water column and potential encounter with the nutrient-rich small zooplankton.

The water current brings water past the blood-containing respiratory membrane that lines the carapace, and carries in suspended leaf scrapings and all other potential food items. Several features of Lynceus anatomy relate to the production and direction of this current. The dorsal lobes of the exopods of the anterior legs are large with a width that is considerably increased by lateral setae with back-turned tips (Fig. 3A). They press laterally against the valves and must draw in water during their metachronal beat. The sub-spherical carapace (width about two-thirds its length) provides space for the large dorsal lobes and allows more water to be moved with each stroke than would be possible within the flatter valves of a bivalved spinicaudatan or anomopod (e.g. Daphnia) branchiopod of similar size. The enlarged head (Figs. 1A, 2A, 3A) can form an anterior seal with the valve edges. When the valves are barely open, the sides of the rostrum with their lateral fringes of fine hairs (Martin and Belk, 1988) must press against the valve membranes. At slightly wider valve opening, the body straightens, raising the head and maintaining contact between valves and rostrum. When open further, and in swimming (Fig. 6D, E), the head protrudes and contacts the valves at a thickened region behind the antennal notch (Figs. 2B, 6D, E). In all positions, the smooth, flat, ventral surface of the head (Fig. 3A) directs water dorsally and posteriorly. When the legs are beating gently, posterior entry or exit of water seems to be restricted by the last legs and the lateral lobes of the anal segment which press laterally against the valve membranes, and by the opercular lamella which blocks the back of the food groove (Figs. 5F, 9D, 10A). Finally, there are the peculiar patches of spines and filaments on the back of the posterior trunk segments (Figs. 9A-C, E, F, 10A, C, D). Although smaller, they have a similar shape and orientation to the denticles (riblets) on shark skin that reduce drag and increase shark swimming speed (Oeffner and Lauder, 2012). These patches may also reduce drag, speed water movement out the rather narrow space between the Lynceus hinge and trunk (Fig. 6C, F) and strengthen the incoming current. It must be emphasized that the above comments relate to the readily observed animals lying on one valve on the substrate. Water movement should be different in a swimming animal where the valves are more open and separated from the trunk.

Fryer and Boxshall (2009) discuss feeding in Lynceus, with observations on living L. gracilicornis and detailed discussion and illustration of leg structure and function in L. simiaefacies. They conclude that Lynceus does not possess filters and does not filter feed. However, particles as small as |$20\;\mu {\rm{m}}$| in diameter are retained, and the water current passes through various setal arrangements that might intercept them. Foremost are the long posterior setae (screen setae of Fryer and Boxshall, 2009) on the E2, E3 and portions of the E4 of the anterior legs that characterize the medial surface of the trunk (Figs. 4A, 5A, B). Although they are stated to be unlike typical crustacean filter setae (Fryer and Boxshall, 2009), they do have lateral setules with rear-facing tips that contain attached particles. The water current seems to chiefly flow into a medial “filter chamber” in the middle third of the trunk (Figs. 1A, 2A). Based on leg anatomy and the manipulation of preserved material, the following sequence of events could occur. The pairs of anterior scoop-like legs (Fig. 5A) T1-6 are separated by a substantial food groove. As they move backwards during their metachronal beat, water resistance causes them to flatten somewhat, so that the dorsal lobe of the exopod presses laterally against the valve, and the large, flap-like E2 and E3 (Figs. 4A, 5A) are forced inward. The long, flexible, posterior setae on E2 and E3 of opposing leg pairs now overlap to form a net. The net drives water posteriorly and dorsally, but is also pulled through it. The posterior setae are supported by each other and, at their base, by the row of stiff, anterior spines lying behind them (Fig. 5H).

On the anterior recovery stroke, water resistance causes the scoops to reform. The dorsal lobes are forced away from the valves, and the E2 and E3 away from those of the opposite leg. Their flexible posterior setae no longer intermesh or are supported by the anterior spines and now bend backwards. As the leg pair moves away from the following one during the metachronal beat, their anterior setae are dragged through, and scraped by, the anterior toothed spines on the E2 and E3 behind it. Since these spines are much shorter than the posterior setae, material is now concentrated and more easily transferred to the food groove.

