BEHAVIOR, FUNCTIONAL MORPHOLOGY, AND ECOLOGY RELATED TO
FEEDING IN AQUATIC INSECTS WITH PARTICULAR REFERENCE TO
STENACRON INTERPUNCTATUM, RHITHROGENA PELLUCIDA
(EPHEMEROPTERA: HEPTAGENIIDAE), AND
EPHEMERELLA NEEDHAMI (EPHEMEROPTERA: EPHEMERELLIDAE)
A Thesis Submitted to the Faculty of
Purdue University
by
In Partial Fulfillment of the Requirements for the Degree
of
Doctor of Philosophy
August 1988
CHAPTER III - FEEDING BEHAVIOR OF STENACRON INTERPUNCTATUM
CHAPTER IV - FEEDING BEHAVIOR OF RHITHROGENA PELLUCIDA
CHAPTER VI - FEEDING BEHAVIOR AND RELATED FUNCTIONAL MORPHOLOGY OF EPHEMERELLA NEEDHAMI
CHAPTER VII - ECOLOGICAL ASSOCIATION OF EPHEMERELLA NEEDHAMI AND THE GREEN ALGA CLADOPHORA
CHAPTER VIII - MISCELLANEOUS FEEDING OBSERVATIONS
CHAPTER IX - CONCLUSIONS AND HORIZONS
It didn't really take me long to convert my dissertation from Wordstar version 1.ancient_history to HTML. However, as I look at the translation, I am bothered by a few things. Most important, there seem to be a number of errors in the literature citations, so please use these with care. I have noticed some inconsistencies in either the date of publication or in the spelling of names that I will attempt to correct in later versions. Also, it will take some time to get the figures scanned in. Finally, I noticed many other errors that have gotten past me and some good editors. Students - proofread, proofread, proofread, then set it aside for a few years and proofread again.
I have made a conscious decision, for now at least, to leave the manuscript pretty much as it was. Things cited as being in press that have since been published will be updated when I get a chance, but I will not revise my thoughts from the original. I may, however, add links to additional pages in the future where hindsight and other research has led me to reconsider.
McShaffrey, Dave. Ph.D., Purdue University, August 1988. Behavior, functional morphology, and ecology related to feeding in aquatic insects, with particular reference to Stenacron interpunctatum, Rhithrogena pellucida (Ephemeroptera: Heptageniidae), and Ephemerella needhami (Ephemeroptera: Ephemerellidae). Major Professor: W.P. McCafferty.
KEYWORDS: Videomacroscopy, scanning electron microscopy, gut content analysis, and other techniques were used to investigate the diet, feeding behavior, and functional morphology of three species of mayflies: Stenacron interpunctatum, Rhithrogena pellucida (Ephemeroptera: Heptageniidae), and Ephemerella needhami (Ephemeroptera: Ephemerellidae).
Stenacron interpunctatum was found to feed on detritus, which it collected using three distinct stereotypic feeding cycles; a brushing cycle was used to feed on loosely attached detritus, a gathering cycle was used to feed on unattached detritus, and a filtering cycle was used to feed on suspended material. Stenacron interpunctatum was classified as a collector-gatherer.
Rhithrogena pellucida used two feeding cycles to ingest a diet consisting mostly of periphyton. One cycle utilized the labial palps to brush up loosely attached material, the other cycle utilized the maxillary palps to scrape up tightly accreted material. Rhithrogena pellucida was classified as a scraper.
Ephemerella needhami used two feeding cycles to feed on filaments of Cladophora and associated material. One cycle used the maxillae to remove detritus from the filaments, the other used the mandibles to bite off filaments. Ephemerella needhami was classified as a collector-gatherer.
Comparative studies of the mouthpart morphology of all three species at the ultrastructural level revealed morphological differences closely associated with microhabitat and diet.
Brief accounts of feeding behavior are listed for gastropods, amphipods, other mayflies including: Callibaetis, Isonychia, Potamanthus, Heptagenia, Stenonema, Ephoron, and Caenidae; the caddisfly Macrostemmum, water pennies, larval blackflies, and the midge Cricotopus. The application of the techniques developed in this study to such diverse organisms proves the utility of these techniques for future studies.
This dissertation concerns natural history, a scientific style at the heart of the scientific method. Here observational methods are emphasized and pushed to their limits by state-of-the-art technology.
The central question addressed here is simple: How do organisms feed? That topic in itself is rather broad, and here it is applied to three organisms: Stenacron interpunctatum, Rhithrogena pellucida, and Ephemerella needhami. These organisms play an important, but as yet poorly understood, role in stream ecology. In addition, understanding how these organisms feed may shed light on evolutionary relationships. The last two points deserve elaboration.
An important idea surfaced in 1980 in the form of the river continuum concept (RCC) (Vannote et al. 1980). Essentially, this was an attempt to describe the association between changing physical characteristics and differing biological communities along a river from headwaters to mouth. Rooted in earlier works, particularly that of Cummins (e.g. 1973, 1974), the RCC attempted to link a number of ecological concepts and observations.
A basic component of the RCC was the description of energy flow through the macroinvertebrate component of the ecosystem in terms of functional feeding groups (FFGs). These groupings attempted to reduce the manifold feeding adaptations of freshwater crustacea, mollusca, and insecta into a few basic groups. With these groupings in hand, ecologists could then formulate ecosystem-level theories. From an ecological perspective, two of the goals of this study are: 1) to expand our knowledge of the feeding ecology of a few organisms; 2) to provide techniques for a systematic approach to the feeding ecology of a wider range of macroinvertebrates. An overview of the techniques used in this study is presented in Chapter II, and accounts of the feeding ecology of the 3 organisms studied are given in Chapters III, IV, and VII.
Aside from the ecological ramifications, this study is important from an evolutionary and taxonomic standpoint. The feeding structures of aquatic macroinvertebrates are often used as identifying characters in taxonomy, or as a basis for comparison in systematics. An underlying knowledge of the function of these structures, why they take on a particular shape or configuration, will obviously enhance their use or disuse as taxonomic or phylogenetic characters. In particular, an ability to distinguish between homology and convergence is critical, this question is addressed in Chapter VI.
The three organisms studied were not chosen at random. Stenacron interpunctatum was chosen as the prototypic subject at a time when the study included 11 different species, all of which were selected for local availability. Because of its secretive ways and prognathous mouthparts, S. interpunctatum was viewed as the most challenging to film, hence the decision to study it first. Rhithrogena pellucida was selected next because it was related to S. interpunctatum, was found in similar habitat, but occupied a different microhabitat and had different mouthpart morphology. Together these species represent microhabitat extremes among the non-predaceous Heptageniinae, and provide a good basis for explaining some of the peculiarities of heptageniid mouthpart morphology. The comparative mouthpart ultra-morphology of these species is the subject of Chapter V.
Ephemerella needhami possesses more generalized, hypognathous mouthpart morphology, and was chosen to round out this study for comparative purposes. Ephemerella needhami turned out to be a fascinating subject because of its intimate association with the filamentous green alga Cladophora. The feeding behavior and morphology of E. needhami, and comparisons of E. needhami with other hypognathous mayflies and with heptageniids is covered in Chapter VI. The association with Cladophora is described in Chapter VII.
During the course of this study, a wide range of organisms were filmed in addition to the 3 species studied in detail. The casual observations of these other species may provide useful information to specialists in those groups, or act as a springboard for future research. A collection of anecdotal accounts appears in Chapter VIII to fill these needs.
Most of this dissertation has already been published or is currently in press. Chapter III was published as McShaffrey and McCafferty (1986), Chapter IV as McShaffrey and McCafferty (1988), and Chapters V, VI, and VII are in press as separate papers. Slight differences in style between the chapters reflect the different journal formats. All of the references and figures have been collected at the end of the paper. At this point, I would like to especially acknowledge the contributions of Pat McCafferty, who was truly a co-author on these chapters.
Since each of the main chapters was written to stand on its own, the traditional introduction, and methods sections are somewhat condensed, and pertinent literature reviews will be found in the chapters. Despite this, there is some redundancy in several areas, but overall it makes the dissertation much more readable by placing the appropriate background, methodology, and conclusions near the data concerned.
Because this dissertation has almost as much to do with techniques as with results, Chapter II has been written to guide the reader through the evolution of the observational protocol and associated equipment. It is historical in style; again, the rigorous description of the methods is contained in each chapter as appropriate. Some readers may wish to read the methods chapter last, after they have a feel for the methodology and equipment that were used.
Chapter IX is a brief discussion of the future applications of this study. It focuses on the techniques evolved during the course of the project, and how they might be used in future studies. It also looks at the types of information generated by these techniques and how they may be applied to different areas of science.
