J. N. Am. Benthol. Soc., 1988, 7(2):87-99 O 1988 by The North American Benthological Society Feeding behavior of Rhithrogena pellucida (Ephemer0ptera:)

D. MCSHAFFREYAND W. P. MCCAFFERTY Department of Entomology, Purdue University, West Lafayette, lndiana 47907 USA

Abstract. Larvae of Rhithrogena pellucida (Ephemer0ptera:Heptageniidae) were observed feeding in observation flow cells using macroscopic video techniques. The stereotypic feeding behavior is described as cycles delineated by specific movements of the labial palps and consisting of stages of coordinated movement of the entire suite of mandibulate mouthparts. When employing the Labial Brushing Cycle, larvae used the labial palps to brush loosely accreted material from the substrate. When employing the Maxillary Scraping Cycle, larvae used special pectinate setae of the maxillary palps to remove material more tightly bound to the substrate. In both cycles, the legs of the larvae also removed material from the substrate. Larvae did not perform gathering cycles or feed on deposits of detritus. Food processing by successive fields of setae on various mouthparts to the point of ingestion is described in detail. Feeding observations combined with field and laboratory data on gut content and microhabitat distribution suggest that R. pellucida feeds primarily on periphyton. Comparisons are drawn between the primary scraping function of R. pellucida and the primary brushing function of Stenacron interpunctatum. A spectrum of adaptational strategies for obtaining food materials along a gradient of substrate association from seston to the most tightly attached materials is suggested and may be used to describe the feeding roles of many macroinvertebrates. Key words: Ephemeroptera, Heptageniidae, Rhithrogena pellucida, Stenacron interpunctatum, feeding behavior, morphology, functional feeding groups, ecology.

The feeding behavior and functional mor- sides the comparative data that may have rel- phology of the heptageniid Stenacron in- evance to the study of adaptation and evolution, ferpuncfatum, and new methods for studying there are also obvious ecological applications. them, were reported by McShaffrey and Mc- One that we have pursued involves the com- Cafferty (1986).To allow these data to be viewed mon practice of assigning functional feeding within a larger comparative context we subse- groups (FFG) (Cummins 1973) to the various quently studied another midwestern mayfly, members of aquatic communities. Rhithrogena pellucida. This mayfly provides an Strenger (1953) studied the functional mor- interesting comparison because it too is a stream- phology of European species of Rhithrogena and dwelling member of the family Heptageniidae included observations of behavior and inter- but lives in different habitats and possesses pretation of morphology. Strenger's work, al- somewhat different mouthpart morphology. As though important, was limited by her inability with S. interpunctatum, our objectives with R, to film behavior for detailed analysis and the pellucida were to document both the feeding fact that mouthpart morphology was not stud- behavior and feeding structures. We did not ied at the ultrastructural level. We are fortunate attempt to document the entire range of be- to have been able to extend such research by havior of the species, or to determine the nu- having SEM and videomacroscopic techniques tritional significance of the material ingested. available to us. Rhithrogena have pre- Our functional morphology/behavioral ap- viously been regarded as both collector-gath- proach to studying feeding incorporates field erers and scrapers (Cummins et al. 1984), with observations, gut content analysis, scanning the primary designation being collectors. electron microscopy (SEM), and videomacros- copy into a comprehensive research protocol. Methods This allows critical examination of questions concerning diet and microhabitat and the re- Organisms were collected in April 1986 from lated adaptations of feeding behavior and cobble substrate in the Tippecanoe River in mouthpart morphology. An important feature north-central Indiana. They were maintained of this system is its redundancy; all questions in the laboratory at room temperature (20-25°C) are investigated using multiple techniques. Be- in 75 x 30 x 30-cm aquaria filled with substrate 87 D. MCSHAFFREYAND W. P. MCCAFFERTY [Volume 7

