Sphinx Moth Pollinators for the Endangered Western Prairie Fringed Orchid Platanthera praeclara In Manitoba, Canada (2004)

By Richard Westwood and Christie L. Borkowsky

Christie Borkowsky is a graduate student in the Department of Entomology at the University of Manitoba in Winnipeg. Prairie technician for the Manitoba Tall Grass Prairie Preserve, Christie conducts floral and faunal surveys for rare and endangered species. Christie’s interests lie the pollinators of prairie plants, in particular, the hawkmoths that pollinate the western prairie fringed orchid.

Richard Westwood is an Associate Professor, Dept. of Biology and Dept. of Environmental Studies at the University of Winnipeg, and member of the Center for Forest Interdisciplinary Research University of Winnipeg. His research interests include: Testing an improved insect pollination strategy for the endangered Western Prairie Fringed Orchid in the Manitoba Tall Grass Prairie Preserve at Vita, Manitoba. Funding supporters: Manitoba Sustainable Development Fund, Manitoba Conservation - Conservation Data Centre, World Wildlife Fund and the Manitoba Orchid Society. Graduate student - Christie Borowsky (MSc) and Karen Budnick undergraduate Honours Thesis student.

References

Abstract

The western prairie fringed orchid, Platanthera proeclara (Sheviak & Bowles), is an endangered species in North America. In Manitoba orchids produce lower numbers of seed capsules than more southern populations. Exploration of the pollination biology for P. praeclara is critical to preserve this endangered species. This study identified the pollinators of P. praeclara using cone and malaise traps and tested the effectiveness of marking pollinators with traces of Day-Glow® orange marker powder. Although Lepidoptera were numerous in orchid plots during daily observation periods, including day flying Sphingidae, none were pollinators for P. praeclara. Among the 5856 insects from 49 families captured over 45 trapping days, six sphinx moths, two specimens of Hyles gallii (Rottenburg) and four specimens of Sphinx drupiferarum J.E. Smith (Sphingidae), were found with two or more orchid pollinia attached to their eyes and were confirmed as pollinators of P. praeclara. S. drupiferar1lrn is uncommon in southern Manitoba and H. gallii appears to be a less efficient pollinator than S. drupiferarum. Proboscis length, eye width and flight period may influence the efficiency of biotic pollination of the orchid.

Additional key words

Sphingidae, pollinia, pollination, P. praeclora, S. drnpfirarnm, H. gallii.

The western prairie fringed orchid, Platanthera pracciara (Sheviak & Bowles 1986), occurs in areas of remnant tall grass prairie in southeastern Manitoba. Prior to its discovery in the mid 1980’s near Tolsoi, Manitoba, P. praeclara was not known to exist in Canada (Johnson 1985, 1991). The Manitoba population, at its maximum has consisted of approximately 21,000 individual plants (although the population may fluctuate widely from year to year) and it is the largest of four metapopulations (over 300 plants) in North America. The remaining populations occur in North Dakota, South Dakota, Minnesota, Kansas, Nebraska, and Iowa (Sheviak & Bowles 1986, Bjugstad & Fortune 1989, Bray & Wilson 1992, Bjugstad-Porter 1993, Pleasants & Moe 1993, Davis 1994, Sieg & King 1995, U.S. Fish & Wildlife Service 1996, Hof et al. 1999).

The orchid is protected under Manitoba’s Endangered Species Act and has been placed on the endangered species list in both Canada and United States (Collicutt 1993, Davis 1995). In Manitoba, P. praeclara often has low fruit set and subsequent seed production (Borkowsky 1998). Although the western prairie fringed orchid will produce a vegetative form, there is little evidence that vegetative reproduction occurs in P. praeclara (Bowles 1983, Sather 1991, Sieg & King 1995, Hof et al. 1999). A plant may also go dormant, as is typical in other species of orchids (Nilsson 1992). Therefore, the recruitment of new plants is dependant primarily upon successful pollination and subsequent seen production. In surveys of over 1000 plants Borkowsky (1998) found that only 2.1% of orchid stems produced one or more seed capsules annually in Manitoba between 1994 and 1998. In more southern orchid populations the percent of stems that produce seed capsules can range up to 49% (Bowles 1983, Sheviak & Bowles 1986, Cuthrell 1994). The authors have hypothesized that reduced seedpod production may be linked to low pollination success in Manitoba. The level of pollination success for P. praeclara and the identity of tile orchid pollinators are unknown in Manitoba.