A number of different techniques were used in this study; together they constitute a much more holistic approach to the study of feeding ecology and functional morphology than other previous studies. Figure 1 depicts the various methods used and the areas of investigation that the various methodologies treat. The overall design of this project was to deal with the central question - How do these organisms feed? - by investigating their microhabitat, diet, functional morphology, and feeding behavior. These areas of investigation were studied by means of field observations, gut content analysis, scanning electron microscopy (SEM) and videomacroscopy, as appropriate. Each area of investigation was thus treated by multiple techniques; this helped to eliminate the bias inherent in any one technique. The exact methods used in the study of each species is included with the chapters dealing with that species; this chapter is designed only to provide on overview of the techniques used, as well as some historical background on their evolution over the lifespan of this project. For purpose of discussion, the methods are divided into field observations, videomacroscopy, behavioral analysis, scanning electron microscopy, gut content analysis, and synthesis.
Careful field collection and observation were a critical component of this project; much of the ecological data acquired was obtained in this way. Most of the organisms studied, including all of the S. interpunctatum, R. pellucida, and E. needhami, were collected at one site, the Tippecanoe River, at the point where it intersects state highway 18 at the Carroll County - White County border. This site was initially chosen for accessibility, but a number of advantages grew out of continued use. Repeated sampling and observation revealed ecological patterns, particularly seasonal patterns, which might have gone undetected otherwise. From May 1985 to November 1985 weekly presence/absence records of all species at the site were made, and the results entered into a computerized database. Additional sampling of this type, but on a more irregular basis, continued through May 1986. These records provided the baseline data from which patterns of seasonal activity were discerned. Such data led to the study of E. needhami, which had an interesting seasonal activity pattern closely linked to that of the alga Cladophora.
For all species, careful observations were made of microhabitat associations. Organisms were often collected individually, usually by picking up a piece of substrate. Location of the individual was noted, as well as its behavior; i.e. did the larva immediately move to the underside of the stone? Considerable time was spent closely examining the substrate with the aid of mask and snorkel to determine exactly where the larvae were located when undisturbed. All field observations were made in daylight.
Videomacroscopy is the use of electronic image gathering, processing, and recording devices (video) to observe, analyze, and record moving images at magnifications greater than 1x. It has its origins in cinematography, where moving images are captured on film. Videomacroscopy has existed almost as long as video cameras, but the recent availability of high-quality, low-cost consumer equipment has greatly enhanced its appeal. Videomacroscopy has several advantages over cinematography: 1) Costs are much lower for video; whereas the initial investment for a video or cine set-up may be similar, videotape is less expensive than film, and there are no processing costs. 2) Videotape can easily be copied; its electronic format makes editing, image processing, titling, etc. much simpler. 3) Since the video signal is monitored in real time, the exposure can be corrected immediately with little loss of data; with film, exposure problems are only detected when the film is processed. 4) The video signal is routed electronically through wires, thus the camera operator may be some distance away; this enhances observational flexibility. 5) Video image brightness can be amplified electronically, making observations at low light levels possible. 6) Video image size is also amplified electronically, thus reducing the amount of magnification that must be accomplished by optics.
Video does have drawbacks; its main disadvantages are poor resolution and blurring of single-frame images. Both are related to the scan lines that make up a video image. The number of scan lines determines resolution to a great extent, the more scan lines there are, and the smaller they are, the better the resolution.
The equipment used in this study was limited by the resolution of the videocassette recorder (VCR). Resolution in video work is measured in terms of the number of horizontal lines that can be distinguished on a test pattern. The camera used had a resolution of 800 lines, the monitors had resolution up to 700 lines, and the VCR had resolution of about 240 lines.
In terms of obtaining clear single-frame images, the main limitation is the scan speed. Each scan of the image area takes 1/60 of a second, but two scans are needed to make a complete picture. Thus, the effective speed corresponds to a 1/30 second shutter speed in 35mm photography. This in turn means that almost any image will be blurred to some extent. Unlike 35mm photography, where shutter speeds can be increased, the speed of the standard video camera is fixed.
There are improvements in sight. High-speed shutters are already available in cameras. Although they require more light, they are able to freeze motion by allowing only a brief exposure of The video equipment used in this study was assembled by J. Keltner for his study of burrowing in mayflies (Keltner 1983, Keltner and McCafferty 1986). During the course of this study, the equipment was modified and augmented for observing feeding behavior. These modifications came in four areas: electronics, optics, observational arenas, and documentation.
Few changes were made in this area. The WV-1850 video camera and its Extended Red Newvicon pickup tube (S4119) are extremely effective at low light levels (down to 0.1 foot-candles), and, as already noted, have relatively good resolution for video equipment. The basic configuration used by Keltner (1983), with the video signal routed through a Panasonic time-date generator (WV-810) to the VCR (Panasonic NV-8950) and monitors, was not changed. By 1987, a color monitor and a second VCR had been added to the basic configuration, but were not used in making observations. The time-date generator was modified so that its stopwatch functions were fully enabled and could be operated remotely.
The optical system used by Keltner consisted of a 50 mm Macro-Switar lens and 75 mm of extension tubes. The 35x magnification (on a 9" monitor) was not sufficient for observing mouthpart movement. Several different configurations were tested to increase magnification. A Fujinon 20-100mm remotely controlled zoom lens gave acceptable performance only when fitted with close-up lenses or extension tubes, both of which eliminated the remote zoom function. Ultimately, this lens, coupled with the low-light capability of the WV-1850 surveillance camera, found its best use in covert monitoring of such nocturnal laboratory events as students previewing examinations (Keltner, personal communication), or janitors surreptitiously using the telephone for personal calls (McShaffrey, unpublished data). Another lens, a modified Fujinon 50 mm lens, gave performance similar to the Macro-Switar, with the added convenience of remote control of the iris.
The main improvement to optics configuration in the system was the purchase of an adapter that permitted Nikon F mount equipment to be mated to the "c" mount of the videocamera. Magnifications up to 80x were achieved using the extension tubes, adapter, a bellows unit, and 50 mm Nikon lens. The bellows unit was adapted by adding a remotely controlled electric motor with elastic drive to the forward focusing knob, this made remote zooming possible.
The high magnification achieved by using this configuration reduced working distance and depth-of-field. At maximum extension, the lens was so close to the observation cell that it was very difficult to get light around the lens to the subject. At the same time, increased light was needed because lens apertures had to be reduced to increase depth-of-field. While adequate light could be obtained for visual-spectrum observations by a difficult arrangement of the fiber optics light pipes, infrared illumination was insufficient. A partial solution was achieved by attaching infrared light-emitting diodes (LED's) to the surface of the observation cell to provide spot illumination.
The Nikon to "C" adapter also made it possible to mount the videocamera on a trinocular-head dissecting microscope via a Nikon AFM photomicrography unit. This required that the shutter in the photomicrography unit be kept open by setting it on the "bulb" position. The stereomicroscope provided high-magnification views (up to 120x) with ample working distance, but focusing was difficult since the heavy camera had to be moved. In addition, depth-of-field and observational flexibility were severely limited. Still, it was used with great success in the study of E. needhami.
A final comment should be made concerning the effects various sources of light used in this study. Keltner (1983) used near infrared light almost exclusively. His organisms lived in low-light situations and he was concerned that high levels of human visual-spectrum light would affect their behavior. There is general agreement that most insects are not sensitive to infrared light, and in particular, aquatic insects would not be expected to sense infrared since it is absorbed quickly in water. Still, there have been few if any studies on the visual spectrum of aquatic insects; the assumptions about their visual spectra are based on extrapolations from other insects such as the honeybee Apis mellifera. The fiber optics lighting system used in this study provided infrared light by filtration, thus reducing the amount of light available. When the enhancements were made to the optical system to increase magnification, an increase in light was also necessary, and there was no longer sufficient infrared light available. Close examination of videotapes made with infrared and human visual spectrum light revealed little difference in the stereotypic feeding behavior of S. interpunctatum, although its overall behavior was certainly affected, since this species normally avoids light. Thus, while feeding activity was diminished (as the larvae sought to escape the light), the feeding that did occur was normal, and the resulting observations could be made at higher magnification and with greater clarity and depth-of-field. After the study of S. interpunctatum, the decision to use infrared or human visual spectrum lighting was made on a case-by-case basis; for instance, R. pellucida, which feeds on exposed rock surfaces in daylight, was filmed with human visual spectrum light, while Ephoron sp., a more nocturnal bottom dweller, was observed with infrared light.
Keltner (1983) stated that the key to detailed behavioral observations was confinement of the organism to a small area, and he developed a series of small observation cells (aquaria) for this purpose. These observation cells were not suitable for observations of S. interpunctatum feeding; there were two problems that had to be solved: a requirement for flowing water (to make conditions more natural) and the need to view feeding underneath the head capsule. These problems were solved by the development of the flow cell described in detail in the next chapter. The flow cell essentially simulated the crevice microhabitat of S. interpunctatum, and the microscope slide with cultured periphyton which made up the front wall/viewing port of the observation cell simulated the rock surface with its biofilm food source.
Culture of periphyton and other material onto the glass slides was problematic; slides used often were essentially a monoculture of Cocconeis sp. diatoms. This problem was turned to an advantage by the observation that Cocconeis is among the periphyton most resistant to removal by macroinvertebrates. Cocconeis removal by an organism under laboratory conditions became the test criterion for scraping function.