TOP VlEW

FRONT VlEW 19881 FEEDING OF RHITHROGENAPELLUClDA 89 and water from the collection site; circulation This enclosure measured 1cm long, 2 mm wide, was provided by airstones. Only mature larvae, with a water depth up to 3 cm. A hemicylinder as judged by relative wingpad development in of black plastic (radius 1.5 cm), with 30 2.5-mm- Ephemeroptera larvae (McCafferty and Huff diameter holes, was placed around the enclo- 1978), were used for observation and videotap- sure on the inside of the tank. This plastic served ing. to isolate the organism visually and reduce un- Observations were carried out using the tech- wanted reflections during filming. niques outlined in McShaffrey and ~c~affert~Water flow was provided as in the shallow (1986), with some modifications. When study- flow cell; the water entered the tank cell through ing R. pellucida, all observations were made us- a tube attached with its outlet below the water- ing visible light supplied by a fiber optic illu- line inside the plastic hemicylinder directly be- minator to maximize resolution. Our experience hind the enclosure, and exited through another showed that S. interpunctatum feeding move- tube attached to one side of the tank cell. Water ments were the same in visible and infrared level in the cell was controlled by regulating light (McShaffrey and McCafferty 1986). inflow and outflow with valves. There was no Most observations were made in an obser- noticeable water current in the enclosure. The vation flow cell as described and illustrated in tank cell was mounted in place of the shallow McShaffrey and McCafferty (1986). With R. pel- flow cell on the stage of the observational the- Iucida, onl) the shallow (1 mm deep) flow cell ater. The final position of the organism, vertical was used. Water at room temperature (20-25OC) with the mouthparts facing the glass slide and was fed into the cell by gravity from an aquar- the videocamera, was the same as in the flow ium mounted on the roof of the observational cell. theater. Water leaving the cell collected in Food for the experiments was provided as another aquarium; a pump controlled automat- described in McShaffrey and McCafferty (1986). ically by a depth-sensing switch periodically Diatoms and other periphyton collected in the returned water to the upper tank. Water in both Tippecanoe River were cultured on glass slides aquaria was aerated. The flow of water was reg- in the laboratory. The composition of the pe- ulated by a valve; current speed was adjusted riphyton community on the slides was the same as described in the study of S, interpunctatum, as the previous study-a 95% covering of the with the same inherent difficulties in deter- diatoms Cocconeis and Achnanthes, numerous mining precise current speed (McShaffrey and bacteria between the diatoms, and the remain- McCafferty 1986). In most cases the flow was ing 5% consisting of scattered fungal hyphae, increased until the individuals began to orient Oscillatoria, Cladophora, Melosira, Fragilaria, Chlo- to the current, this usually occurred at low cur- rella, rotifers, and protozoa. Scanning electron rent velocities (1-5 cm/s). micrographs of typical growth on a slide surface In addition to the shallow flow cell, a tank are shown in Figure 2. Detritus taken from the cell was also employed. The tank cell (Fig. 1) holding tanks was introduced into the water was constructed of 2-mm-thick acrylic, 10.5 cm flow in the shallow flow cell at both high (>5 long, 4 cm wide, and 4 cm deep, and held a cm/s) and low (<5 cm/s) flow rates; detritus volume of 150 ml of water when in use. The was placed directly into the enclosure of the front face had a 1 x 3-cm rectangular section tank cell. Detritus settled to the bottom of the removed from the center; a portion of a stan- tank cell and the shallow flow cell at low flow dard glass microscope slide was attached over rates; at high flow rates detritus moved through the outside of this hole, and a 1-mm2mesh screen the shallow flow cell in suspension. We consid- was attached over the inside to form an enclo- ered feeding effective if ingestion of the ma- sure into which the individuals were placed. terial was observed.

FIG.1. Tank cell used in feeding experiments with R. pellucida, drawn to scale. Unlabelled arrows show direction of water flow. Abbreviations: BP = black plastic, EO = experimental organism, FP = front plate, MS = microscope slide (viewing port), SC = screen (1mmz),WI = water inlet, WL = water line, WO = water outlet, WR = water reservoir. 90 D. MCSHAFFREYAND W.P. MCCAFFERTY [Volume 7

FIG. 2. SEM micrographs of the surface of a glass microscope slide used in feeding experiments. a. Plan view, bar = 100 pm. b. Oblique view showing vertical distribution, bar = 10 pm.