Pollination is the process in which pollen grains are transferred to the stigma, which is followed by fertilization of tile ovules and development of seeds (Proctor et al. 1996). Many orchids require a biotic organism (a pollination agent or pollen vector) to transport the pollen to the stigma (van der Pijl & Dodson 1966, Faegri & van der Fiji 1979). In P. pracclara the most striking visual characteristic of the flower is a large, deeply fringed, tri-lobed lower lip and a long, slender spur containing nectar that suspends backward from the flower (Figs. 1, 2). The nectar spur may be 36–55 mm long with a maximum diameter of 2.7 ± 0.5 mm (Sheviak & Bowles 1986). Orchids grow to 38–85 cm tall (Sheviak & I3owles 1986) with the determinant inflorescence containing 7 to 12 flowers (Sheviak & Bowies 1986, Pleasants 1993). The pollinium (Fig. 1) is a specialized structure that consists of pollen, a column and viscidium (Nilsson 1992). In P. praeclara the minute grains of pollen are arranged into subunits called massulae (Pleasants & Moe 1993). These subunits form a hi-lobed mass that is attached to the column, which is secured to the viscidium (the entire structure is termed the pollinium). The pollinium is sheathed, with the exception of the viscidium, which is exposed and adapted to cement itself to the pollinator (Bowles 1983). Each flower has one pollinium located on either side of the stiginatic surface. This allows for a 6 to 7 mm separation between each viscidium (Sheviak & Bowles 1986). The opening to the nectar spur is located immediately below the stigmatic surface.

The small opening to the nectar spur restricts the position of a potential nectar seeking insect, and increases the likelihood that one or both viscidia will come into contact with a likely pollinator. The pollinium is pulled from the sheath when a pollinator contacts the viscidium and then withdraws from the flower (Darwin 1904, Dressler 1981, 1993).

Orchids also produce a scent or fragrance which may be weak or absent in the daytime, increases in intensity at dusk and may remain strong until sunrise depending on the age and condition of the flower (Sheviak & Bowles 1986). The relationship between the attractiveness of the fragrance to pollinators and the intensity of the scent is unknown in P. praeclara. To be identified as a pollinator and to rule out indiscriminate visitors to the flower, the organism must make regular visits to the flowers during its lifetime arid effectively deposit the pollen on the stigma (Faegri & van der Pijl 1979). The flower-pollinator relationship may also be controlled by pollinator behavior, mouthpart morphology or taxonomy and by flower morphology (Faegri & van der Pijl 1979, Wyatt 1983). Wyatt (1983) identified nine forms of biotic pollination in orchids including sphingophily (Sphingidae—hawkmoths), phalaenophiiy (small moths), psycophily (butterflies), melittophily (bees), myophily (syrphid and bee flies), sapromyophily (carrion and dung flies), cantharophily (beetles), ornithophily (birds) and chiropterophily (bats).

The potential pollinators for P. praeclara in Manitoba are not known. Although the literature suggests that night-flying members of the Sphingidae may be key pollinators of P. praeclara, other members of this orchid genus and several closely related genera may be pollinated by butterflies, other moth species, certain Diptera, Coleoptera or Hynienoptera (van der Pijl & Dodson 1966, Patt et al. 1989, Robertson & Wyatt 1990, Bowles et al. 1992, Larson 1992).

The floral characteristics of P. praeclara indicate a moth pollination method (Faegri & van der PijI 1979, Sheviak & Bowles 1986, Luyt & Johnson 2001), most likely sphingophily or phalaenophily. P. praeclara has no developed landing platform on the flowers, therefore hovering pollinators are favored and butterflies may be excluded (van der Pijl & Dodson 1966, Dressier 1981). To retrieve nectar, the insect must correctly align itself with the flower as it inserts its proboscis, which increases the likelihood that the pollinator will contact one or both of the viscidia which will adhere to the proboscis or eyes of the insect (Sheviak & Bowles 1986). Considering the length of the nectar spur of P. praeclara arid position of the viscidia, the list of potential pollinators is further reduced to Lepidoptera with a very long proboscis (e.g., moths belonging to the Sphingidae or hawkinoth family). Few, if any, observations of pollination of fringed orchids by hawkmoths have been made in the field (Bowles 1983, Sheviak & Bowles 1986, Pleasants & Moe 1993). Based on proboscis length, Sheviak & Bowles (1986) suggested that the following species of moths could be potential pollinators of P. praeclarn in the United States: Eunwrpha achenwn (Drury), Hyles lineata (F.), Sphinx drupferamm J.E. Smith and Sphinx kalmiac J.E. Smith. Cuthreil (1994) collected two specimens, one each of Sphinx drupferarum and Enmnorpha acheinon with viscidia from P. praeclara attached to the head, from a light trap adjacent to a field of orchids in the United States.

