Contrasting diel activity and feeding patterns of four instars of Rhyacophila dorsalis (Trichoptera)

1. Ontogenetic shifts in prey choice and predator behaviour can affect food‐web structure. Therefore, it is important to establish if the diet and feeding activity differ between life‐stages of the same species. This hypothesis was tested for second, third, fourth and fifth larval instars of Rhyacophila dorsalis by comparing their diel activity and feeding patterns. Second to fifth instars collected from two streams were used either for gut analyses or for observations of their activity and feeding patterns in three stream tanks. Food was provided in excess; being organisms living in bryophytes on top of a large stone in each tank, augmented by different‐sized larvae of Ephemeroptera, Simuliidae and Chironomidae. As few first instars for gut analyses were found in the field, the diet of first instars reared in the laboratory was also studied. 2. Larvae for gut analyses were taken 1 h before dusk or dawn (n = 50 larvae per instar for each day or night sample). First and second instars fed on the smaller food items with no significant day‐night differences in diet. Gut contents indicated a progressive trend from feeding chiefly at night in third instars to almost exclusively at night in fifth instars. Fourth and fifth instars fed on the larger food items, whilst the diet of the third instar larvae overlapped with that of both the earlier and later instars. 3. Diel activity patterns of single larvae differed between instars but not within each instar (n = 20 larvae per instar). Second instars were active throughout the 24 h, with peaks at dusk, around midnight, dawn and around midday. A similar pattern was shown by third instars but the peak of activity at midday was less than the other three peaks. Prey were captured only during these peaks for both instars. Fourth and fifth instars were most active, and fed only, at night. They used an ambush strategy to capture more active prey at dusk and dawn (e.g. Baetis, Gammarus), and a searching strategy to capture more sedentary prey during the night (e.g. chironomids, simuliids). These experiments provided support for the hypothesis under test. If competition and/or interference occur between instars, then it could be reduced between earlier and later instars because of differences in their diet and diel pattern of feeding activity.     

Studies on the biology of Chaoborus flavicans (Meigen) (Diptera: Chaoboridae) in a fish-free eutrophic pond,Japan

Population dynamics of Chaoborus flavicans larvae of various instars was studied from November 1986 to December 1987 in a eutrophic, fish-free pond, Japan. First and 2nd instar larvae were observed from late April to late October, indicating a reproductive period of about half a year. C. flavicans overwintered in the 4th instar larvae. In water column samples, total density of all instars was 680–23680 m^-2, and pupal density 0–2600 m^-2; larvae of the 1st, 2nd, and 3rd instars showed 5–6 density peaks in 1987, suggesting that 5–6 generations occur during a year (peaks of the 4th instar larvae were not clear, probably due to their longer development than those of younger instars). In sediment samples, no 1st and 2nd instar larvae were found, 3rd instar larvae were found occasionally but density of the 4th instar larvae was 280–18600 m^-2, and pupal density varied between 0–502 m^-2. Fouth instar larvae accumulated in sediment in the cold season and in the water column in the warm season; high temperature and low oxygen concentration were the most important factors limiting the distribution of larvae in the sediment in summer in the NIES pond. The dry weight of total C. flavicans larvae was 0.08–4.2 g m^-2 in sediment samples and 24–599 μg l^-1 (0.10–2.40 g m^-2) in water column samples. Comparisons of maximum densities in the NIES pond in different years and in waters of different trophic status show that density is generally higher in eutrophic than in oligotrophic habitats. This revised version was published online in July 2006 with corrections to the Cover Date.     

Effects of oxygen deficiency on survival and glycogen content of Chironomus anthracinus (Diptera,Chironomidae) under laboratory and field conditions