The flow cell, as used in the study of S. interpunctatum proved useful, with modifications, in studying R. pellucida, E. needhami, and other organisms. The cell was easily adapted for tests of filtering and gathering functions by adding detritus in suspension and adjusting water flow. Flow cells were constructed in various sizes and enclosure depths. Current flow cells have been refined by increasing the width and bolt spacing so that the camera lens can approach the viewing port more closely in high magnification/short working distance situations. The front surface surrounding the viewing port was beveled in later models to allow more light to reach the viewing port.
Despite improvements to the flow cells, they are not suitable at maximum magnifications where the short working distance demands a flush front surface with no protruding bolt heads. The tank cell described in detail in Chapter IV was developed for that purpose. Since the front surface could not be removed, it was only suitable for tests of gathering function, where the food material could be placed directly into the enclosure. The tank cell also had little flow or aeration. The tank cell was employed to some extent in the study of R. pellucida, but its major use was in studying lentic organisms.
Although it was much easier to film organisms when they were confined to small observation cells, it was necessary to make observations of organisms in more natural surroundings in order to gauge the effects of the enclosures on behavior. For S. interpunctatum, such observations were made by placing the videocamera, bellows unit, and lens on a tripod next to the holding aquaria. Similar observations were made of R. pellucida and E. needhami in an artificial stream described in Chapter VI.
This area was a constant problem. It is very difficult to get a publication-quality photograph from a video source, and presentation of moving video images in a published format is even more challenging. I have had to rely on written descriptions of feeding movements and line illustrations to document behavioral observations made using video. Other methods were attempted: plastic 3-D models, animated drawings, computer-generated drawings, and various techniques for reproducing video images on film. No method has approached the quality of well-drawn illustrations of each stage of feeding. A recent addition to the video equipment, a Polaroid Freeze-Frame unit, produces relatively good images from video, but images reproduced from videotape are not of sufficient quality for publication.
Behavioral analysis was carried out in 3 stages. The first was a careful planning of observational conditions, the second a detailed analysis of recorded behavior, and the third an evaluation of results. Often the third stage prompted additional observations, sometimes under different conditions, and the process was repeated. Observational design meant insuring that each species was placed in appropriate circumstances so that all possible feeding modes might be observed. Normal practice was to initially pair the organism with a food source that the organism had been reported to feed on in other studies. After such initial tests, other food sources were introduced. The importance of testing organisms with food sources other than those reported in the literature cannot be overemphasized. Tests of filtering function in S. interpunctatum showed that this was a possible but totally unsuspected method of feeding. A similar situation occurred with the observation of a predatory function in E. needhami.
In this study, only mature larvae were studied. This minimized problems associated with accurate species identification and filming of the smaller, less mature larvae. Unfortunately, since organisms often change their feeding habits as they develop, these data are only applicable to a small segment of the total life span of the species studied.
Once appropriate observations had been made on as many organisms as practical, detailed analysis began. Feeding sequences were chosen, on a basis of clarity and completeness, for each feeding function. Feeding sequences were divided into cycles based on the movement of a key mouthpart (i.e. the labium of S. interpunctatum and R. pellucida, or the maxillae of E. needhami.
Each cycle was analyzed separately by freeze-frame techniques. In this process, the positions of all mouthparts at the beginning of each cycle were marked on a clear acetate sheet laid over the monitor screen. The tape was advanced one frame (typically 0.03 second) and the new mouthpart positions were noted. Time data were recorded from the on-screen display (accurate to 0.01 second).
Once composite behavioral cycle descriptions for each type of feeding had been assembled, these descriptions were checked against remaining videotape that was of lesser quality. Particular care was taken to determine if the behavioral descriptions, which were often based on observations from the more restricted observational arenas (hence higher quality), matched observations from more natural conditions (aquaria, artificial streams), which were usually of lower quality.
With mayflies, it is vital to understand the form and function of the setae on the mouthparts if one is to understand the functional morphology of the mouthparts. Examination of the setae with a light microscope is not sufficient; SEM is necessary. Over 500 SEM micrographs were made on 3 different instruments during the course of this project. The following is a summary of the techniques in use at the end of the project.
Larvae were killed and stored in 70% ethanol. Periods of storage ranged up to 1 year; some specimens examined for comparative purposes were taken from the museum where they had been stored for up to 20 years. Larvae were dehydrated by placing them in 9ml vials with successive solutions of 80%, 90%, and 100% ethanol for 10 minutes at each step. They were then placed in 50% ethanol, 50% xylene overnight; the xylene increased hardness of the specimens. The ethanol/xylene solution was replaced by several successive 9 ml aliquots of ethyl acetate. If it was desirable to remove food material from the mouthparts, they were sonicated for 30 seconds in an ultrasonic cleaning bath after the first aliquot of ethyl acetate was added.
If dissection was required, this was performed with the mouthparts submerged in ethyl acetate. Mouthparts were transferred to SEM stubs coated with a thin layer of adhesive made by dissolving the adhesive from transparent tape in ethyl acetate, painting this material onto the stub, and allowing it to dry. Care was taken to insure that the mouthparts transferred to the stubs were dry before being set in place. This method of dehydrating specimens has the advantage of being much faster than critical-point drying, and it produces few artifacts.
Specimens were coated with gold-palladium for 5-8 minutes at 10 microvolts in a Hummer 1 evaporative coater; then the stubs were placed on their sides (approximately 70o) and coated for an additional 3-5 minutes.
Specimens were examined via secondary electron emission with a JSM 840 SEM, normally operating at an accelerating voltage of 15 kv. Micrographs were made using Polaroid Type 55 film. MIcrographs were made of whole larvae, head capsules, mouthparts, and all setal types. Dorsal, ventral, anterior, lateral, posterior, medial and other views were employed as needed. As a historical note, some of the initial SEM work in this project was done on a JSM 1 instrument at the University of Akron.
Gut content analysis, used alone, has been the chief method of determining both diet and FFG classification; the problems associated with this are discussed in Chapters III and IV. Still, gut content analysis is an important tool in determining diet, and, when coupled with behavioral observations, in determining FFG classification.
Because gut content analysis was a supporting method and not the primary means of investigation, qualitative methods were chosen initially. This in essence meant macerating the gut contents of individual larvae in euparol on a glass microscope slide, and examining the slide for content.
For the study of E. needhami, a more rigorous methodology was employed. Gut contents were analyzed quantitatively with the aid of an ocular grid and computerized data entry and summarization. In all cases, gut contents from individuals collected in the field were compared with those from organisms that had been videotaped to determine if laboratory conditions were affecting diet.
An adjunct to gut content analysis was the examination of substrates taken from the field to see what food items were available. Such examinations were conducted by light microscope and SEM. Similar examinations were made of substrates used in behavioral studies in the laboratory.
Once the data from the techniques described above had been gathered and analyzed, it had to be synthesized into a holistic overview of the feeding ecology and functional morphology of the organism in question. Much of the synthesis involved correlation of morphological data with behavioral observations to reach conclusions concerning functional morphology. This meant determining function based on what had been observed (rather than determining function based on pre-conceived notions; the reader should consult Keltner [1983] for an excellent treatment of functional morphology and associated sciences and their methodologies). Examples of situations where function did not follow preconceptions are given in Chapter V.
Other correlations included gut contents vs. material observed to be ingested, and microhabitat data vs. food availability vs. gut contents. Discrepancies were noted and the relevant data re-examined for possible explanations. At times, this stage of analysis suggested further observations, which were made to test the validity of the explanations. An example of this type of feedback is given in Chapter VII where additional gut content and field data was collected to investigate hypotheses of E. needhami's ecology that were generated by the previous year's data.
Keltner's (1983) review of the rich history of this type of study is unsurpassed. His ultimate assessment of the state of what he referred to as adaptive analysis and the methodologies employed in such analysis (functional morphology, biomechanics, etc.) was that no single methodology was sufficient. Whereas the present study does not claim to be a single, inclusive framework for adaptive analysis, it does present a holistic approach to the questions of feeding ecology and the function and form of mouthparts. Although specialists in the fields of ethology, morphology, and ecology may perceive a lack of rigor in the techniques used here, the methodologies borrowed from those disciplines form a synergistic approach that surpasses any of the methodologies alone.
An important aspect of any animal's ecology is its role as a consumer. In freshwater ecology, particularly in studies of running water, the consideration of feeding relationships of benthic macroinvertebrates has been crucial to conceptualizing community dynamics and predicting ecological relationships (Cummins 1973, 1974, Vannote et al. 1980). Despite the fundamental importance of such information, feeding habits are often inferred from reports of similar taxa because most species have not received critical autecological study. Conclusions about feeding have usually been based on gut content analyses and casual observations. Unfortunately, gut contents can be misleading because of seasonality and errors introduced in sampling, dissection, and identification of partially digested material (Shapas and Hilsenhoff 1976). It may not be clear which material in the gut is nutritionally important. Casual observations of behavior can seriously underestimate the complexity of an organism's feeding activities.