Motion analyses were carried out by exami- tortion of mouthparts and is much faster than nation of the videotape recordings of the ob- critical point drying. servation sessions as in McShaffrey and McCaf- Gut contents of living larvae were placed on ferty (1986). Owing to sparseness of specimens glass microscope slides and examined imme- in the field, the need to sacrifice some individ- diately with a compound microscope. Gut con- uals upon return to the laboratory for gut con- tents were classified as being mineral, organic tent analysis, and natural mortality, only 10 in- detritus, diatoms, filamentous algae, unicellular dividuals were videotaped. Of the several green algae, remains, or bacteria. No hundred feeding cycles videotaped, 50 were attempt was made to quantify proportions other analyzed in detail using single frame advance. than to determine what type predominated. Pe- The 50 feeding cycles chosen for detailed anal- riphyton species present in the field were de- ysis were selected for visual clarity and pres- termined by returning substrate to the labora- ence of repeated feeding activity. Once a chro- tory, carefully removing the periphyton, and nology of mouthpart movement was determined examining it under a compound microscope. within these cycles, the chronology was tested against the remaining cycles. The feeding cycles are classified into stereotypic cycles as outlined Results in McShaffrey and McCafferty (1986)and adapt- Field observations ed from the methods used by Trueman (1968) and Keltner and McCafferty (1986). Each stage Our knowledge of the life cycle of R. pellucida is delineated by specific movements of the la- in the Tippecanoe River is based on extensive bial palps. collections from the same site from 1983 to 1987. Mouthpart structures were examined using Diligent searching of all microhabitats during light microscopy and SEM. Specimens were weekly intervals from April 1985 to July 1986 prepared for SEM by dehydration in increasing produced R. pellucida samples only during April concentrations of ethanol from 75% to loo%, and May. Most of these larvae were relatively then transferred to 100% ethyl acetate. Some mature and were always collected off the upper specimens were sonicated in 100% ethyl acetate surfaces of stones or collected in kick screen for one minute in an ultrasonic cleaner to re- samples. The larvae were also always found in move food material from the mouthparts; oth- mid-stream where the substrate receives direct ers were not sonicated so that the position of sunlight and the periphyton community is well material on the mouthparts might be deter- developed. mined. The larvae were then either dissected When R. pellucida larvae were found, the cob- or transferred whole to the SEM stubs after air ble substrate was covered by a diverse aufwuchs drying. This method produces very little dis- community, including diatoms such as Gom- / 1 head capsule paraglossa glossa of \ of lab~urn iabiurn piernenturn FIG. 3. Ventral view of mouthparts of Rhithrogena pellucida, mouthparts have been positioned to allow underlying structures to be seen. phonema, Navicula, Fragilaria, me lo sir^, Cocconeis, vae never attempted to feed on detritus or to Achnanthes, and Synedra. Cladophora, which was filter feed. also present throughout the period when R,pel- lucid~larvae were found, reached a growth peak Stereotypic feeding behavior in early May at the same time R. pellucida adults emerged. It appears that R. pellucida in the Tip- Resting position: When a larva is not feed- pecanoe River is univoltine, with emergence ing, the mouthparts (Figs. 3-5) are held in a and egg-laying in early May. Although we do characteristic resting position (Fig. 6a). The dis- not have data on the egg stage or the early tal segments of the labial palps are held folded instars, it is probable that the eggs hatch in inward so that they lie over (dorsal to) the para- winter, with mature larvae appearing on the aglossae but not the glossae. The labium as a upper surfaces of rocks in April. whole is held adducted dorsally to close the

Effects of experitnental conditions on behavior DORSAL Rhithrogena pellucida larvae were highly ste- head ca sule reotypic in their mouthpart movements. Exter- nal factors such as noise, vibration or change in light intensity often appeared to initiate ga18a!actma feeding movements; however, these move- ments were qualitatively similar to those ob- served under other conditions. Feeding cycles induced by external stimuli usually ceased after oaraglossa glsssa :rementr;r

1-2 cycles, whereas non-induced feeding cycles VENTRAL were usually repeated at least 5 times. S~eci- FIG. 4. Diagrammatic cross-section of the mouth- mens in the holding aquaria were positively parts of Rhithrogeiza pellucida, showing relative ventral rheotactic. During the course of this study, lar- positions. 92 D. MCSHAFFREYAND W. P. MCCAFFERTY [Volume 7

FIG.5. Ventral view of Rhithrogena pellucida mouthparts, ventral to dorsal sequence. a. Labium. b. Maxillae. c. Hypopharynx. d. Mandibles. e. Labrum. Scale (as shown in a) is constant for all figures.

preoral cavity ventrally. The galealaciniae of brushing tools; in the Maxillary Scraping Cycle, the maxillae are positioned with their median the maxillary palps are used as scraping tools. setae touching the lingua of the hypopharynx; Each cycle consists of several stages that can be the galealaciniae also are in contact ventrally delineated by specific movements of the la- with the labial palps and dorsally with both the bium. A single cycle of either type takes 1.2 s superlinguae of the hypopharynx and the man- at 20°C. Mouthpart movements described below dibles. The maxillary palps are held with the are often preceded by and concurrent with basal segments perpendicular to the long axis movement of the forelegs ahead (upstream) of of the body; the distal segments are folded for- the individual. Legs are alternately and repeat- ward at a 45' angle to the basal segments. The edly extended and then retracted, drawing the basal segments of the maxillary palps do not claws over the substrate. These leg movements extend beyond the contour of the head capsule; were capable of removing tightly adhering ma- the distal segments of the maxillary palps par- terial from the substrate and bringing it to the allel the contour, extending just slightly beyond mouthparts. it. The apices of the distal segments of the max- Labial Brushing Cycle: Sfage 1.-Abduction illary palps do not meet but extend only to the of the basal segments of the labial palps: As lateral edges of the labrum. The labrum and the feeding commences, the labium begins to swing maxillary palps together cover the anterior and ventrally towards the substrate. The basal seg- lateral aspects of the preoral cavity. The labrum ments of the labial palps swing posteriorly about is in contact with the denticles of the mandibles 20" (Fig. 6b); this movement draws the distal ventrally. The denticles of the mandibles are segments of the labial palps away from contact held separated, and presumably the molae are with the paraglossae. Once the apices of the closed. labial palps clear the paraglossae, the labium Feeding cycles: Larvae placed in the obser- completes its ventral movement and comes to vation cell with diatoms grown on the glass rest against the substrate. At this point the basal slide performed two different behavioral cycles; segments of the maxillary palps swing poste- we term these brushing and scraping cycles. riorly about 20". The galealaciniae do not move The two cycles differ in the degree to which in Stage 1 of the initial brushing cycle in the the maxillary palps are employed. In the Labial series, but in all subsequent cycles of the series Brushing Cycle, the labial palps are used as they move medially at this point in Stage 1.The FEEDING OF RHITHROGENAPELLUCIDA