Growth and glycogen content of Chironomus anthracinus in Lake Esrom, Denmark was examined during summer stratification in 1992 and 1993. Simultaneously, effects of oxygen deficiency on glycogen utilization and survival were experimentally studied. The population consisted of almost fullgrown 4th instar larvae in 1992 and 2nd and 3rd instar larvae in 1993. Growth rate and glycogen content changed as hypolimnetic oxygen deficiency increased. During a 1st phase of stratification dry weight and glycogen content increased (2nd and 3rd instars) or was almost constant (4th instar) but decreased significantly during the following 2nd phase. This change from growth to degrowth and utilization of endogenous glycogen reserves correlated with a change in the thickness of the microxic layer (<0.2 mg O₂ 1^–1) above the sediment surface. The layer increased from 2–3 m in phase 1 to 4–5 m in phase 2, and we suggest that this deteriorated the oxygen conditions and resulted in a change in larval energy metabolism from fully aerobic during the 1st phase to partly anaerobic in the 2nd phase. During the 2nd phase larval metabolism was estimated at less than 20% of normoxic rate. Experimental exposure of the larvae to anoxia indicated highly different survival of young larvae (2nd and 3rd instars) and older larvae (large 4th instars). The morality of young larvae was 50% after three days in anoxia at 10 °C, whereas only 25% of the older larvae had died after 3–4 weeks under similar conditions. Extending the treatment, however, resulted in increased death rate of the 4th instar larvae with only 10% surviving after seven weeks. The anaerobic metabolism of 4th instar larvae as estimated from glycogen degradation at 10 °C was 5% of normoxia in the interval from 0–5 days but 1.5% in the interval from 20–25 days. It is concluded that survival of C. anthracinus in anoxia is very limited, but traces of oxygen in the environment allowing for faint aerobic metabolism prolong the survival time of the larvae from a few days (2nd and 3rd instars) or a few weeks (4th instar) to probably 3–4 months.     

The growth of Bibio johannis (L.) and Dilophus febrilis (L.) (Diptera: Bibionidae) larvae in the field

Bibionid larvae were extracted from soil samples collected from June 1985 to March 1987. Each larva was identified, weighed and measured for maximum head capsule width. Both Bibio johannis and Dilophus febrilis have six larval instars. Bibio johannis grew slowly in late summer and early autumn but rapidly from October to January. Early and late instar larvae of D. febrilis occurred simultaneously and it was concluded that this species was not bivoltine but may exist as two separate cohorts with distinct flight periods.     

Predation by three hemipterans:Tropiconabis nigrolineatus,Oechalia schellenbergii andCermatulus nasalis,onHeliothis punctiger larvae in two types of searching arenas

Three species of hemipteran predators preyed differently upon 1^st instarHeliothis punctiger Wallengren larvae.Cermatulus nasalis consumed more larvae thanOechalia schellenbergii which consumed more larvae thanTropiconabis nigrolineatus. All the species consumed significantly less 1^st instar larvae on plants than what they consumed in Petri-dishes. Fifth instar predators showed significant differences in terms of prey consumption due to sex independent of searching conditions. Only 4^th and 5^th instars ofT. nigrolineatus attacked and captured 2^nd instars ofH. punctiger larvae. The other 2 species however readily attacked and consumed 2^nd instarH. punctiger larvae. Their prey consumption was similar in Petri-dishes and on plants. Only 5^th instars ofT. nigrolineatus could subdue and capture 3^rd instarH. punctiger larvae. Second instar pentatomids captured just one 3^rd instar larva but older instars killed and ate more. Fourth instarH. punctiger larvae were immune to attacks by allT. nigrolineatus and younger pentatomids due to their defense ploys but 5^th instar pentatomids could subdue and capture them. None of the predators captured 5^th instarH. punctiger larvae except few 5^th instar females ofC. naslis andO. schellenbergii.      

Onshore–offshore variations of copepod community in northern South China Sea

Previous studies which have tested the feeding preferences of shredders for fungal species and the food quality of fungi used detritus uniformly colonized by a fungus, which is not the case for decaying leaves in streams. It is not known whether shredders in different development stages exhibit variations in feeding preference and larval performance. This study examined the feeding preferences and the growth of the third and the fifth instars of Pycnopsyche gentilis larvae using fungal-colonized patches and whole leaves, respectively, having different fungal species compositions (Alatospora acuminata, Anguillospora filiformis, Articulospora tetracladia, Tetrachaetum elegans, and all species combined). The aquatic hyphomycetes used were co-dominant on leaves in the stream inhabited by the caddisfly. During 14 d of feeding, the larvae of both instars did not show significant differences in feeding preferences for the patches growing on oak leaves, although the third instar larvae were slightly more selective than the fifth instar larvae. When fed with maple leaves for 18 d, larval growth rates, gross growth efficiencies, and survivorship were not significantly different among the fungal treatments. However, the larval growth of both instars fed with fungal-colonized leaves was always significantly greater than the growth of larvae fed with diets of uncolonized leaves. The third instar larvae grew faster than the fifth instar larvae, but the growth efficiencies of the two instars were similar. These results suggest that P. gentilis larvae exhibit less selectivity in their feeding than other caddisfly shredders that have been examined and that the dominant fungi colonizing leaves in their habitat are similar in palatability and food quality for this shredder. Handling editor: B. Oertli     

Chaoborus crystallinus predation on Daphnia pulex: can induced morphological changes balance effects of body size on vulnerability?