To better understand the pollination biology of P. praeclara in Manitoba diurnally active Lepidoptera including day flying Sphingidae, nocturnally active Lepidoptera and other insect Orders were surveyed as potential pollinators of P. praeclara. Several types of passive traps were tested to capture P. praeclara pollinators, and a method to mark individual pollinators was investigated.

Study Site And Methods

Five field experiments were established between 1997 and 1999 to determine the identity off P. praeclara pollinators. Three study plots were located in the Manitoba Tall Grass Prairie Preserve, near Tolsoi, Manitoba (49° 05&prime N, 96° 49&prime W) and two additional moth light trapping sites were located near Lonesand, Manitoba (approximately 21 km east of the Preserve) arid near Grunthal, Manitoba (approximately 26 km northwest). The climate is a boreal continental regime with mean temperatures of 19.6°C and –18.8°C for July and January, respectively. Fifty-five percent of the annual precipitation (mean 579.1 mm) falls during the period of May through August (Canadian Climate Program 1993). Drainage in much of the prairie is poor with soil composed of lacustrine parent material, sandy loam to clay loam upper horizons and a thin organic surface layer. Stones, rocks and occasionally boulders are also scattered across the prairie (Canada Soil Inventory 1989, Moore & Fortney 1994).

The Tall Grass Prairie Preserve has an abundance of grasses, including Big bluestem (Andropogon gerardii Vitman) and Little bluestem (Schizachyrium scopariurn (Michx.) Nash) (Gramineae) and is interspersed with bluffs of various deciduous tree species, including willow (Salix spp.) and poplar (Populus spp.) (Henne & Diehl 2002).

Three 50×50 m plots were established in fields were orchids were numerous. Each plot contained a minimum of 25 orchids. Plots were separated by a distance of 500 to 1000 m. To establish that there were no day flying insects pollinating the orchids, 15 individual orchids were tagged in each of the three plots (total of 45 plants) in 1997 and 1998. Each orchid was visually observed for 15 minutes on at least three separate days during the flowering period in 1997 and 1998 between 1100 h to 1500 h to determine if insects contacted plants and were responsible for removal of one or more of the pollinium from the flowers, either by accident or for the purpose of obtaining nectar or pollen. The number and identity of potential pollinators (Order, Family and Genus/species if known) landing or crawling on the orchids was recorded.

Inverted cone insect traps were used to sample potential pollinators from individual orchids. Cone traps were constructed from light gauge steel tubing and wire and covered in fine mesh screen. Cone traps were approximately 40 cm in length and 25 cm in diameter. An inverted wire mesh cone was placed at the base of the trap (much like a minnow trap) and a hinged trap door placed on the top of the trap. Three legs attached to the bottom of trap allowed it to be placed over an individual orchid. The legs were of sufficient length that they could be pushed into the ground around the orchid, allowing the mouth of the inverted cone to be suspended approximately 5 to 10 cm above the terminal flower of the orchid. Traps were emptied daily between 0630 h and 0830 h and again between 1900 h and 2030 h. Captured insects were identified and examined for presence of orchid pollinia. One cone trap was placed in each of the three plots and rotated between plants during the flowering period. Cone traps were placed over orchids that had the most flowers open with intact pollinia available. Traps remained in place over an individual orchid for 24–48 hours and were placed in the plots between 1 July and 15 July in 1997, 1998 and 1998.

Malaise insect traps (Bioquip® Equipment Specialties) were to used to sample for potential pollinators over a groups of orchids (generally five or more individual plants). Each trap was approximately 2 in in height with a glass container at the top to collect trapped insects. A piece of Vapona® insecticide strip (0.2 cm2) was placed at the end of the collecting cylinder to kill captured insects. One malaise insect trap was placed within each of the three plots and rotated among groups of orchids during the flowering period. Traps were emptied daily between 0630 Ii and 0830 h and again between 1900 Ii and 2030 Ii. Insects were identified and examined for presence of orchid pollinia. Traps were placed in the field between 1 July and 15 July in 1997, 1998 and 1999. Traps were rotated within the plot to new groups of orchids every three to five days.