FIG.6. Relative positions (all in ventral view) of Rhithrogena pellucida mouthparts during various stages of a Labial Brushing Cycle. a. Resting Position: Labium and Maxilla, left; Labial and Maxillary Palps, right. b. Stage 1 of first cycle: Labium and Maxilla, left; Labial and Maxillary Palps, right. c. Stage 2 of first cycle: Labium and Maxilla, left; Labial and Maxillary Palps, right. d. Intermediate position of Stage 3 of first cycle: Labium and Maxilla, left; Labial and Maxillary Palps, right. e. Final position of Stage 3 of first cycle: Labium and Maxilla, left; Labial and Maxillary Palps, right. f. Final position of Stage 3 of first cycle: Maxilla and Hypopharynx, left; Maxilla and Mandible, right. g. Stage 1 of second and all subsequent cycles: Labium and Maxilla, left; Labial and Maxillary Palps, right. labrum begins to move down slightly to shield (Fig. 6c). When the labial palps are fully ex- the front of the other mouthparts. tended they extend just beyond the contour of Stage 2.-Abduction of the distal segments of the head capsule. At this point the galealaciniae the labial palps: Once the basal segments of the complete their medial movement and are now labial palps reach their posterior position, the abducted, moving laterally outward. The distal distal segments of the labial palps are abducted, segments of the maxillary palps are extended moving through a range of about 50" until they at the same time as the labial palps, moving are at an angle of 125" to the basal segments through an angle of about 60" to make an angle 94 D. MCSHAFFREYAND W. P. MCCAFFERTY [Volume 7

of 105" with the basal segments. The maxillary and brush across the labial palps as they move palps reach well beyond the lateral borders of outward, and the posterior movement of the the head capsule. The labrum completes its basal segments of the maxillary palps draws the movements ventrally and helps shield the setae on the distal segments of the maxillary mouthparts under the head capsule from the palps through the combs on the denticles of the current. mandibles. If leg movements accompany the Stage 3.-Retraction of the labial palps: The brushing cycle, then the left and right forelegs labial palps are retracted from their extended alternate in moving to the mouthparts. position by a combination of an anterior swing Maxillary Scraping Cycle: This cycle differs of the basal segments of the labial palps and from a Labial Brushing Cycle in that the labial adduction of the distal segments of the palps. palps do not initiate all the movements, and the This movement is continuous and coordinated maxillary palps are brought to the substrate to so that the labial palps always remain under the scrape. Stage 1 is similar to Stage 1 of a Labial head capsule. The maxillary palps follow this Brushing Cycle. In Stage 2, the maxillary palps movement closely; the labial palps lead the are extended and lowered to the substrate; the movement and remain positioned just anterior labial palps are positioned in the articulation to the articulations between the segments of the between the maxillary palp segments. The ven- maxillary palps (Fig. 6d). During this move- tral surfaces of the distal segments of the max- ment the distal segments of the labial palps illary palps are covered with unique pectinate brush (sweep) up material from the substrate setae (Fig. 7) which effectively remove tightly using the profuse setae on their ventral and adhering material from the substrate. In Stage anterior surfaces. The trailing maxillary palps 3, the maxillary palps initiate the closing move- are not in contact with the substrate but are ments, and both sets of palps move over the positioned to capture any particles dislodged substrate. The forelegs may also be active in the by the labial palps. If foreleg movement is con- Maxillary Scraping Cycle. current with feeding, the claw of one foreleg is brought to the corresponding labial palp as Food processing it begins to be adducted during this stage. When the distal segments of the labial palps Once food has been brushed or scraped up reach the paraglossae they are raised dorsally by either the labial palps or the maxillary palps, by a rotation of the basal segments of the labial respectively, it must be positioned for inges- palps so that they pass dorsal to the glossae and tion. This process is carried on continuously paraglossae. The distal segments of the labial during feeding and for several seconds after the palps reach a position that is more median than last material is procured from the substrate. in the resting position (Fig. 6e). The entire la- Transfer is accomplished by successive fields of bium is raised as soon as the distal segments of setae on the various mouthparts (Fig. 3) as de- the labial palps are dorsal to the glossae and scribed below. Food transport is similar for both paraglossae. The galealaciniae follow the labial the Labial and Maxillary Feeding Cycles. palps in moving medially; they also move some- After completion of the first cycle, and when what ventrally. The denticles of the mandibles the labial palps are extended again (Stage I), complete their lateral outward movement and material on the setae of the distal segments of the distal segments of the maxillary palps rotate the labial palps is removed in two ways. Out- and are inserted among these denticles (Fig. 6f). ward lateral movement (Stage 1) of the distal The labrum moves dorsally. segments of the labial palps causes material to As feeding continues, this brushing cycle is be combed from the ventral setae of the labial repeated in a series of these cycles. Subsequent palps by setae on the dorsal surfaces of the glos- cycles in the series are initiated by a slightly sae and paraglossae. Most of the material on the further adduction of the distal segments of the distal segments of the labial palps is carried by labial palps (Fig. 6g), followed by the abduction the lateral and dorsal setae. This material is re- of the basal segments of the labial palps as in moved by the comb setae on the crown of the Stage 1. The brushing cycle is repeated un- galealaciniae as the distal segments of the labial changed except that in Stage 1 of subsequent palps are adducted further and then moved lat- cycles the galealaciniae are moving medially erally outward by the posterior movement of 19881 FEEDING OF RHITHROGENAPELLUCIDA 95