Juvenile Daphnia pulex form neckteeth in reponse to chemicals released by predatory Chaoborus crystallinus larvae. Formation of neckteeth is strongest in the second instar followed by the third instar, whereas only small neckteeth are found in the first and fourth instar of experimental clones. Predation experiments showed that body-size-dependent vulnerability of animals without neckteeth to fourth instar C. crystallinus larvae matched the pattern of neckteeth formation over the four juvenile instars. Predation experiments on D. pulex of the same clone with neckteeth showed that vulnerability to C. crystallinus predation is reduced, and that the induced protection is correlated with the degree of neckteeth formation. The pattern of neckteeth formation in successive instars is probably adaptive, and it can be concluded that neckteeth are formed to different degrees in successive instars as an evolutionary compromise to balance prediation risk and protective costs.     

Larval interactions in aphidophagous predators: effectiveness of wax cover as defence shield of Scymnus larvae against predation from syrphids

Small-sized predators in the aphidophagous guild of Aphis gossypii Glover colonies on hibiscus trees in Japan exploit aphids at low prey abundance. Scymnus (Pullus) posticalis Sicard beetles were the first predatory species to attack aphids in the spring, and their larvae co-occurred with larvae of Eupeodes freguens (Matsumura) syrphids in aphid-infested leaves of hibiscus for 3 weeks in absence of large-sized coccinellid predators. Larval interaction between Scymnus and syrphid predators was examined in relation to effectiveness of wax cover of Scymnus against predation from syrphids. Waxless first instar larvae were not protected but wax-covered larvae of second, third and fourth instars were protected from predation by syrphid larvae. The protection was lower in the second instar which has a thin wax cover and significantly higher in the third and fourth instars having a thick wax cover. In addition, larvae from which the wax was removed were significantly more vulnerable to predation. Vulnerability of Scymnus larvae to predation from syrphids was directly related to the thickness of wax cover. Results suggest that the wax cover of Scymnus larvae act as an effective defence mechanism against predation from syrphid larvae.     

Does superparasitism improve host suitability for parasitoid development? A case study in the Microplitis rufiventris-Spodoptera littoralis system

Superparasitism refers to the oviposition behavior of parasitoid females who lay their eggs in an already parasitized host. Recent studies have shown that allocation of additional eggs to an already parasitized host may be beneficial under certain conditions. In the present work, mortality of Microplitis rufiventris wasps was significantly influenced by both host instar of Spodoptera littoralis larvae at parasitism and level of parasitism. In single parasitization, all host instars (first through sixth) were not equally suitable. Percentage of emergence success of wasp larvae was very high in parasitized first through third (highly suitable hosts), fell to 60% in the fourth instar (moderate suitable) and sharply decreased in the penultimate (5th) instars (marginally suitable). Singly parasitized sixth (last) instar hosts produced no wasp larvae (entirely unsuitable), pupated and eclosed to apparently normal adult moths. The scenario was different under superparasitism, whereas supernumerary individuals in the highly suitable hosts were almost always killed as first instars, superparasitization in unsuitable hosts (4th through 6th) had significant increase in number of emergence success of wasp larvae. Also, significantly greater number of parasitoid larvae successfully developed in unsuitable hosts containing three wasp eggs than counterparts containing two wasp eggs. Moreover, the development of surplus wasp larvae was siblicidal in earlier instars and nonsiblicidal gregarious one in the penultimate and last “sixth” instars. It is suggested that the optimal way for M. rufiventris to deal with high quality hosts (early instars) is to lay a single egg, while the optimal way to deal with low quality hosts (late instars) might be to superparasitize these hosts.     