In 1999, five orchids in each plot under a malaise trap or one orchid under a cone trap were chosen to test the effectiveness of Day-Glow® orange marker dye powder in identifying potential pollinators. A small amount of Day-Glow® orange marker dye powder was applied with a extra fine nylon brush to the centre of each flower of the orchids chosen. Insects captured in the traps were identified and examined for presence of pollinia removed from the orchids arid also examined in a dark room under an ultra-violet light for the presence of the Day-Glow® powder. Insects with marker particles adhering to their bodies were considered to have come in contact with the treated flowers.

To determine moth flight periods, two Wards all- weather insect traps® (Wards Natural Science) were placed approximately 1 km to the west of the orchid plots and 5 km to the south of the plots. Two additional Ward’s all-weather insect traps® were located at Lonesand and Grunthal, Manitoba to augment moth captures at the Tall Grass Prairie Preserve. Traps were operated from 1 May to 31 August, 1997–1999. Traps had an eight watt ultra-violet fluorescent bulb as an attractant. The traps were used to survey nocturnal Lepidoptera and determine the flight periods of potential orchid pollinators. Flight periods were considered to include the period between date of first and last capture in each year. Traps in the Tall Grass Prairie Preserve were placed in open areas surrounded by mature trees and were not visible from orchid plots. Lepiodoptera captured in traps were sorted, pinned and identified to species.

The flowering period for orchids in plots was recorded during the study. Flowering period was defined as the period between the appearance of the first flower and last flower in a plot and peak flowering date was defined as the date when the most flowers were fully open in a plot. The mean overlap in days between moth species flight period and orchid flowering period for all years was calculated. Mean overlap flight period data was square root transformed to satisfy assumptions of normality and heterogeneity of variance for analysis of variance (ANOVA) (SPSS Inc. 1999). Where ANOVA was significant a least-significant difference (LSD) test was used to determine the differences between means (α = 0.05) because of its consistency (Saville 1990). To determine if potential pollinators would fit pollinia distance separation requirements and nectar spur depth requirements of P. praeclara, measurements were made on a minimum of five individuals from all sphinx moth species collected in the various trap types. Pinned moths were softened in a relaxing chamber and the length of the proboscis, the distance between the outer margins of the compound eyes and distance between the inner margins of compound eyes where measured in mm under a dissecting microscope. For data analysis the proboscis length was square root transformed and distance between outer eye edges and distance between inner eye edges were log transformed to satisfy assumptions of normality and heterogeneity of variance. Morphological measurements between species were subject to ANOVA and where ANOVA was significant differences between means were identified using a LSD test. To examine the relationship between proboscis length and ability of pollinators to retrieve nectar, the length of nectar spurs and the depth of nectar within the spur (distance from distal end of spur to top of nectar line) was measured for orchids in each of the three plots.

TABLE 1. Number and identity of insect Orders and Families collected in cone and malaise traps 1997–1999.

Identification of insects was based upon Hodges (1971), Rockburne and Lafontaine (1976), Morris (1980), hodges et al. (1983), Covell (1964), Borror et al. (1989), Klassen et al. (1989), Layberry et al. (1998) and Handfield (1999).
Order Family/Subfamily n
Ephemeroptera Baetidae 1
Heptageniidae 1
Odonata Coenagrionidae 3
Orthoptera Acrididac 5
Plecoptera Perlidac 1
Hemiptera Mindac 21
Rediiviidae 2
Pentatomidae 2
Homoptera Cercopidae 5
Cicadellidae 13
Neuroptera Mantispidae 2
Chrysopidae 9
Coleoptera Carabidac 8
Scarabaeidae 5
Elateridae 2
Lampyridae 12
Cleridae 5
Coccinecllidae 8
Tenebrionidae 1
Chrysomelidae 5
Curculionidae 3
Mecoptera Panorpidae 2
Trichoptera Limnephilidae 6
Diptera Tipulidac 26
Culieidae 55
Chironomidae 24
Simuliidae 15
Tabanidae 4450
Syrphidae 45
Muscidae 546
Calliphoridae 228
Sarcophagidae 11
Lepidoptera Pyralidae 14
Pterophoridae 2
Tortricidae 34
Gelechiidae 40
Geometridae 18
Arctiidae 3
Noctuidae 34
Lasiocampidae 15
Sphingidae 6
Hesperiidae 10
Nymphalidae 15
Satyridae 2
Hymenoptera Lcbneumonidae 23
Vespidae 2
Sphecidae 3
Megachilidae 2
Apidae 19