FIG. 7. SEM micrograph of the setae on the ventral EIG. 8. SEM micrograph of the mandibular molae surface of the maxillary palps of Rhithrogena pellucida, and epipharynx of Rhithrogena pellucida. The molae bar = 10 pm. are at the bottom of the micrograph, the epipharynx is at the top. The pharynx can be seen just above the diatoms, which have passed through the molae. Bar = 10 pm. the basal segments of the labial palps (Stage 1). The comb setae of the galealaciniae are brought into contact with the distal segments of the la- terial to a point where it can be pressed against bial palps by the slight adduction of the distal the mandibular molae for straining and inges- segments of the labial palps at the beginning tion (Fig. 8). Presumably, material left on the of the outward movement and with a ventral denticles of the mandibles is brushed directly movement of the galealaciniae. Material in the onto the dorsum of the hypopharynx or the comb setae is removed by stout setae on the venter of the epipharynx by the movements of dorsal sides of the distal segments of the labial the mandibles; we have not been able to observe palps when the comb setae are drawn base-first this directly since this area is obscured by the through them as the distal segments of the la- other mouthparts. bial palps are adducted medially (Stage 3). Ma- In both types of cycles the maxillary palps terial on the comb setae may also be removed are used to displace larger particles from the when the galealaciniae move laterally outward preoral cavity and to brush material from the (Stage 2); this draws the comb setae base-first comb setae on the galealaciniae. Once food ac- through the setae on the dorsal sides of the quisition ceases, the mouthparts remain in mo- glossae and paraglossae. In both types of cycles, tion as described above for several cycles, mov- any material on the maxillary palps is removed ing material into position to be ingested. During by combs on the mandibular denticles as the this phase, the labial palps make only small palps are withdrawn (Stage 2). movements, primarily slight adductions of the Material left on the glossae and paraglossae distal segments, and the maxillary palps only and on the dorsal surfaces of the distal segments brush against the mandibular denticles and the of the labial palps is moved further through a galealaciniae (they do not sweep the substrate). number of processes. As the labial palps return to the preoral cavity (Stage 3), this material is Gut content analysis displaced posteriorly, medially, and dorsally by the labial palps and the new load of material Gut contents of four larvae collected in the they bear. As the galealaciniae move medially field were analyzed along with those of one (Stage 1 and Stage 2), their median setae push larva that had been videotaped. All five guts the material between the superlinguae and the were similar in content; they contained dia- lingua of the hypopharynx (Fig. 6f). The lingua toms, organic detritus and sediment. The pre- beam dorsal rows of setae which guide the ma- dominant material was intact diatom frustules; 96 D. MCSHAFFREYAND W. P. MCCAFFERTY [Volume 7 there was also considerable organic detritus ap- observe these movements but inferred them parently of algal origin. Diatoms found in the based on her extensive morphological studies. gut included: Navicula, Gomphonema, Melosira, We were able to observe some opening and clos- Cocconeis, Achnanthes, Fragilaria, and Rhoico- ing of the setae of the distal segments of the sphenia curvata; Navicula species were the most labial palps, but we cannot eliminate the pos- common. None of the guts contained intact fil- sibility that such movements are passive re- amentous algae or recognizable animal re- sponses to drag as the palps are moved against mains. the substrate, the water, or the other mouth- parts. The molar region of the mandibles is com- pletely obscured by the other mouthparts, and Discussion we were not able to observe this area directly. We had to estimate molar action based on the Our observational procedures required the movement of the denticles, which were visible. organisms to be confined inside a small cell Because R. pellucida and S. interpunctatum are made of synthetic materials, and it exposed them both flatheaded mayflies of the subfamily Hep- to unnatural light and water flow regimes. In tageniinae, and can be found coexisting in the our study of S, interpunctatum (McShaffrey and same general area of streams, it is interesting ~c~affert~1986), we found that our experi- to compare their morphology, microhabitat, and mental conditions did affect the overall behav- feeding movements. Stenacron interpunctatum is ior of that species, but did not affect the actual found in slow-current areas, on the bottom of feeding movements reported. With S, inter- stones, away from the current (Flowers and Hil- punctatum, we were able to study individuals senhoff 1978, McShaffrey and McCafferty 1986, not confined to an observation cell; and we made Wiley and Kohler 1980, Wodsedalek 1912). We observations using infrared light to simulate observed R. pellucida living on the tops of stones the low-light nature of the crevices in which in high-current areas; Flowers and Hilsenhoff we collected the larvae in the field (McShaffrey (1978) also found this species in high-current and McCafferty 1986). We were not able to col- microhabitats. Strenger (1953) reported that the lect enough R.. pellucida larvae to permit obser- European species she studied were also collect- vations outside the observation cell, but our ex- ed on the tops of stones. The difference in mi- perience with S. interpunctatum, Ephemerella crohabitat is further reflected in the gut