Movement and feeding patterns of Epilachna cucurbitae Richards (Coleoptera:Coccinellidae) on pumpkin and zucchini plants2

The oviposition patterns of adults and the movement and feeding patterns of larvae of Epilachna cucurbitae on two species of cucurbits, Cucurbita maxima cv Queensland Blue and C. pepo cv Blackjack, were studied in the field and laboratory. The physical and nutritional characteristics of host plant leaves of different ages were described. Younger leaves had higher nitrogen contents but were less abundant, smaller and had higher trichome densities than older leaves. The development of first instar larvae was delayed by the leaf hairs on young and mature pumpkin leaves which prevented larvae from reaching the leaf surface to feed First instal larvae developed more quickly on leaves rich in nitrogen. Neither the total developmental time of larvae nor the size of pupae was affected by leafage because larvae on poor quality leaves compensated by eating more. Female beetles oviposited on all but the youngest and oldest leaves of the host plant. The trichomes on young leaves prevented females from attaching eggs to the leaf surface. First instar larvae remained where they hatched, but older larvae were more mobile, Changing feeding sites frequently and moving progressively to younger, more nutritious leaves. Final instar larvae moved onto adjacent vegetation to pupate. The adaptive significance of these patterns is discussed in relation to the nutritional value, hairiness and abundance of host plant leaves of different ages and the physical limitations of different larval instars.     

Some observations on the food of larvae of Procladius bellus (Diptera: Chironomidae)

Analysis of the gut contents of Procladius bellus (Loew) larvae collected in field samples indicated that it was a detrivitore‐omnivore. The observed dietary change from first and second instars (detritivore‐herbivores) to third and fourth instars (omnivores) did not suggest a change in feeding behaviour from detritivore to predator, but rather that the animal material was consumed along with larger detrital material. Small detrital material (1–15 μm) was an important dietary component and detritus in general contributed 50–70% of the overall diet. Diatoms and green algae contributed 15–20%, while blue‐green algae contributed about 10%. Animal material contributed about 5% to second and third instar larvae and 20% to fourth instar larvae.     

Life History of Bittacomorpha clavipes (Fabricius) (Diptera: Ptychopteridae) in an Ozark Spring,U.S.A.

A population of Bittacomorpha clavipes was studied in an Ozark spring from July 1987- August 1988. Measurement of larval head capsules (N=4,544) showed B. clavipes has four distinct larval instars. Seasonal distribution of immature and adult life stages suggests that this population has an asynchronous, non-seasonal, multivoltine life history with at least three cohorts. First instar larvae and pupae were collected during all months except December through March and December through February, respectively. Adults and larval instars II-IV were collected throughout the year. Larval recruitment generally paralleled adult flight periods, and the presence of adults during all months sampled indicates a correlation with ovipositional periods. The life history documented here for B. clavipes suggests that limiting factors for growth and reproduction are not seasonally dependent for this population.     

Mate recognition inCryptomyzus aphids: copulation and insemination

The acceptability of three widely distributed Australian Menispermaceae,Tinospora smilacina Benth.,Sarcopetalum harveyanum F. Muell. andStephania japonica (Thunb.) Miers, as food for larvae of the fruitpiercing moth,Othreis fullonia (Clerck), was examined in three laboratory experiments. When larvae were presented with plant species individually total development times were shortest onT. smilacina and longest onS. japonica, despite relatively similar consumption rates within most instars.T. smilacina elicited greater (P<0.05) relative growth rates thanS. japonica in all instars except the 6th. In the second experiment, when larvae were allowed to select from each of the 3 plants, noS. japonica was chosen by 1 st instars and it represented only 3.7% of food consumed by 2nd instars. Significantly moreT. smilacina was eaten in each instar thanS. japonica, and more thanS. harveyanum except in the 2nd and 4th instars. The final experiment examined the abilities of larvae to switch hosts when forced after the 1st and 3rd instars. After the first or second food change largest average headcapsule widths were associated with feeding onT. smilacina as the most recent food. Feeding by final instars onT. smilacina also resulted in the shortest development time and highest puparial weights. While some larvae survived irrespective of plant sequence 83.3% of the recorded mortality occurred while larvae were exposed toS. japonica, principally during the 1st instar. These experiments lend support to field observations which suggest thatT. smilacina is a major host ofO. fullonia whileS. japonica is notS. harveyanum is probably an important alternate host whenT. smilacina is scarce.     