Results

Observational survey

In 1997 and 1998 orchids were observed between 1100 h and 1500 h for approximately 70 hours. Although numerous insects (many belonging to pollinating families or genera) frequented the plots during the daily observation periods no individuals were observed to seek nectar or pollen from the orchids, to use the flowers as a resting platform or to sun themselves on the orchids. Individuals of the sphinx moths Hemaris thysbe (Fabricus) (16 individuals) and Hemaris diffins (Boisduval) (28 individuals) were observed in the orchid plots during the observation periods but they were not attracted to orchids, despite repeatedly visiting nearby flowering herbs for nectar.

Trapping experiments

Between 1997 and 1999 the cone and malaise traps caught 5856 individual insects from 49 families over 45 trapping days (Table 1). The only insects found to have pollinia attached to their bodies belonged to the family Sphingidae (Table 2). Six sphingid moths, two specimens of the Bedstraw hawkmoth, Hyles gallii (Rottenburg), and four specimens of the Wild cherry sphinx, S. drupiferarum were collected with two or more pollinia attached to the eyes (Figs. 3, 4). H. gallii had 2 pollinia per moth, while S. drupiferarum had 3 to 11 pollinia per moth.

Marking experiment

In 1999 one sphinx moth, H. gallii, was collected from a cone trap with traces of Day-Glow® orange marker dye powder on both eyes and the proboscis. Both pollinia attached to the eyes also had traces of powder, primarily on the massulae and viscidium.

Flight period

The flight periods for sphingid species found in the vicinity of the Tall Grass Prairie Preserve were based on the catches from the four black light traps during the period of 1997 to 1999 (Table 3). Flowering dates for P. praeclara are also shown in Table 3. Generally the moths collected were most abundant in the first several weeks of the orchid flowering period. The flight periods of S. drupfera rum and H. gallii overlapped with orchid flowering by 34.6% and 45.3% respectively.

Morphological measurements

The proboscis length, width across the eyes and distance between the inner eye margins were measured for the 15 species of Sphingidae collected in the vicinity of the Tall Grass Prairie Preserve (Table 4). The mean orchid nectar spur length was 45.27mm (n = 1016, SE = 0.134) and the mean depth of nectar within the spur was 12.44 mm (n = 1016, SE 0.201).

Table 2. Summary of six Sphingidae collected from cone and malaise traps with attached pollinina 1997-1999.
Date Species Type # of pollinia Pollinia location Marker powder
11 July 1997 Hyles gallii Cone 2 Head n/a
11 July 1997 Sphinx drupiferarum Malaise 5 Head n/a
15 July 1998 Sphinx drupiferum Malaise 11 Head n/a
15 July 1998 Sphinx drupiferum Malaise 3 Head n/a
20 July 1998 Sphinx drupiferum Cone 7 Head n/a
6 July 1999 Hyles gallii Cone 2 Head present

Discussion

The results confirm observations by Faegri and van der Piji (1979), Sheviak and Bowles (1986), Cuthrell (1994) and Luyt and Johnson (2001) that P. praeclara is pollinated by nocturnal Lepidoptera confined to the family Sphingidae. It appears that there are no diurnally active insects that seek nectar or pollen from P. praeclara. The presence of pollinia on one H. gallii specimen and four S. drupiferarum specimens and the marker powder and pollinia on another specimen of H. gallii confirm these species as pollinating agents ofF. praeclara in Manitoba. Despite the larger size of be malaise trap covering more orchids, there was little difference in the capture rates of moths with pollinia between trap types. The Day-Glow® orange marker dye powder successfully marked one pollinator.

Pollinia were attached to the center of each eye on both moth species. S. drupiferarum had more pollinia per specimen. The distance between the outer edges of the eyes was greater in S. drupiferarum than H. gallii (Table 4), but it is unknown if this small difference in eye separation (0.43 mm) would affect the attachment of pollinia. Perhaps S. drupiferarum is more aggressive at retrieving nectar and presses harder into the centre of the orchid thus capturing more pollinia. Alternately H. gallii may visit the orchids less frequently resulting in fewer opportunities to remove pollinia, or be discouraged from visiting orchids once several pollinia have been attached to the eyes. H. gallii is more numerous in the vicinity of the orchids than S. drupiferarum, thus it seems likely that they are less attracted to the orchids.