con- needhami, and various other Ephemeroptera, tents of these species. Whereas S. interpunctatum Trichoptera, Diptera, Odonata, Coleoptera, Am- is primarily a detritivore (Lamp and Britt 1981, phipoda, and other freshwater macroinverte- McShaffrey and McCafferty 1986), the R, pel- brates (unpublished data) suggests that the cy- lucid~studied here are primarily periphyton clic movements of the mouthparts are feeders. stereotypic and vary little between conspecifics The short developmental period for R. pel- of the same age class. This conclusion is con- lucid~in the spring (see also Clifford 1982, Flow- sistent with data from other feeding stud- ers and Hilsenhoff 1978) is associated with its ies (Brown 1960, Chance 1970, Devitt and Smith feeding habits. Rhithrogena pellucida in the Tip- 1985, Froehlich 1964, Pucat 1965). pecanoe River is only found in early spring Our observations are similar to those made before Cladophora growth is significant, whereas by Strenger (1953) for a European species of S. interpunctatum occurs year-round at the same Rhithro~ena.We were not able to confirm her site. In this river, a narrow temporal window suggestion that the distal segments of the labial in which there is good periphyton develop- palps deform their surfaces; this reportedly ment on the rock substrate occurs during the would enable their setal patches to open and latter part of April to early May. The end of this close and thus increase feeding efficiency period is marked by Cladophora overgrowth of (Strenger 1953). We also were not able to doc- the substrate; presumably the filaments of the ument the function of the brush on the mola Cladophora interfere with the feeding of R. pel- of the right mandible; Strenger (1953) stated lucid~,which requires a relatively flat substrate that it is used to carry food from the hypo- free of obstructions to the movements of the pharynx to the left mola. In both cases, it is maxillary palps. Stenacron interpunctatum feeds possible that Strenger (1953) did not actually on detritus, which is always available, and has a complex life history with larvae present year- ployed. Similarly, although none of the R. pel- round (McCafferty and Huff 1978). lucid~were observed filter-feeding, the large, Morphological differences between the well-muscled maxillary palps, with their fringe mouthparts of R. pellucida and those of S. inter- of bipectinate setae (similar to those present on punctatum correspond to the different micro- the small maxillary palps of S. interpunctatum) habitats. In R, pellucida they are apparently ad- appear adequate for filtering. At this time we aptations allowing the removal of tightly-bound cannot discount the possibility of R, pellucida material on the upper surfaces of rocks. The filter-feeding or deposit-feeding (gathering). We primary difference is the increased develop- had been unable to confirm a scraping function ment of the maxillary palps of R. pellucida (com- for the forelegs of S. interpunctaturn (McShaffrey pare Fig. 5b with figs. 2 and 4b in McShaffrey and McCafferty 1986); however, we were able and McCafferty [1986]). Corresponding to in- to document such a capabiltiy for R, pellucida in creased development of the maxillary palps is the present study. a flattening and broadening of the mandibular With regard to FFG classification, R. pellucida denticles (compare Fig. 5d with figs. 2, 3, and is best defined as a scraper based on the ability 4d in McShaffrey and McCafferty [1986]) to ac- of the maxillary palps and legs to remove tight- commodate the larger maxillary palps involved ly adhering material from the substrate, and the in removal of food particles. Other secondary presence of significant amounts of periphyton morphological differences between R. pellucida in the gut contents. It also, however, brushes, and S, interpunctaturn reflect a decrease in im- particularly when using the Labial Brushing portance of the labial palps to R. pellucida as Cycle. food-gathering organs. Rhithrogena pellucida has The ability of an organism to use more than fewer comb setae on the galealaciniae; and the one mode of feeding-whether it be R, pellucida hypopharynx, which is important in transfer- scraping and brushing, S. interpunctaturn using ring material from the labium to the molae, is the same mouthparts to brush, collect-gather, greatly reduced (compare Fig. 5c with fig. 4c in and filter, or different developmental stages of McShaffrey and McCafferty [1986]). Strenger a single species using different feeding methods (1953) noted similar differences when compar- (Cummins 1973, Cummins and Merrit 1984)- ing Rhithrogena and Ecdyonurus in Europe. complicates the problem of pigeonholing species The brushing cycle of S. interpunctatum is very into FFGs. This complexity is aggravated by at- similar to the Labial Brushing Cycle of R. pel- tempting to assign FFGs based on gut contents lucid~,the major difference being an improved that are often unidentifiable. Our study of R. coordination of the movements of the maxillary pellucida and other aquatic suggests that and labial palps in R. pellucida. This improved a more mechanical approach to delineating FFGs coordination places the maxillary palps in a po- for benthic insects that feed on relatively small sition closely following the labial palps to cap- materials (microvores, sensu McCafferty 1981) ture any material swept up by the labial palps may be appropriate. In this scheme, potential but not captured by them. Such improved ef- food material, regardless of origin, is viewed in ficiency may be vital in a higher-current mi- terms of a continuum ranging from material croenvironment. The Maxillary Scraping Cycle suspended in the water, to material settled on of