Perhaps the most difficult task in organizing feeding ecology data is the grouping of organisms into categories descriptive of similar feeding habits. Classifications based on food types are the oldest. Cummins (1973), however, has argued for the need of categories based on functional characteristics. As an example, larvae of the mayfly genus Stenacron, which is the subject of the research reported here, have been placed in a "scraper" functional category by Cummins et al. (1984), defined as consisting of organisms feeding on "attached algae and associated material" (Cummins and Merritt 1984). "Gatherers", defined as organisms that collect detritus from sedimentary deposits, is another functional category in which Stenacron was placed (Cummins et al. 1984). Other authors have defined these groups differently; for example Lamberti and Moore (1984) broaden the definition of scrapers to include organisms that collect material other than algae from crevices in the substrate.
I have studied the functional feeding role of Stenacron interpunctatum by analyzing functional morphology and feeding behavior. This approach allows me to define more accurately the feeding role of the animal, and also to make informed estimates about the functional roles of other organisms with similar morphology. Other studies of mayflies have combined observation and morphological analyses (Brown 1960, 1961, Froehlich 1964, Rawlinson 1939, Schonmann 1981, Strenger 1953, 1975). My study, however, differs considerably by incorporating a recording device for detailed motion analyses, scanning electron microscopy (SEM) for analysis of fine structure and topology, and novel viewing angles. In addition, my primary observation system was designed to duplicate natural stream conditions (i.e. flow, light levels, orientation) more closely than the studies cited above.
I have chosen S. interpunctatum as my prototypic subject because it is common in the eastern half of North America and easily cultured. Furthermore, mature larvae are large enough (head capsule width >2.5mm) to observe with ease, and individuals continue to feed even under the high illumination often required for documentation. I view this study as ecologically important because Stenacron and its relatives in the family Heptageniidae are often important or even dominant in certain lotic habitats (Bednarik and McCafferty, 1979). My objectives in this study were to document the feeding behavior of S. interpunctatum and identify the structures utilized in feeding. Because this study is part of a larger, comparative investigation of the feeding behavior and associated morphology of larval mayflies, I have deferred critical study of the nutrition of this species, and I have not attempted to prepare a complete ethogram of its behavioral repertoire.
Organisms were collected from early spring to late fall from cobble substrate in the Tippecanoe River in north-central Indiana, and were maintained in the laboratory in large aquaria filled with substrate and water from the collection site. Circulation in the aquaria was provided by airstones placed on the bottom. Only large (head capsule width >2.5mm) larvae were used for observation and videotaping.
Individuals were observed by the following techniques. Initially, to minimize possible distractions, individuals were observed in a large light proof observational theater insulated to reduce sound vibrations. The observer was outside the theater, and by remote control operated the video camera (Panasonic WV-1850 with 1" Extended Red Newvicon S4119 pickup tube) mounted inside the theater for observation. The theater and monitoring system, which were designed for a study of burrowing behavior, were described in more detail by Keltner and McCafferty (1986). Light was provided either by an externally mounted fiber optic illuminator equipped with Kodak Wratten No. 87c filters or by infrared light-emitting diodes (LEDs). The filters transmit less than 0.1% of the light below 780nm, the LEDs transmit at 880nm; thus the light used was essentially non-visible to humans. High reflectiveness by Stenacron eyes and the lack of observed responses to changes in infrared light intensity indicated that the infrared light sources were non-visible to the study organisms. Other observations were carried out under visible light which allowed smaller lens diaphragm openings to be used, thus increasing resolution and depth of field. A variety of lenses and lens extensions were employed to allow magnification up to 80x on a 9" (23cm) high resolution black and white video monitor. A 1/2" (12.7mm) VHS format Panasonic NV-8950 video recorder with freeze-frame, slow motion, and reverse motion capabilities allowed images to be recorded and analyzed.
Organisms were placed for observation in a cell built of 1/2" (12.7mm) thick clear acrylic (Fig. 2), its center excavated to form a 1 x 3cm rectangular depression. Several cells were constructed with depths ranging from 1 to 5mm. At each end of the depression several small holes were drilled down into the material to join with larger holes leading out to the edge of the acrylic. Aquarium tubing was cemented into these larger holes. The depression was surrounded with self-adhesive automotive gasket material and covered with a 75 x 25mm glass microscope slide. The slide was held in place by a second piece of clear acrylic, 1/4" (6.35mm) thick, that had its center removed to correspond with the depression; this hole was also surrounded by gasket material. The apparatus was held together by bolts inserted through holes in both pieces and secured with wing nuts. The observation cell was placed in the theater with the long axis in a vertical position and the glass slide facing the camera. The stage on which the cell was mounted was moved in three axes by electric motors.
Water at room temperature (20oC) was circulated through the cell by pumping or siphoning it into the cell through the lower tube and out the upper. The flow was controlled by a valve; however, current velocities around the larvae (as measured by timing the movement of small particles) varied even within a given experiment. The larvae could occupy a significant portion of the cross-sectional space within the shallowest observation cells; this had a dynamic effect on local current velocities and produced a complicated pattern of streamlines of varying current velocities within the cell. For most experiments flow was adjusted until the larvae began to orient. This adjustment was necessary because of the difficulty in measuring current velocity (or deciding which velocity was significant to the organism). Orientation usually occurred at "low" current velocities (0-5cm/s).
Food was introduced into the observation cell by several means. Algae taken from the Tippecanoe River were cultured in the laboratory directly on the glass slide forming the viewing wall of the observation cell. The slides were covered primarily (95%) with Cocconeis; bacteria were common between these diatoms. The remaining 5% was covered by scattered fungal hyphae, Oscillatoria, Cladophora, Melosira, Fragilaria, Chlorella, rotifers, and protozoa. Slides were examined microscopically before and after the observation period. Feeding was considered effective if the slides contained patches where the algae had been removed, or if actual removal and ingestion were observed. Detritus taken from the holding tanks was placed in the cell and feeding was observed at low and high flow rates. Feeding on detritus was considered effective if ingestion was observed.
Motion analyses were carried out primarily by examination of the videotape recordings of the observation sessions of the organisms in the cells; however, this was supplemented by direct video observation. Over 20 individuals of S. interpunctatum were observed for a total period of over 50 hours in the course of this study. Ten of these individuals were videotaped in the observation cells; 236 feeding cycles from 8 of these individuals (4 larvae feeding on attached algae and 4 larvae feeding on detritus) were analyzed in detail using the single frame advance feature of the videotape recorder. This information was used to assemble a chronology of mouthpart movements which was checked against the remaining videotape. In all, several thousand feeding cycles were examined. Following the protocol of Keltner and McCafferty (1986), the observed feeding behavior was classified into stereotypic cycles describing different methods of obtaining and handling possible food material. Each cycle was divided into stages to describe the choreography of movement of the suite of mouthparts For our purposes the stages are delineated by changes in the movements of the labium and the labial palps.
Additional observations were made of organisms on more natural substrate in aquaria and in small artificial streams to assess the effect of the theater and observation cell on behavior, and to check the accuracy of the descriptions of the feeding cycles. These observations were made by the unaided eye, through a stereomicroscope, and with the video camera. These methods did not provide the detail and/or repeatability necessary for critical analysis of mouthpart movement, however, resolution is greater during the initial, direct observation than on videotape playback. I have found this also to be true for other species of mayflies and other aquatic invertebrates that have or are being studied using similar techniques.
Mouthpart structures were examined by light microscopy and SEM. This information was used to construct a movable plastic model of the mouthparts. Gut contents of living larvae were placed on glass microscope slides and examined immediately with a compound microscope. Gut contents of individuals freshly collected in the field were compared to those held in the laboratory. Gut contents were classified as being mineral, organic detritus, diatoms, filamentous algae, unicellular algae, animal remains, or bacteria. No attempt was made to quantify proportions other than to note what type predominated.
The open canopy at the collection site permitted algal growth to cover the substrate. The aufwuchs community was more diverse than that cultured on slides in the laboratory. Additional taxa of diatoms such as Gomphonema, Gyrosigma, Navicula, and Synedra were also present. Certain taxa such as Cladophora, Melosira, and Fragilaria exhibited seasonal blooms to a point where they visibly dominated the periphyton community. At times large areas of the substrate were covered by blackfly larvae and pupae, or by the silk retreats of Petrophila (Lepidoptera) or chironomid larvae.
The larvae collected in the field were, without exception, collected from the bottoms of stones. Although sampling was not quantitative, comparable numbers of S. interpunctatum larvae were found among the cobble substrate both in the open canopy areas and under an adjacent highway bridge where shading visibly reduced algal growth. Larvae were more commonly collected from areas of slower current.
Mature S. interpunctatum was highly stereotypic in the movement of its mouthparts in all experimental enclosures (observation cells, artificial streams, holding aquaria with natural substrate). The few variations in feeding movements observed were apparently associated with food type and are described below. In every case, movements were independent of experimental enclosures.
External factors such as noise, vibration, changes in visible light intensity or water velocity did not affect the qualitative performance of feeding movements of S. interpunctatum. They did, however, affect the duration of movements and the release of stereotypic feeding movements. Feeding movements induced by external stimuli ceased after 1-2 cycles of the mouthparts; those that appeared to be induced by the organism typically continued for at least 5 cycles. Feeding movements observed under visible light were identical to those viewed with infrared light. Specimens observed in the artificial streams and holding aquaria were negatively phototactic, positively thigmotactic and positively rheotactic.