Date 1997 1998 1999 n¹ % Overlap of flight period & flowering period days ± SE Peak orchid flowering date 2 July 4 July 7 July Sphingidae-flight period dates Ceratomia undulosa Harria 12 June-12 July 7 June-18 July 9 June-18 July 14 92.0 ± 4.9c² Sphinx cheris (Hubner) 31 May-2 July 11 July Absent 6 14.1 ± 13.9a Sphinx kalmiae J. E. Smith 1-26 July 5-24 June 10 June- 18 July 8 58.0 ± 23.2abc Sphinx luscitiosa Clemens Absent 20 May-30 June 19 June-11 July 7 34.6 ± 19.9abc Sphinx drupiferarum J. E. Smith 11 July 1 June-25 July 24 June 5 34.6 ± 32.7abc Smerinthus cerisyi Kirby 28 May-27 June 21 May-29 July 16 May-25 July 58 73.7 ± 26.3bc Smerinthus jamaicensis (Drury) 25 May-30 June 1 June-10 July 29 May18 July 123 65.3 ± 18.2bc Poanes exaecatus (J. E. Smith) 3 July-20 July 11 June-15 July 5 June-20 July 37 94.6 ± 9.2c Poanes myops (J. E. Smith) 11-28 June 9 June-4 July 1-27 June 34 29.3 ± 16.9abc Cressonia juglandis (J. E. Smith) Absent 8-30 June Absent 5 11.7 ± 11.6a Pachysphinx modesta (Harris) 27 May-28 June 11 June-5 Aug 9 June-9 July 61 63.0 ± 20.8bc Hemaris thysbe (Fabricus)³ 31 May-25 June 5-28 June 4 June-2 July 16 25.3 ± 6.9abc Hemaris diffinis (Boisduval)4 21 May-2 July 10 June-10 July 1 June-13 July 28 62.0 ± 10.4bc Darpsa myron (Cramer) Absent 15 May-6 July Absent 5 18.0 ± 16.0a Hyles gallii (Rottenburg) 15 June-31 July 31 May-29 June 6 July 18 45.3 ± 28.5abc p = 0.040 F14.44 = 2.86
TABLE 3. Flight period dates, orchid flowering periods, number of individual sphinx moths collected in black light traps and percent overlap between flight and flowering dates in the vicinity of the Tall Grass Prairie Preserve and for orchids 1997–1999.

Cuthrell (1994) described two specimens of S. drupiferarum (one caught by light trap and one a museum specimen) and one specimen (Drury) (collected by light trap) with P. praeclara pollinia attached to the eyes. The present study adds H. gallii to this list of pollinators while of E. achemon does not occur in Manitoba.

The mean proboscis lengths for S. drupiferarum and H. gallii captured in this study were 38.40 mm and 33.50 mm, respectively. Subtracting the mean nectar depth from total nectar spur length provides a distance of 32.83 mm, thus it appears that a proboscis length of 30–35 mm is required to reach the nectar in P. praeclara. Based on proboscis length, Sphinx chersis (Hubner) and S. kalmiae may also be potential pollinators in Manitoba (Table 4), with no other species having a mean proboscis length greater than 23 mm. The separation of inner and outer eye margins between species was less distinct with some species having a short proboscis yet still having similar eye positioning and separation (Table 4).

Orchid flowering periods and moth flight periods are restricted in terms of overlap in Manitoba. The overlap in more southern areas of the range of P. praeclara may be greater and therefore higher levels of pollination may occur resulting in more seed capsules per plant. S. drupiferarum is uncommon in southern Manitoba and the most numerous sphingid species in the Tall Grass Prairie Preserve do not possess a proboscis of sufficient length to take nectar from P. pracciara. H. gallii populations fluctuate greatly from year to year in southern Manitoba, often with very few individuals appearing in some years. Therefore, low pollinator populations may be restricting seed production in some years for orchids in Manitoba.