R. pellucida is much more effective than either the substrate, to material bound or growing at- the Labial Brushing Cycle of R. pellucida or tached to the substrate. Various feeding strat- Brushing Cycle of S, interpunctatum in removing egies are adapted to deal with various sections tightly adhering material from the substrate of this continuum. owing to the use of the large maxillary palps In our scheme, benthic microvores can be di- with their unique setae. vided into two basic groups, Filterers and Col- We did not observe R. pellucida performing lectors. Filterers derive food material from the cycles similar to the filtering or gathering cycles water, and the filter may consist of either parts we found in S. interpunctatum (McShaffrey and of the body or manufactured devices such as McCafferty 1986). Although none of the R. pel- silk nets. Passive filterers rely on seston already lucid~in this study would feed on deposits of moving to their filtering apparatus, whereas ac- detritus, such feeding probably would be effec- tive filterers generally resuspend deposits to fil- tive when the Labial Brushing Cycle is em- ter. Thus, depending on the immediate source 98 D. MCSHAFFREYAND W. P. MCCAFFERTY [Volume 7 of the food material, we classify benthic filterers chose to use the term brushers to describe a as either seston filterers or deposit filterers. Col- feeding system found in certain Culicidae. Those lectors remove deposits or attached material authors were evidently unaware of the brusher from a substrate by direct contact of the feeding concept of McShaffrey and McCafferty (1986); structures with the food material. Collectors can from their discussion it is not clear whether be divided into three groups: brushers, gatherers, their mosquito "brushers" feed by direct con- and scrapers. Brushers use setae to obtain loose tact of the mouthparts with deposits (as is the or lightly attached material and are often mor- case with S. interpunctatum),or if the mouthparts phologically, functionally, and behaviorally create a current which carries the material to close to deposit filterers. Gatherers feed on sim- the mouthparts (deposit filtering, herein), or ilar materials but primarily use structures other both. than setae for food-gathering. Scrapers have ad- This study of R. pellucida shows the utility of aptations that allow them to feed on tightly our approach in investigating questions about accreted material. the functional morphology of feeding and ad- To summarize, benthic filterers and collectors aptations to differing environments. Compari- exhibit a range of feeding strategies for feeding sons between this research and our work on S. on a spectrum of small particles ranging from interpunctatum illustrate how these investigative suspended materials (seston filterers) to depos- techniques, when applied to coexisting species, its of various integrity (deposit filterers, brush- increase understanding of problems related to ers, gatherers) to firmly attached materials resource partitioning. (scrapers). As is the case for R, pellucida and S. interpunctatum, species may not be limited to one strategy. This scheme is attractive because Acknowledgements habitats can be characterized hydraulically, and Arwin Provonsha provided Figures 3 and 4; the FFG composition of the microvore com- Annn Delleur provided Figure 1. Dan Blood- munity can be estimated based on the physical good assisted with the construction of the ob- state in which the hydraulic forces will place servation flow cell. The scanning electron mi- the food material. In this context, concepts of croscope was made available by the Electron community development can be based on the Microscopy Center in Agriculture at Purdue relative amounts of microhabitat available to University with support from NSF grant PCM- provide food material in different positions in 8400133. the environment and with different propensi- Purdue University Agricultural Experiment ties for attachment. One drawback of such a Station Journal No. 11,359. classification system is that it provides little in- formation about the origin of the material (i.e., primary production, detritus, etc.); however, Literature Cited under natural conditions many benthic mac- BROWN,D. S. 1960. The ingestion and digestion of roinvertebrate microvores feed on a variety of algae by Cloeon dipteru~nL. (Ephemeroptera).Hy- materials and only a relatively few species are drobiologia 1631-96. dependent on a single trophic level for their CHANCE,M. M. 1970. The functional morphology food resources. As indicated above, the system of the mouthparts of black fly larvae (Diptera: excludes macrovores such as shredders, miners, Simuliidae). Quaestiones Entomologicae 6:245- engulfers, and many predators. 284. Erecting a new FFG classification system is CLIFFORD,H. F. 1982. Life cycles of mayflies (Ephem- perhaps premature; however, such a system may eroptera), with special reference to voltinism. become necessary as the behavior of more Quaestiones Entomologicae 18:15-90. species of aquatic insects is studied in detail. CUMMINS,K. W. 1973. Trophic relations of aquatic insects. Annual Review of Entomology 18:183- For instance, Dahl et al. (1988) recently pre- 206. sented a classification scheme of FFGs for the CUMMINS,K. W., G. F. EDMUNDS,AND R. W. MERRITT. Culicidae that also further divided the FFGs of 1984. Summary of ecological and distributional Cummins and Klug (1979) into more discrete data for Ephemeroptera (mayflies). Table 10A, groupings based on the mechanism of food ac- pages 122-125 in R. W. Merritt and K. W. Cum- quisition. It is notable that Dahl et al. (1988) mins (editors). An introduction to the aquatic in- 19881 FEEDING OF RHITHROGENAPELLUCIDA 99