Regulation ofHelicoverpa zea larval behavior by the parasitoidEucelatoria bryani

The parasitoidEucelatoria bryani Sabrosky regulates the larval behavior of its hostHelicoverpa zea (Boddie). Parasitized third, fourth and fifth instars burrow into the soil 0.7–3.4 days earlier than unparasitized larvae that normally enter the soil to pupate at the end of the fifth and final larval instar. Parasitized third instars molt once then burrow as fourth instars, one instar earlier than normal. WhenE. bryani pupariated on the soil surface in the field, none survived to the adult stage. However,E. bryani adults emerged from 49.2% of hosts that had burrowed into the soil. By accelerating the timing ofH. zea burrowing behavior and causing host larvae to enter the soil before death,E. bryani ensures its pupariation in an environment with improved protection against natural enemies and lethal temperatures.     

Ontogenetic shifts in the functional response and interference interactions of Rhyacophila dorsalis larvae (Trichoptera)

1. Ontogenetic shifts in predator behaviour can affect the assessment of food‐web structure and the development of predator–prey models. Therefore, it is important to establish if the functional response and interference interactions differ between life‐stages. These hypotheses were tested by (i) comparing the functional response of second, third, fourth and fifth larval instars of Rhyacophila dorsalis, using three stream tanks with one Rhyacophila larva per tank and one of 10 prey densities between 20 and 200 larvae of Chironomus sp.; (ii) using other experiments to assess interference within instars (two to five larvae of the same instar per tank), and between pairs of different instars (one, two or three larvae per instar; total predator densities of two, four or six larvae per tank). 2. The first hypothesis was supported. The number of prey eaten by each instar increased with prey density, the relationship being described by a type II model. The curvilinear response was stronger for fourth and fifth instars than for second and third instars. Mean handling time did not change significantly with prey density, and increased with decreasing instar number from 169 s for fifth instars to 200 s for second instars. Attack rate decreased progressively with decreasing instar number. Handling time varied considerably for each predator–prey encounter, but was normally distributed for each predator instar. Variations in attack rate and handling time were related to differences in activity between instars, fourth and fifth instars being more active and aggressive than second and third instars, and having a higher food intake. 3. The second hypothesis was partially supported. In the interference experiments between larvae of the same instar or different instars, mean handling time did not change significantly with increasing predator density, and attack rate did not change for second and third instars but decreased curvilinearly for fourth and fifth instars. Interference between some instars could not be studied because insufficient second instars were available at the same time as fourth and fifth instars, and most third instars were eaten by fourth and fifth instars in the experiments. Prey capture always decreased with decreasing attack rate. Therefore, interference reduced prey consumption in fourth and fifth instars, but not in second and third instars. The varying feeding responses of different instars should be taken into account when assessing their role in predator–prey relationships in the field.     

Host-age-dependent parasitism and reproductive success of Apanteles stantoni (Hymenoptera: Braconidae), a larval parasitoid of Diaphania indica (Lepidoptera: Pyralidae)

The relative suitability of five instars of Diaphania indica (Saunders) (Lepidoptera: Pyralidae) as a substrate for the development of a larval parasitoid, Apanteles stantoni Ashmead (Hymenoptera: Braconidae), was investigated. Maximum parasitism (22.25?±?1.21%) under laboratory conditions was observed in the early larval instars. The highest parasitoid emergence was recorded from the second (86.07?±?0.70%) and third (98.93?±?0.72%) instar larvae of D. indica, and that from the first larvae was 71.43?±?1.18%. The number of cocoons in each cluster, length and width of single cocoons, percentage emergence, sex ratio and adult longevity of A. stantoni collected from different instars of D. indica were also recorded. These results indicated that the life stage of the host when the parasitoid larvae complete their final instar is particularly important for their development. Therefore, considering the efficiency of parasitism and reproduction, the second-instar larvae of D. indica is the most suitable stage for mass rearing A. stantoni in the laboratory.     

Instar sizes,life cycles,and food habits of five Rhyacophila (Trichoptera: Rhyacophilidae) species from the Appalachian Mountains of South Carolina,U.S.A.

The larval head widths at each instar, life cycles, and food habits of late instars were determined for five species of Rhyacophila from two Appalachian mountain streams in South Carolina, U.S.A. Rhyacophila acutiloba Morse & Ross was univoltine with two cohorts, one emerging in the spring and another presumably emerging in early autumn. Rhyacophila fuscula (Walker), R. nigrita Banks, and R. carolina Banks were apparently multicohort, univoltine species with extended flight periods. Rhyacophila minor Banks was univoltine with a spring emergence. All species were predaceous and consumed mainly Plecoptera nymphs and Trichoptera larvae.     