Resting position: When the subjects are not feeding, their mouthparts (Figs. 3, 4, 5-9) are held in a characteristic resting position (Fig. 10). The distal segments of the labial palps are held folded inward so that they lie over the glossae and paraglossae. The labium as a whole is held adducted dorsally to close the preoral cavity ventrally. The galealaciniae of the maxillae are positioned with their median setae touching the lingua of the hypopharynx; the galealaciniae also are in contact ventrally with the labial palps and dorsally with the superlinguae of the hypopharynx. The distal segments of the maxillary palps are held folded over the anterior opening of the preoral cavity. The mandibles are held with the denticles separated, and the molae presumably closed. The labrum is held pressed against the head capsule dorsally and the mandibles ventrally.
Brushing cycle: Individuals placed in the observation cell with algae grown on the glass slide performed a behavioral cycle I term brushing. A brushing cycle is comprised of several stages delineated by the movements of the labium. A single brushing cycle, as described below, takes approximately 1 second at 20o C. The mouthpart movements described below are often preceded by vigorous movement of the forelegs ahead (upstream) of the individual. The legs are repeatedly extended and then retracted, drawing the claws over the substrate; this movement ceases during the first cycle. I did not find any evidence that such movement of the forelegs displaces tightly accreted material by the forelegs.
Stage 1: -- Abduction of the basal segments of the labial palps: As feeding commences, the labium begins to swing ventrally towards the substrate. The basal segments of the labial palps swing posteriorly about 10o (Fig. 11); this movement partially displaces the distal segments of the labial palps from contact with the glossae and paraglossae. At this point the basal segments of the maxillary palps swing posteriorly. The galealaciniae do not move in Stage 1 of the first brushing cycle, but in all subsequent brushing cycles they move medially at this point. The labrum begins to move down and backwards to shield the front of the other mouthparts.
Stage 2: -- Abduction of the distal segments of the labial palps: Once the basal segments of the labial palps reach their posterior position, the distal segments of the labial palps are abducted, moving through a range of about 30o. As the apical tips of the labial palps clear the glossae and paraglossae, the labium completes its ventral swing coming to rest against the substrate with the labial palps fully extended, yet remaining under the contour of the head capsule (Fig. 12). At this point the galealaciniae are abducted, moving laterally outward. The maxillary palps have been extended and reach beyond the lateral borders of the head capsule. The denticles of the mandibles begin to move laterally outward. The labrum completes its movement posteriorly and ventrally, effectively shielding the mouthparts under the head capsule from the current.
Stage 3: -- Retraction of the labial palps: The labial palps are retracted from their extended position by a combination of an anterior swing of the basal segments of the palps and adduction of the distal segments of the palps. This movement is continuous and coordinated so that the palps always remain under the head capsule. As the distal segments of the labial palps move they brush (sweep) up material from the substrate with the profuse setae on the ventral and anterior surfaces. When the distal segments of the labial palps reach the paraglossae they are raised dorsally by a rotation of the basal segments of the palps so that they pass dorsal to the paraglossae and glossae. The distal segments of the palps reach a position that is more median than in the resting position (Fig. 13). The entire labium is raised as soon as the distal segments of the labial palps are dorsal to the glossae and paraglossae. The galealaciniae follow the labial palps in moving medially; they also move somewhat ventrally. The maxillary palps sweep inward, brush over the substrate, and enter the preoral cavity dorsal to the superlinguae. The denticles of the mandibles complete their lateral outward movement and the maxillary palps are inserted among these denticles (Fig. 14). The labrum moves dorsally.
As feeding continues, this brushing cycle is repeated with the abduction of the basal segments of the labial palps as in Stage 1 (foreleg movement is not repeated). The brushing cycle is repeated unchanged except that in Stage 1 of subsequent cycles the galealaciniae are adducted medially, and the posterior movement of the basal segments of the maxillary palps draws the apical setae of the maxillary palps through the combs on the denticles of the mandibles.
Gathering cycle: When S. interpunctatum is presented with loose detritus, a gathering cycle is performed. The gathering cycle is similar to the brushing cycle except for the following. The mouthparts are not pressed tightly against the substrate in Stage 2, and the legs, particularly the forelegs, may bring detritus to the preoral cavity where it is swept up by the labial palps as they move inward in Stage 3. The lateral setae on the distal segments of the labial palps carry more material from the substrate than the ventral setae. In Stage 3 the apical setae of the maxillary palps bring a larger volume of detritus to the preoral cavity than in the brushing cycle; this material is combed out by the mandibular denticles when the maxillary palps are withdrawn in Stage 1.
When presented with filamentous algae such as Cladophora sp., the mouthparts are positioned in Stages 2 and 3 so that the labial palps brush over the filaments, removing epiphytes. The maxillary palps play a role in manipulating the filaments. It is possible for the organism to manipulate filaments into the preoral cavity; however, what could be interpreted as attempts to bite off and swallow pieces of such filaments were ineffective.
Filtering cycle: When a heavy load of detritus is moving in the current, S. interpunctatum alters its feeding behavior radically and becomes a passive filter feeder. The filtering cycle is less complex than the other cycles, and the labial palps are not used to gather food. The larva orients so that its head faces into the current and the longitudinal axis of the body is parallel to the flow. The mouthparts are held in the resting position except for the maxillary palps; these are extended until they are perpendicular to the flow. Particles of detritus adhere to the apical setae. Periodically the maxillary palps are retracted into the preoral cavity as in Stage 3 of the brushing cycle. The detritus brought to the preoral cavity is combed out from the apical setae of the maxillary palps by the denticles of the mandibles as in Stage 1 of the brushing and gathering cycles; the detritus is then further processed in the manner described below, except that the labial palps do not gather additional food.
Once food material has been acquired by either the palps of the labium or maxillae in any of the cycles above, it must be positioned for ingestion. This process is carried on continuously during feeding and for several seconds after the last material is acquired from the substrate. The transfer is accomplished by successive fields of setae on the various mouthparts as described below (Fig. 3).
After completion of at least one brushing or gathering cycle, and when the labial palps are extended again (Stage 1), the material on their setae is removed in two ways. The outward lateral movement (Stage 1) of the distal segments of the labial palps causes material to be combed from the ventral setae of the labial palps by the setae on the dorsal surfaces of the glossae and paraglossae. Material on the lateral setae of the distal segments of the labial palps is removed by the comb setae of the galealaciniae as they move medially in opposition to the outward movement of the labial palps (Stage 1). The material in the comb setae is brushed out as these setae are drawn base-first through the setae on the glossae and paraglossae by the ensuing outward lateral movement of the galealaciniae (Stage 2).
In all cycles, material on the maxillary palps is removed by the combs on the mandibular denticles as the palps are withdrawn (Stage 2). In the filtering cycle all material is brought to the mouth in this way.
Material left on the glossae and paraglossae is moved further through a number of processes. As the labial palps return to the preoral cavity (Stage 3), this material is displaced posteriorly, medially, and dorsally by the palps and the new load of material they bear. As the galealaciniae move medially (Stage 1), their median setae push the material between the superlinguae and the lingua (Fig. 15). The lingua bears rows of setae on its dorsum which guides the material to a point where it can be pressed against the mandibular molae for straining and ingestion. Presumably, material left on the denticles is brushed directly onto the dorsum of the hypopharynx by movements of the mandibles; I have not been able to observe this since this area is obscured by the other mouthparts.
The maxillary palps are used in all cycles to displace larger particles from the preoral cavity and to brush material from the comb setae on the galealaciniae. Once food acquisition ceases, the mouthparts (except for the labial palps) remain in motion as described above for several cycles, moving material into position to be ingested.
Organisms collected in the field had little variety in their gut contents. All guts contained mineral material and organic detritus. Diatoms were observed in the guts of only a few individuals collected early in November, 1985, when a dense growth of diatoms, predominately species of Melosira and Fragilaria, covered the substrate in the Tippecanoe River. Whereas some of these diatoms were present in the gut, they were not predominant, and the diatom frustules were intact. None of the larvae collected in the field had filamentous algae, unicellular algae, or animal remains in their guts. Individuals held in the lab had gut contents similar to those newly collected in the field, although some of the individuals held in the laboratory did contain the diatom Cocconeis sp.
In any behavioral study, it is important to assess the impact of experimental protocol on behavior. In this study the need to confine the organisms to a small area to allow for high magnification observation required an enclosure that did affect the overall behavior of the organism. Since my objective was only to document a small portion of the species' overall behavioral repertoire, this was not a significant constraint because the portion of the behavior in which I was interested in was not affected. The brushing cycle and the gathering cycle were performed in an identical fashion in the observation cell as well as in aquaria and in the artificial stream. My observations of other Ephemeroptera, Trichoptera, Diptera, Odonata, Coleoptera, Amphipoda, Gastropoda, and other freshwater macroinvertebrates (See Chapter VIII) suggest that the cyclic movements of the mouthparts are stereotypic and vary little if any between individuals of the same age class of any one species. This conclusion is consistent with other insect feeding studies (Brown 1960, Chance 1970, Devitt and Smith 1985, Froehlich 1964, Pucat 1965).