sects of North America. Kendall-Hunt Publish- MCCAFFERTY,W. P., AND B. L. HUFF. 1978. The life ing Company, Dubuque, Iowa. cycle of the mayfly Stenacron interpunctatum CUMMINS,K. W., AND M. J. KLUG. 1979. Feeding (Ephemer0ptera:Heptageniidae). Great Lakes ecology of stream invertebrates. Annual Review Entomologist 11:209-216. of Ecology and Systematics 10:147-172. MCSHAFFREY,D., AND W. P. MCCAFFERTY.1986. Feed- CUMMINS,K. W., AND R. W. MERRITT.1984. Ecology ing behavior of Stenacron interpunctatum (Ephem- and distribution of aquatic insects. Pages 59-65 er0ptera:Heptageniidae). Journal of the North in R. W. Merritt and K. W. Cummins (editors). American Benthological Society 5:200-210. An introduction to the aquatic insects of North PUCAT,A. M. 1965. The functional morphology of America. Kendall-Hunt Publishing Company, the mouthparts of some mosquito larvae. Quaes- Dubuque, Iowa. tiones Entomologicae 1:41-86. DAHL,C., L. E. WIDAHL,AND C. NILSSON.1988. Func- STRENGER,A. 1953. Zur Kopfmorphologie der tional analysis of the suspension feeding system Ephemeridenlarven. I. Teil Ecdyonurus und Rhith- in mosquitoes (Diptera:Culicidae). Annals of the rogena. ~sterreichischeZoologische Zeitschrift 4: Entomological Society of America 81:105-127. 191-228. DEVITT,B. D., AND J. J. B. SMITH. 1985. Action of TRUEMAN,E. R. 1968. A comparative account of the mouthparts during feeding in the dark-sided cut- burrowing process of species of Mactra and of worm, Euxoa messoria (Lepidoptera:Noctuidae). other bivalves. Proceedings of the Malacological Canadian Entomologist 117:343-349. Society of London 38:139-151. FLOWERS,R. W., AND W. L. HILSENHOFF.1978. Life WILEY,M. J., AND S. L. KOHLER.1980. Positioning cycles and habitats of Wisconsin Heptageniidae changes of mayfly nymphs due to behavioral reg- (Ephemeroptera). Hydrobiologia 60:159-171. ulation of oxygen consumption. Canadian Jour- FROEHLICH,C. G. 1964. The feeding apparatus of the nal of Zoology 58:618-622. nymph of Arthroplea cogener Bengtsson (Ephem- WODSEDALEK,J. E. 1912. Natural history and general eroptera). Opuscula Entomologica 29:188-208. behavior of the Ephemeridae nymphs Heptagenia KELTNER,J., AND M7. P. MCCAFFERTY.1986. Function- interpunctata. Annals of the Entomological Soci- al morphology of burrowing in the mayflies Hex- ety of America 5:31-40. ageilia limbata and Peiltagenia vittigera. Zoological Journal of the Linnaean Society 87:139-162. Received: 6 November 1987 LAMP,W. O., AND N. W. BRITT. 1981. Resource par- Accepted: 5 May 1988 titioning by two species of stream mayflies (Ephemer0ptera:Heptageniidae). Great Lakes Entomologist 14:151-157.