Morphology and development of <Emphasis Type="Italic">Pteromalus cerealellae</Emphasis> (Ashmead) (Hymenoptera: Pteromalidae) on <Emphasis Type="Italic">Callosobruchus maculatus</Emphasis> (F.) (Coleoptera: Chrysomelidae)

Pteromalus cerealellae (Ashmead) (Hymenoptera: Pteromalidae) is an ectoparasitoid of several stored-product insect pests. Very little information has been published on its biology and development in host larvae, which typically are concealed within seeds. We documented the development of P. cerealellae within fourth instar larvae of its concealed host, Callosobruchus maculatus (F.) (Coleoptera: Chrysomelidae) infesting cowpea seeds. The preimaginal life stages of the parasitoid were characterized for the first time using morphological structures revealed by microscopic techniques including scanning electron microscopy. Pteromalus cerealellae produces hymenopteriform eggs and larvae. Eggs hatch into 13-segmented first instar larvae with peripneustic condition of spiracles. The larvae have simple, tusk-like mandibles, whereas the mandibles of the pupae and the adults are of the conventional toothed types. Using statistical analyses of the sizes of the larval mandibles and head capsules in conjunction with reliable characters such as the number of exuviae on the body of parasitoid larvae, cuticular folding, and excretion of the meconium, we recorded four larval instars for P. cerealellae. The data showed significant positive correlations between larval mandible lengths and widths of larval head capsules, as well as between mandible lengths and larval instars, suggesting that mandible length is a good predictor of the number of instars in P. cerealellae. Developmental time from egg to adult emergence was ∼12 d for females and ∼11 days for males at 30 ± 1°C, 70 ± 5% r.h. and 12L:12D photoperiod.     

Identification of instars and species in some larval Polypedilum (Polypedilum) (Diptera: Chironomidae)

Instars II and III of Polypedilum aviceps Townes, Polypedilum convictum (Walker), and Polypedilum illinoense (Malloch) can be identified to species by associating them with instar IV because key taxonomic characters remain relatively unchanged from instar to instar. Instars I cannot be identified to species or genus unless they are associated with older, identifiable larvae reared from the same egg masses. No single character evaluated on slide material can be used to clearly separate instars in all three species. Larvae of P. aviceps can be separated into instars based on any four of seven characters; P. convictum by either of two characters; and, P. illinoense by a combination of two characters. Changes in structures of instars II, III, and IV are described for all three species. Growth ratios for some structures are compared and discussed with regard to Dyar's Rule.     

Selection of healthy and nuclear polyhedrosis virus infectedAnticarsia gemmatalis [Lep.: Noctuidae] as prey by nymphalNabis roseipennis [Hemiptera: Nabidae] in laboratory and on soybean

Nabis roseipennis Reuter nymphs demonstrated a preference for nuclear polyhedrosis virus (NPV) — infected over healthyAnticarsia gemmatalis Hübner larvae when offered a choice of larval prey in Petri dishes and on soybean. In Petri dishes, small (second-third instar) and large (fifth-sixth instar) nymphs attacked a significantly greater number of diseased than healthy larvae at all larval instars tested (first-fifth instars) and exposure periods (2, 5 and 24 h), except that at 2 h the number of 1^st and 3^rd instar larvae attacked by large nymphs did not differ significantly (P≤0.05). Nabis roseipennis caged with larvae on individual soybean plants in the greenhouse resulted in a generally low percentage of attack by small and large nymphs after 2 days, ranging from 5.6 to 36.7%. As in the Petri dishes, the nabids showed a significant preference for diseased larvae over healthy larvae attacked for all nabid and larval sizes on soybean, with the percentage of diseased larvae attacked ranging from 28.0 to 65.4% (P≤0.05). This preference for diseased larvae on soybean as well as in Petri dishes demonstrates that the preference was not due to the close proximity in which the host and prey were found in the Petri dishes. The preference for diseased larvae may be due to a reduction in a defensive response in late stages of disease. This material is based upon work supported in part by USDA Grant No. 83-CRCR-1-1212.