This study shows that S. interpunctatum is an opportunistic feeder with a behavioral repertoire sufficient to allow it to feed in a number of different ways. The importance of detritus as a food source for this species has been documented by Lamp and Britt (1981); their quantitative gut content data agree with my qualitative observations. My laboratory observations combined with ecological observations from the field suggest that this species is better classified as a collector (gatherer) than as a scraper according to the categories of Cummins and Merritt (1984). In laboratory tests S. interpunctatum was not able to remove tightly adhering diatoms such as Cocconeis sp. from a glass slide; however, individuals on more natural substrates are apparently able to remove this diatom to some extent. The only alga that S. interpunctatum was observed to ingest was an extremely flocculent mass of diatoms with physical characteristics similar to loose detritus. Although the organisms are capable of manipulating filamentous algae and removing epiphytes from surfaces of these plants, they have difficulty biting off sections of the filaments. Filamentous algae, however, have been reported in the gut of Stenonema modestum, a closely related and morphologically similar species (Kondratieff and Voshell 1980).
Microhabitat data also suggest that this species is a collector. I consistently found these larvae on the undersides of stones in crevices where detritus was deposited. Other studies (Wiley and Kohler 1980, Wodsedalek 1912) also stressed the occurrence of S. interpunctatum larvae on the undersides of stones.
Although direct observation and analysis of feeding behavior of earlier instars of S. interpunctatum were not conducted in this study, earlier instars of many species may be detritivores even if they later specialize on other food sources (Cummins 1973, Cummins and Merritt 1984). Since mature S. interpunctatum larvae primarily ingest detritus (Lamp and Britt 1981), the importance of this food source in the diet of this species is significant.
The ability of S. interpunctatum to vary its feeding behavior to either filter or gather depending on the circumstances demonstrates its independence from a single resource such as periphyton. A more tenable classification of S. interpunctatum as primarily a collector (gatherer) has important consequences for ecological investigations incorporating or extrapolating from the river continuum concept.
On the basis of their mouthpart structure, heptageniids, in general, have sometimes been assumed to scrape their food. For example, Brown (1960) stated that the crown setae of the maxillae are used to rake up diatoms; however, this actually was a misinterpretation of Strenger (1953) (in German) who described the function of these setae ("Kammborsten") as combs for the labial palps. I found that the orientation of these setae aligns them precisely with the labial palps, and that they cannot reach the substrate because of the position of the large labium between the maxillae and the substrate. The external chitinized "scraping bars" reported by Morgan (1911) on the distal segments of the labial palps in the heptageniid Epeorus fragilis are not present in S. interpunctatum. Examination of the distal segments of the labial palps of S. interpunctatum by light microscopy revealed structures which could be interpreted as scraping bars, but examination with SEM revealed that they were internal structures (probably apodemes). Further investigations will be needed to see if this is not the case for other heptageniids.
The only mouthparts of S. interpunctatum that reach the substrate are the labial palps and the tips of the maxillary palps. Both of these structures are heavily setose and not fit for scraping in the manner I have observed for radulae of gastropods and the mandibles of water penny beetle larvae (Psephenus) (Chapter VII). I have, however, seen the heptageniid Heptagenia flavescens scrape bacterial film from the substrate (Chapter VII). The terminal ends of the maxillary palps of that species are reinforced and modified as a scraping tool, quite unlike the brush-like maxillary palps of S. interpunctatum. It is possible that the forelegs of S. interpunctatum may be capable of some scraping. They are active before the brushing cycle and are important in bringing food to the mouth in the collecting cycle. The fact that I did not observe them to be effective in removing material from the substrate may be an experimental artifact.
From a functional-morphological standpoint, it is desirable to erect a new category of feeding -- 'brushing', distinguished from scraping by the morphological structures utilized in acquiring material from the substrate. I propose that brushers include organisms that remove material from the substrate using setae, and that scrapers include organisms that use hardened structures such as mandibles or radulae to remove accreted material. Whereas such a distinction is valuable from a functional standpoint, its ecological significance may be less apparent and awaits data about what precise components of the aufwuchs can be removed by each of these functions and what components are nutritionally significant.
Analyses of the feeding behavior and functional morphology of a species are important tools in describing the ecological relationships of that species. Such studies are necessary before ecological assumptions based on morphology are made, and provide additional insight into the relevance of gut content data. As more of these studies are completed within taxonomic groupings and among taxa with similar morphology their usefulness will increase as an aid in understanding community dynamics and evolutionary trends.
I thank John Keltner for his technical assistance and advice; Arwin Provonsha, whose illustrations translate complex three-dimensional relationships to the two-dimensional world of the journal page; and Dan Bloodgood for assisting with field work, construction of the observation cells, and preparation of Figure 2. Purdue University Agricultural Experiment Station Journal No. 10,178.
The feeding behavior and functional morphology of the heptageniid mayfly Stenacron interpunctatum and new methodology for studying such was reported by McShaffrey and McCafferty (1986). In order that these data can begin to be viewed within a larger comparative context I subsequently studied another midwestern mayfly, Rhithrogena pellucida. This mayfly provides an interesting basis of comparison because it is also a stream-dwelling member of the family Heptageniidae but lives in different habitats and possesses somewhat different mouthpart morphology. As with S. interpunctatum, my objectives with R. pellucida were to document both the feeding behavior and feeding structures. I did not attempt to document the entire range of behavior of the species, or to determine the nutritional significance of the material ingested.
My functional morphology/behavioral approach to studying feeding incorporates field observations, gut content analysis, scanning electron microscopy, and videomacroscopy into a comprehensive research protocol. This allows critical examination of questions concerning diet and microhabitat and the related adaptations of feeding behavior and mouthpart morphology. An important feature of this system is its redundancy; all questions are investigated using multiple techniques. Besides the comparative data that may have relevance to the study of adaptation and evolution, there are also obvious ecological applications. One that I have pursued involves the common practice of assigning functional feeding groups (FFG) (Cummins 1973) to the various members of aquatic communities.
Strenger (1953) studied the functional morphology of European species of Rhithrogena and included observations of behavior and interpretation of morphology. Strenger's work, although important, was limited by her inability to film behavior for detailed analysis and the fact that mouthpart morphology was not studied at the ultrastructural level. I was fortunate to have been able to extend such research by having SEM and videomacroscopic techniques available to me. Rhithrogena mayflies have previously been regarded as both collector-gatherers and scrapers (Cummins et al. 1984), with the primary designation being collectors.
Organisms were collected in April, 1986, from cobble substrate in the Tippecanoe River in north-central Indiana. They were maintained in the laboratory at room temperature (20-25oC) in 75x30x30 cm aquaria filled with substrate and water from the collection site; circulation was provided by airstones. Only mature larvae, as judged by relative wingpad development in Ephemeroptera larvae (McCafferty and Huff 1978), were used for observation and videotaping.
Observations were carried out using the techniques outlined in McShaffrey and McCafferty (1986), with some modifications. When studying R. pellucida, all observations were made using visible light supplied by a fiber optic illuminator to maximize resolution. My experience showed that S. interpunctatum feeding movements were the same in visible and infrared light (McShaffrey and McCafferty 1986).
Most observations were made in an observation flow cell as described and illustrated in Figure 2. With R. pellucida, only a shallow (1mm deep) flow cell was used. Water at room temperature (20-25oC) was fed into the cell by gravity from an aquarium mounted on the roof of the observational theater. Water leaving the cell collected in another aquarium; a pump controlled automatically by a depth-sensing switch periodically returned water to the upper tank. Water in both aquaria was aerated. The flow of water was regulated by a valve; current speed was adjusted as described in the study of S. interpunctatum, with the same inherent difficulties in determining precise current speed (McShaffrey and McCafferty 1986). In most cases the flow was increased until the individuals began to orient to the current, this usually occurred at low current velocities (1-5 cm/s).
In addition to the shallow flow cell, a tank cell was also employed. The tank cell (Fig. 16) was constructed of 2 mm thick acrylic, 10.5 cm long, 4 cm wide, and 4 cm deep, and held a volume of 150 ml of water when in use. The front face had a 1 X 3 cm rectangular section removed from the center; a portion of a standard glass microscope slide was attached over the outside of this hole, and a 1 mm2 mesh screen was attached over the inside to produce an enclosure into which the individuals were placed. This enclosure measured 1 cm long, 2mm wide, with a water depth up to 3 cm. A hemicylinder of black plastic (radius 1.5 cm), with 30 2.5 mm diameter holes, was placed around the enclosure on the inside of the tank. This plastic served to isolate the organism visually and reduce unwanted reflections during filming.
Water flow was provided as in the shallow flow cell; the water entered the tank cell through a tube attached with its outlet below the waterline inside the plastic hemicylinder directly behind the enclosure, and exited through another tube attached to one side of the tank cell. Water level in the cell was controlled by regulating inflow and outflow with valves. There was no noticeable water current in the enclosure. The tank cell was mounted in place of the shallow flow cell on the stage of the observational theater. The final position of the organism, vertical with the mouthparts facing the glass slide and the videocamera, was the same as in the flow cell.