Observations supported the suggestion of Martin (1989) that the ventral lobes of the exopods of the anterior legs are filters. Lastly, when the legs are beating gently, the antennae lie back within, or slightly extend from, the valves. Their plumose setae (Fig. 3B) may be passive filters that intercept entering material and are scraped by leg movements. The copepod nauplii that were seen impaled on the medial spines on the dorsal lobes of an exopod may have been incidental or part of a feeding mechanism. Obviously, a great deal remains to be learned about the details of feeding and food processing in L. brachyurus.

Three types of particle agglomeration leave an animal posteriorly a) compressed feces that have passed through the gut, b) pieces of substrate, sometimes combined with mucus, that have been held by the posterior legs without entering the food groove, probed, and discarded, and c) mucus-coated strings or particles of “pseudo-feces” that have been in the food groove. It might be argued that the pseudo-feces is being expelled because particle collection is an automatic consequence of respiratory leg movement and that gut processing capacity has been exceeded (Fig. 6A). However, concentrated particle suspensions were commonly brushed away, or moved out in the departing water current without being collected. In the animal’s “normal” position of lying on the substrate with the legs beating gently, most entering material (which must be mainly suspended particulate matter) is removed. Since it presumably could easily pass on through, it is therefore being actively collected and evaluated. Even when environmental particle concentrations were low, only some moved to the gut, while much seemingly identical material was expelled posteriorly as pseudofeces (Fig. 6E). There is therefore a continual collection and evaluation of presumably low quality material in the surrounding water, with only some being sent to the gut.

Reproduction

In the laboratory, a group of recently collected Lynceus generally shows some sexual activity. In a typical situation, a male approaches an ovigerous female lying on the substrate and initiates chemical and tactile interaction. The male repeatedly turns the female, probably aided by the atypical setae on the male T2 (Fig. 10D) which seem well suited to catch the narrow edge of the closed female valves. The male valves open and the antennae and anterior legs beat vigorously. This must transfer chemical products, including perhaps material from the brown coating seen on the tips of the endites of many males (Fig. 11C, E), between animals. In a successful mating the female valves open slightly, the male clamps a valve with claws and rostrum, opens his valves broadly, and bends the trunk downward with the legs continuing to beat. Eventually there is periodic insertion of the male trunk and extensive brushing of the female legs. Finally eggs and sperm are released and fertilization occurs within the female branchial cavity. There is “female choice,” since when females are separated from males until all existing eggs are shed, and then placed with a male, the whole mating process speeds up, and egg extrusion can occur within a few minutes of contact.

The release of free sperm by male Lynceus is consistent with the much larger number of sperm found in the testes of L. brachyurus than in those of spinicaudatan clam shrimp (Wingstrand, 1978), including Leptestheria dahalacensis, where a peculiar spermatophore is present (Scanabissi and Mondini, 2000). Grube (1853) described and figured sac-like objects adhering to setae of the various legs of both sexes of L. brachyurus, and for several convincing reasons rejected the possibility that they might be spermatophores and concluded that they must be epibionts. Solowiow (1951) figured attached sacs of various sizes that he identified as spermatophores. However, this is unlikely, since the |$22\;\mu {\rm{m}}$| length indicated for a presumed sperm is much greater than the |$4 - 5\;\mu {\rm{m}}$| diameter of Lynceus sperm (Wingstrand, 1978). In L. graciliformis attached pairs were typically seen swimming (Martin et al., 1986). Swimming pairs of L. brachyurus were seen in both lab and field, but were uncommon and seemed to be an attempt at escape by the female. Perhaps the female eventually settles and mating proceeds. Since eggs seem to be extruded into previously deposited sperm beneath the opercular lamellae, it seems impossible that fertilization could occur with the open female valves and water movement of a swimming animal.