TABLE 4. Length of proboscis, the distance between the outer margins of the compound eyes and distance between the inner margins of compound eyes for Sphinx moths collected in the vicinity of the Tall Grass Prairie Preserve, Manitoba.
N¹ Mean proboscis length (mm ±SE) Mean distance between outer eye margins (mm ±SE) Mean distance between inner eye margins (mm ±SE)
1 Ceratomia undulosa Harris 7 9.14 ±0.76c² 5.05 ±0.05h 2.04 ±0.02d
2 Sphinx cheris (Hubner) 5 40.32 ±0.48h 5.89 ±0.03k 2.40 ± 0.11ef
3 Sphinx kalmiae J. E. Smith 5 33.64 ±1.31g 5.31 ±0.17bi 2.05 ±0.18d
4 Sphinx luscitiosa Clemens 5 22.56 ±0.17f 4.46 ±0.05ef 2.34 ±0.01bc
5 Sphinx drupiferarum J.E. Smith 5 38.40 ±0.97h 5.59 ±0.06i 2.34 ±0.02e
6 Smerinthus cerisyi Kirby 8 2.21 ±0.06a 4.76 ±0.13g 1.88 ±0.3c
7 Smerinthus jamaicensis (Dury) 8 1.73 ±0.75a 3.85 ±0.03c 1.70 ±0.02b
8 Poanes excaecatus (J.E. Smith) 7 2.89 ±0.03b 4.63 ±0.05f 2.09 ±0.03d
9 Poanes myops (J.E. Smith) 7 1.87 ±0.04a 4.07 ±0.03d 1.83 ±0.06bc
10 Cressonia juglandis (J.E. Smith) 5 2.10 ±0.05aa 3.31 ±0.08a 1.37 ±0.53a
11 Pachysphinx modesta (Harris) 8 1.95 ±0.17a 5.73 ±0.07j 2.53 ±0.02f
12 Hemaris thysbe (Fabricus) 8 12.75 ±0.52d 3.88 ±0.02c 1.88 ±0.08c
13 Hemaris diffinis (Boisduval) 8 9.48 ±0.20c 3.58 ±0.06b 1.89 ±0.03c
14 Darapsa myron (Cramer) 5 17.68 ±0.18e 4.28 ±0.09e 2.12 ±0.04d
15 Hyles gallii (Rottenburg) 8 33.50 ±0.96g 4.88 ±0.03gh 2.44 ±0.02ef
p=0.001 F14, 98=37.1 p=0.001 F14, 98=35.2 p=0.001 F14, 98=40.1

The Tall Grass Prairie Preserve is surrounded by agricultural lands, a mixture of intensively farmed grains and oilseeds and livestock production. The larval host plants for S. drupiferarum can very regionally and include Malus spp., Prunsus spp. and lilac, Syringa vulgaris L. (Hodges et al. 1983). These plants are present in the vicinity but their distribution is patchy. It is unknown whether insecticide usage (Suzán et al. 1994) on adjacent farmlands may restrict the population of larval forms of S. drupiferarum and H. gallii or if weed control may restrict access to host plants. S. drupiferarum may be a more efficient pollinator than H. gallii, but the number of S. dupiferaruut adults frequenting the orchid fields was low in this study. There may be considerable competition between nectar sources given the amount of intensively managed lands adjacent to the Tall Grass Prairie Preserve, which would further reduce the time spent by moths pollinating orchids. Strong light sources from farms surrounding the orchid fields may also attract months, preventing them from visiting orchids. S. drupiferarum occupies the northernmost extension of its range in Manitoba, which may also contribute to sporadic population occurrence and lower pollination rates of P. praeclara.

Baker (1961) stated that sphingophilous flowers often have a low frequency of pollinator visitation and compensate by producing numerous seeds. Further research is required to determine if other sphinx moth species are pollinating orchids and if the current level of seedpod production is abnormally low or normal for P. praeclara in Manitoba.

Acknowledgments

This project was funded by the following agencies: Manitoba Conservation, the University of Winnipeg, Manitoba Orchid Society, Manitoba Sustainable Development Fund and the World Wildlife Fund. The authors would like to acknowledge Jason Creenall, Peggy Westhorpe, Lorne Heska, Liz Reimer amsd Tracy Loewen in providing assistance and support to this study and Dr. C. Waisg and Rachel Boone for reviewing the manuscript. Photographs by R. Westwood.

Westwood, Richard A. and Borkowsky, Christie L. "Sphinx Moth Pollinators for the Endangered Western Prairie Fringed Orchid Platanthera praeclara in Manitoba, Canada". Previously published in Journal of the Lepidopterists’ Society 58:1 (2004). 13-20.

Received for publication 30 January 2003; revised and accepted 29 July 2003.