Food for the experiments was provided as described in McShaffrey and McCafferty (1986). Diatoms and other periphyton collected in the Tippecanoe River were cultured on glass slides in the laboratory. The composition of the periphyton community on the slides was the same as the previous study - a 95% covering of the diatoms Cocconeis and Achnanthes, numerous bacteria between the diatoms, and the remaining 5% comprised of scattered fungal hyphae, Oscillatoria, Cladophora, Melosira, Fragilaria, Chlorella, rotifers and protozoa. Scanning electron micrographs of typical growth on a slide surface are shown in Figures 17 and 18. Detritus taken from the holding tanks was introduced into the water flow in the shallow flow cell at both high (>5 cm/s) and low (<5 cm/s) flow rates; detritus was placed directly into the enclosure of the tank cell. Detritus settled to the bottom of the tank cell and the shallow flow cell at low flow rates; at high flow rates detritus moved through the shallow flow cell in suspension. I considered feeding effective if ingestion of the material was observed.
Motion analyses were carried out by examination of the videotape recordings of the observation sessions as in McShaffrey and McCafferty (1986). Owing to sparseness of specimens in the field, the need to sacrifice some individuals upon return to the laboratory for gut content analysis, and natural mortality, only 10 individuals were videotaped. Of the several hundred feeding cycles videotaped, 50 were analyzed in detail using single frame advance. The 50 feeding cycles chosen for detailed analysis were selected for visual clarity and presence of repeated feeding activity. Once a chronology of mouthpart movement was determined within these cycles, the chronology was tested against the remaining cycles. The feeding cycles are classified into stereotypic cycles as outlined in McShaffrey and McCafferty (1986) and adapted from the methods used by Trueman (1968) and Keltner and McCafferty (1986). Each stage is delineated by specific movements of the labial palps.
Mouthpart structures were examined using light microscopy and SEM. Specimens were prepared for SEM by dehydration in increasing concentrations of ethanol from 75% to 100%, then transferred to 100% ethyl acetate. Some specimens were sonicated in 100% ethyl acetate for one minute in an ultrasonic cleaner to remove food material from the mouthparts, others were not sonicated so that the position of material on the mouthparts might be determined. The larvae were then either dissected or transferred whole to the SEM stubs after air drying. This method produces very little distortion of mouthparts and is much faster than critical point drying.
Gut contents of living larvae were placed on glass microscope slides and examined immediately with a compound microscope. Gut contents were classified as being mineral, organic detritus, diatoms, filamentous algae, unicellular green algae, animal remains, or bacteria. No attempt was made to quantify proportions other than to determine what type predominated. Periphyton species present in the field were determined by returning substrate to the laboratory, carefully removing the periphyton, and examining it under a compound microscope.
My knowledge of the life cycle of R. pellucida in the Tippecanoe River is based on extensive collections from the same site from 1983 to 1987. Diligent searching of all microhabitats during weekly intervals from April 1985 to July 1986 produced R. pellucida samples only during April and May. Most of these larvae were relatively mature and were always collected off the upper surfaces of stones or collected in kick screen samples. The larvae were also always found in mid-stream where the substrate receives direct sunlight and the periphyton community is well developed.
When R. pellucida larvae were found, the cobble substrate was covered by a diverse aufwuchs community, including diatoms such as Gomphonema, Navicula, Fragilaria, Melosira, Cocconeis, Achnanthes, and Synedra. Cladophora, which was also present throughout the period when R. pellucida larvae were found, reached a growth peak in early May at the same time R. pellucida adults emerge. It appears that R. pellucida in the Tippecanoe River is univoltine, with emergence and egg-laying in early May. Although I do not have data on the egg stage or the early instars, it is probable that the eggs hatch in winter, with mature larvae appearing on the upper surfaces of rocks in April.
Rhithrogena pellucida larvae were highly stereotypic in their mouthpart movements. External factors such as noise, vibration or change in light intensity often appeared to initiate feeding movements, however, these movements were qualitatively similar to those observed under other conditions. Feeding cycles induced by external stimuli usually ceased after 1-2 cycles, whereas non-induced feeding cycles were usually repeated at least 5 times. Specimens in the holding aquaria were positively rheotactic. During the course of this study, larvae never attempted to feed on detritus or to filter feed.
Resting position: When a larva is not feeding, the mouthparts (Figs. 19, 20, 21-25) are held in a characteristic resting position (Fig. 26). The distal segments of the labial palps are held folded inward so that they lie over (dorsal to) the paraglossae but not the glossae. The labium as a whole is held adducted dorsally to close the preoral cavity ventrally. The galealaciniae of the maxillae are positioned with their median setae touching the lingua of the hypopharynx; the galealaciniae also are in contact ventrally with the labial palps and dorsally with both the superlinguae of the hypopharynx and the mandibles. The maxillary palps are held with the basal segments perpendicular to the long axis of the body; the distal segments are folded forward at a 45o angle to the basal segments. The basal segments of the maxillary palps do not extend beyond the contour of the head capsule; the distal segments of the maxillary palps parallel the contour, extending just slightly beyond it. The apices of the distal segments of the maxillary palps do not meet but extend only to the lateral edges of the labrum. The labrum and the maxillary palps together cover the anterior and lateral aspects of the preoral cavity. The labrum is in contact with the denticles of the mandibles ventrally. The denticles of the mandibles are held separated, and presumably the molae are closed.
Feeding cycles: Larvae placed in the observation cell with diatoms grown on the glass slide performed two different behavioral cycles; I term these brushing and scraping cycles. The two cycles differ in the degree to which the maxillary palps are employed. In the Labial Brushing Cycle, the labial palps are used as brushing tools; in the Maxillary Scraping Cycle, the maxillary palps are used as scraping tools. Each cycle consists of several stages that can be delineated by specific movements of the labium. A single cycle of either type takes 1.2 seconds at 20oC. Mouthpart movements described below are often preceded by and concurrent with movement of the forelegs ahead (upstream) of the individual. Legs are alternately and repeatedly extended and then retracted, drawing the claws over the substratum. These leg movements were capable of removing tightly adhering material from the substrate and bringing it to the mouthparts.
Stage 1: -- Abduction of the basal segments of the labial palps: As feeding commences, the labium begins to swing ventrally towards the substrate. The basal segments of the labial palps swing posteriorly about 20o (Fig. 27); this movement draws the distal segments of the labial palps away from contact with the paraglossae. Once the apices of the labial palps clear the paraglossae, the labium completes its ventral movement and comes to rest against the substrate. At this point the basal segments of the maxillary palps swing posteriorly about 20o. The galealaciniae do not move in Stage 1 of the initial brushing cycle in the series, but in all subsequent cycles of the series they move medially at this point in Stage 1. The labrum begins to move down slightly to shield the front of the other mouthparts.
Stage 2: -- Abduction of the distal segments of the labial palps: Once the basal segments of the labial palps reach their posterior position, the distal segments of the labial palps are abducted, moving through a range of about 50o until they are at an angle of 125o to the basal segments (Fig. 28). When the labial palps are fully extended they extend just beyond the contour of the head capsule. At this point the galealaciniae complete their medial movement and are now abducted, moving laterally outward. The distal segments of the maxillary palps are extended at the same time as the labial palps, moving through an angle of about 60o to make an angle of 105o with the basal segments. The maxillary palps reach well beyond the lateral borders of the head capsule. The labrum completes its movements ventrally and helps shield the mouthparts under the head capsule from the current.
Stage 3: -- Retraction of the labial palps: The labial palps are retracted from their extended position by a combination of an anterior swing of the basal segments of the labial palps and adduction of the distal segments of the palps. This movement is continuous and coordinated so that the labial palps always remain under the head capsule. The maxillary palps follow this movement closely; the labial palps lead the movement and remain positioned just anterior to the articulations between the segments of the maxillary palps (Fig. 29). During this movement the distal segments of the labial palps brush (sweep) up material from the substratum with the profuse setae on their ventral and anterior surfaces. The trailing maxillary palps are not in contact with the substratum but are positioned to capture any particles dislodged by the labial palps. If foreleg movement is concurrent with feeding, the claw of one foreleg is brought to the corresponding labial palp as it begins to be adducted during this stage.
When the distal segments of the labial palps reach the paraglossae they are raised dorsally by a rotation of the basal segments of the labial palps so that they pass dorsal to the glossae and paraglossae. The distal segments of the labial palps reach a position that is more median than in the resting position (Fig. 30). The entire labium is raised as soon as the distal segments of the labial palps are dorsal to the glossae and paraglossae. The galealaciniae follow the labial palps in moving medially; they also move somewhat ventrally. The denticles of the mandibles complete their lateral outward movement and the distal segments of the maxillary palps rotate and are inserted among these denticles (Fig. 31). The labrum moves dorsally.