Although females almost always carried eggs, the highly clumped egg release, and the four instances where a large number of recently released eggs were found with a non-ovigerous female, suggest that egg masses are shed together. Aeration of the newly fertilized eggs as they develop into the dormant cyst stage is clearly important. Except for small clusters, the eggs are in a flat, “cake-like” (Sars, 1986), disc only a few eggs deep. The departing water current passes out above and beneath the disc as it is held above the trunk and moved by the dorsal lobes of the exopods of T9 and T10. These form muscular, twisted cylinders that are mobile even when cut off (Grube, 1853). The disc is brushed from beneath by the long distal setae on the dorsal lobes of the exopods of the anterior legs (Fig. 2A). In some positions, the dorsal lamellae must redirect the departing water current back over the eggs at the back of the disc (Fig. 3A). The water current is continually flowing past the egg mass and it might be expected that there would be a continuous shedding of older eggs from the front of the disc as the material holding them weakened with age. This was not the case. Although some small egg releases were seen in the egg production study, |$92\%$| of the eggs were found in 17 releases of 88 or more eggs. Since male-female claspings, and therefore possible ovulations, occurred in several cases just prior to the appearance of a large egg mass, some of the released eggs may have been retained for no more than 12 hours. It may not be a coincidence that the large egg releases clustered around 160 eggs. Aeration would be expected to be less effective with larger egg numbers, so eggs may accumulate until a certain number is reached, which triggers a mass release of eggs of different ages. A female would increase reproductive success by maximally aerating fertilized eggs, and adding to the egg mass up to the point where the number of held eggs interferes with their successful aeration. Egg retention is limited, however; eight of the ovigerous females that were isolated upon collection had all shed their eggs two days later.

There is clear evidence for the accumulation of eggs from separate mating events, and in all probability of different paternities. Ovulation only occurs during mating, and, as also observed by Sigvardt and Olesen (2014), no single attachment lasts long enough to produce the large egg masses observed. In the present study, a field collected female contained 223 eggs. In the laboratory, contact between two individuals only lasted long enough for the production of a fraction of that number. Dobrynina (2011) found egg clutches of over a thousand eggs in L. brachyurus, but none approaching that size were found here. It is interesting that a male breaks off contact with a captured female only to reunite later to produce a large egg mass. The simplest possibility is that a female only holds a certain number of eggs ready for immediate fertilization. Males would maximize reproduction by abandoning a depleted female, and chancing an encounter with one having an immediately available egg supply. The male will return to a previous mate when mature ova are again present. There is evidence that supports this idea. In the egg production experiment, the only occurrences of a non-ovigerous female were in containers with recently released large egg masses, and of these, only one of the four was mating. In all eight cases of multiple releases of a large egg mass, there were intervening observations when the pair were separated. Perhaps coincidentally, a video of ovulation by a non-ovigerous female showed the extrusion of about 40 eggs, roughly comparable to the number of well-formed ova apparent in the posterior ovary. If there is a latency period in the formation of fertilizable eggs, it does not last long. Thirteen of the 17 females found with recently released large egg masses were already ovigerous, and the most productive pair released 159 eggs within 24 hours of a previous large release.

A female collected over half way through the reproductive season released 500 eggs during a four day period. If egg production should continue at that level for the four weeks that egg-bearing females are abundant (Table 3), then the lifetime fecundity of a female L. brachyurus could exceed 3000 eggs.

Life Cycle

Lynceus brachyurus has a lengthy period of development. In the years studied, the earliest larval stage was found in late February, the first juvenile in early April, and the first egg-bearing female in late April. For the 2009 data (Fig. 12, Table 3), there were 69 days between the first collection of larvae and that of an adult. In France, Rabet et al. (2005) found an even longer 90 days from the first larval stage until maturity. The clear progression in larval size over time and the rather brief period when larvae were found (Fig. 12), suggest a narrow period of egg hatching and larval recruitment. Still, plankton samples were small and larvae could easily have been missed.

The first bivalve stage is noteworthy. It does not seem possible that their long, flexible exopods with well-spaced setae (Fig. 9E) could generate the in-flowing water current that leads to food capture in adults. Further evidence of a different feeding style is that unlike either the larval, or later bivalve stages, the valves of these animals were covered with small adhering particles. Potential food material appears to stick to the animal or its setae and somehow be transferred to the food groove.

In 2008 and 2009, the first egg-bearing females were seen in early May, 42 and 43 days after the last collection of a larva (Fig. 12, Table 3). In the laboratory, Monakov and Dobrynina (1977) found seven post-larval stages lasting about a month before the appearance of the adult. The clear molt of maturity (Fig. 13, Week 20) indicates some developmental synchrony. While distinct size classes are not apparent later on (Fig. 13), there is some post-maturity molting, as egg-bearing females of distinct sizes were collected. The much larger samples of Retallack and Clifford (1980) showed a marked increase in carapace length in both sexes three weeks after the first appearance of ovigerous females. However, sequential egg releases do occur without an intervening molt. The predominance of males in three large June collections, noted to a lesser degree by Grube (1853), is interesting. Females can become totally quiescent, which should contribute to potential longevity, and survived longer than males in some laboratory experiments.

The presence of L. brachyurus in nine of the eleven years of this study contrasts markedly with their sporadic annual occurrence in the vernal pools investigated by Dexter and Kuehnle (1951) and Colburn (2004). The two years in which Lynceus was not found (2004, 2012) were preceded by the two wettest years in the period studied (2003, 2011), and the low numbers of 2006 by the third wettest year (Table 3). Presumably, the pool substrate remained too moist for the Lynceus eggs to receive the desiccation stimulus required for hatching in the following year. While by no means quantitative, there was a fairly standard netting path, so two other large differences between collections are probably meaningful. In 2006 and 2009, the North Pool contained markedly fewer animals than the South Pool. This could relate to the low January-February rainfall these years (Table 4) which may not have produced water levels that reached the cysts on the shallow edge areas that form a larger proportion of the North Pool than they do on the steeper-sided and larger South Pool. Lynceus was never found in the short-lived West Pool. This agrees with the observation of Schneider and Frost (1996) that L. brachyurus did not occur in a short duration (53 day) pool. Its absence is interesting though, since the West Pool fills at about the same time as the nearby North and South Pools, and must regularly receive Lynceus eggs from them by movement of deer and other animals. In the years 2005-9, the water level in the West Pool was much reduced by the end of May, but still held water past the dates when the other pools contained ovigerous Lynceus (Table 3), so an established population would seem possible.

Rising water temperature is clearly the chief cause of mortality. Probably associated with later lethal temperatures, there is a latitudinal increase in reported survival dates for L. brachyurus: 28 June at |$40^\circ {\rm{N}}$| in the present study, 21 July at |$47^\circ {\rm{N}}$| in Minnesota (Martin and Belk, 1988), 1 August at |$51^\circ {\rm{N}}$| in Alberta (Retallack and Clifford, 1980), and 24 September at |$65^\circ {\rm{N}}$| in Finland (Kaisila et al., 1963). The effect of factors other than temperature on L. brachyurus mortality is less clear. The slippery, flexible, and sub-spherical, carapace of post-larval Lynceus is not easily grasped with forceps and presumably would be equally difficult for predators. Although not obvious in my experiments, Schneider and Frost (1996) found effective insect predation on Lynceus and suggest that it limits its population size. However it does not seem possible that a sudden increase in predation would cause the observed sharp declines in numbers (Table 3).

Lynceus brachyurus has a long period of larval and juvenile development. However, development could start earlier as it does with the co-existing anostracan branchiopods. Thus an obvious question on the life cycle is what factors lead to the late appearance of the adults. When they arrive in early May, much of the pool’s seemingly suitable life span has past, potential predators are at their most numerous, and fatal warm temperatures are fast approaching.

Acknowledgements

The Stratford Ecological Center, Delaware, Ohio allowed me to study the vernal pools under their management. I am grateful to L. Tuhela for continuing facilitation of the electron microscopy and plate preparation, to C. Wolverton for aid with macrophotography and manuscript submission, to J. Gatz and D. Johnson for helpful suggestions on the manuscript, and to other colleagues for contributions to various aspects of the project, including A. Downing, R. Carreno, H. Grunkmeyer, D. Hamill, D. Johnson, J. Krehbiel, S. Waterhouse and M. Zavar. The electron micrographs were taken by N. Constantino, M. Fris, C. Hagen, K. Haines, C. Jagger, D. Ordosch and D. Reznik. The figures were assembled by M. Fris and K. Goughenour.