As feeding continues, this brushing cycle is repeated in a series of these cycles. Subsequent cycles in the series are initiated by a slightly further adduction of the distal segments of the labial palps (Fig. 32), followed by the abduction of the basal segments of the labial palps as in Stage 1. The brushing cycle is repeated unchanged except that in Stage 1 of subsequent cycles the galealaciniae are moving medially and brush across the labial palps as they move outward, and the posterior movement of the basal segments of the maxillary palps draws the setae on the distal segments of the maxillary palps through the combs on the denticles of the mandibles. If leg movements accompany the brushing cycle, then the left and right forelegs alternate in moving to the mouthparts.
Maxillary scraping cycle: This cycle differs from a Labial Brushing Cycle in that the labial palps do not initiate all the movements, and the maxillary palps are brought to the substrate to scrape. Stage 1 is similar to Stage 1 of a Labial Brushing Cycle. In Stage 2, the maxillary palps are extended and lowered to the substrate; the labial palps are positioned in the articulation between the maxillary palp segments. The ventral surfaces of the distal segments of the maxillary palps are covered with unique pectinate setae (Fig. 33) which effectively remove tightly adhering material from the substrate. In Stage 3, the maxillary palps initiate the closing movements, and both sets of palps move over the substrate. The forelegs may also be active in the Maxillary Scraping Cycle.
Once food has been brushed or scraped up by either the labial palps or the maxillary palps, respectively, it must be positioned for ingestion. This process is carried on continuously during feeding and for several seconds after the last material is procured from the substrate. Transfer is accomplished by successive fields of setae on the various mouthparts (Fig. 19) as described below. Food transport is similar for both the Labial and Maxillary Feeding Cycles.
After completion of the first cycle, and when the labial palps are extended again (Stage 1), material on the setae of the distal segments of the labial palps is removed in two ways. Outward lateral movement (Stage 1) of the distal segments of the labial palps causes material to be combed from the ventral setae of the labial palps by setae on the dorsal surfaces of the glossae and paraglossae. Most of the material on the distal segments of the labial palps is carried by the lateral and dorsal setae. This material is removed by the comb setae on the crown of the galealaciniae as the distal segments of the labial palps are adducted further and then moved laterally outward by the posterior movement of 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 labial palps by the slight adduction of the distal segments of the labial palps at the beginning of the outward movement and with a ventral movement of the galealaciniae. Material in the comb setae is removed by stout setae on the dorsal sides of the distal segments of the labial palps when the comb setae are drawn base-first through them as the distal segments of the labial palps are adducted medially (Stage 3). Material on the comb setae may also be removed when the galealaciniae move laterally outward (Stage 2); this draws the comb setae base-first through the setae on the dorsal sides of the glossae and paraglossae. In both types of cycles, any material on the maxillary palps is removed by combs on the mandibular denticles as the palps are withdrawn (Stage 2).
Material left on the glossae and paraglossae and on the dorsal surfaces of the distal segments of the labial palps is moved further through a number of processes. As the labial palps return to the preoral cavity (Stage 3), this material is displaced posteriorly, medially, and dorsally by the labial palps and the new load of material they bear. As the galealaciniae move medially (Stage 1 and Stage 2), their median setae push the material between the superlinguae and the lingua of the hypopharynx (Fig. 31). The lingua bears dorsal rows of setae which guide the material to a point where it can be pressed against the mandibular molae for straining and ingestion (Fig 34). Presumably, material left on the denticles of the mandibles is brushed directly onto the dorsum of the hypopharynx or the venter of the epipharynx by the movements of the mandibles; I have not been able to observe this directly since this area is obscured by the other mouthparts.
In both types of cycles the maxillary palps are used to displace larger particles from the preoral cavity and to brush material from the comb setae on the galealaciniae. Once food acquisition ceases, the mouthparts remain in motion as described above for several cycles, moving material into position to be ingested. During this phase, the labial palps make only small movements, primarily slight adductions of the distal segments, and the maxillary palps only brush against the mandibular denticles and the galealaciniae (they do not sweep the substratum).
Gut contents of four larvae collected in the field were analyzed along with those of one larva that had been videotaped. All five guts were similar in content; they contained diatoms, organic detritus and sediment. The predominate material was intact diatom frustules; there was also considerable organic detritus apparently of algal origin. Diatoms found in the gut included: Navicula, Gomphonema, Melosira, Cocconeis, Achnanthes, Fragilaria, and Rhoicosphenia curvata; Navicula species were the most common. None of the guts contained intact filamentous algae or recognizable animal remains.
My observational procedures required the organisms to be confined inside a small cell made of synthetic materials, and it exposed them to unnatural light and water flow regimes. In my study of S. interpunctatum (McShaffrey and McCafferty 1986), I found that the experimental conditions did affect the overall behavior of that species, but did not affect the actual feeding movements reported. With S. interpunctatum, I was able to study individuals not confined to an observation cell, and I made observations using infrared light to simulate the low-light nature of the crevices in which I collected the larvae in the field (McShaffrey and McCafferty 1986). I was not able to collect enough R. pellucida larvae to permit observations outside the observation cell, but experience with S. interpunctatum, Ephemerella needhami, and various other Ephemeroptera, Trichoptera, Diptera, Odonata, Coleoptera, Amphipoda, and other freshwater macroinvertebrates (unpublished data) suggests that the cyclic movements of the mouthparts are stereotypic and vary little between conspecifics of the same age class. This conclusion is consistent with data from other insect feeding studies (Brown 1960, Chance 1970, Devitt and Smith 1985, Froehlich 1964, Pucat 1965).
My observations are similar to those made by Strenger (1953) for a European species of Rhithrogena. I was not able to confirm her suggestion that the distal segments of the labial palps deform their surfaces; this reportedly would enable their setal patches to open and close and thus increase feeding efficiency (Strenger 1953). I also was not able to document the function of the brush on the mola of the right mandible; Strenger (1953) stated that it is used to carry food from the hypopharynx to the left mola. In both cases, it is possible that Strenger (1953) did not actually observe these movements but inferred them based on her extensive morphological studies. I was able to observe some opening and closing of the setae of the distal segments of the labial palps, but can not eliminate the possibility that such movements are passive responses to drag as the palps are moved against the substrate, the water, or the other mouthparts. The molar region of the mandibles is completely obscured by the other mouthparts, and I was not able to observe this area directly. I had to estimate molar action based on the movement of the denticles, which were visible.
Because R. pellucida and S. interpunctatum are both flatheaded mayflies of the subfamily Heptageniinae, and can be found coexisting in the same general area of streams, it is interesting to compare their morphology, microhabitat, and feeding movements. Stenacron interpunctatum is found in slow-current areas, on the bottom of stones, away from the current (McShaffrey and McCafferty 1986, Flowers and Hilsenhoff 1978, Wiley and Kohler 1980, Wodsedalek 1912). I observed R. pellucida living on the tops of stones in high-current areas; Flowers and Hilsenhoff (1978) also found this species in high-current microhabitats. Strenger (1953) reported that the European species she studied were also collected on the tops of stones. The difference in microhabitat is further reflected when comparing the gut contents of these species. Whereas S. interpunctatum is primarily a detritivore (Lamp and Britt 1981, McShaffrey and McCafferty 1986), the R. pellucida studied here are primarily periphyton feeders.
The short developmental period for R. pellucida in the spring (see also Flowers and Hilsenhoff 1978, Clifford 1982) is associated with its feeding habits. Rhithrogena pellucida in the Tippecanoe River is only found in early spring before Cladophora growth is significant, while S. interpunctatum occurs year-round at the same site. There is a narrow temporal window in which there is good periphyton development on the rock substrate. In the Tippecanoe river, this period is during the later part of April to early May. The end of this period is marked by Cladophora overgrowth of the substrate; presumably the filaments of the Cladophora interfere with the feeding of R. pellucida, which requires a relatively flat substrate free of obstructions to the movements of the maxillary palps. Stenacron interpunctatum feeds on detritus, which is always available, and exhibits a complex life history with larvae present year-round (McCafferty and Huff 1978).
Morphological differences in R. pellucida correspond to its different microhabitat and are apparently adaptations allowing it to remove the tightly-bound material on the upper surfaces of rocks. The primary difference is the increased development of the maxillary palps [compare Fig. 22 with Figs. 3 and 6]. Corresponding to increased development of the maxillary palps is a flattening and broadening of the mandibular denticles [compare Fig. 24 with Figs. 3, 4, and 8] to accommodate the larger maxillary palps involved in removal of food particles. Other secondary morphological differences between R. pellucida and S. interpunctatum reflect a decrease in importance of the labial palps as food-gathering organs. There are fewer comb setae on the galealaciniae. Also, the hypopharynx, which is important in transferring material from the labium to the molae, is greatly reduced [compare Fig. 23 with Fig. 7]. Strenger (1953) noted similar differences when comparing Rhithrogena and Ecdyonurus in Europe.
The brushing cycle of S. interpunctatum is very similar to the Labial Brushing Cycle of R. pellucida, the major difference being an improved coordination of the movements of the maxillary and labial palps in R. pellucida. This improved coordination places the maxillary palps in a position closely following the labial palps to capture any material swept up by the labial palps but not captured by them. Such improved efficiency may be vital in a higher-current microenvironment. The Maxillary Scraping Cycle of R. pellucida is much more effective tha
