ArticlePDF Available

European Skipper Butterfly (Thymelicus lineola) Associated with Reduced Seed Development of Showy Lady’s-slipper Orchid (Cypripedium reginae)

Authors:

Abstract

It has been suggested that European Skipper butterflies (Thymelicus lineola) trapped in the lips of the Showy Lady’s-slipper orchid (Cypripedium reginae) may interfere with pollination. This could occur through blockage of the pollinator pathway, facilitation of pollinator escape without pollination, and/or disturbance of the normal pollinators. A large population of the orchid at an Ottawa Valley site provided an opportunity to test the interference hypothesis. The number of trapped skippers was compared in 475 post-blooming flowers with regard to capsule development and thus seed development. The presence of any skippers within flowers was associated with reduced capsule development (P = 0.0075), and the probability of capsule development was found to decrease with increasing numbers of skippers (P = 0.0271). The extent of a negative effect will depend on the abundance of the butterflies and the coincidence of flowering time and other factors. Counts of skippers trapped in flowers were found to follow closely a negative binomial distribution (P = 0.8656).
Introduction
The potential for pollination interference by Euro-
pean Skippers (Thymelicus lineola Ochsenheimer)
caught in the lips of Showy Lady’s-slippers (Cypri-
pedium reginae Walter), has been alluded to by a num-
ber of authors (e.g., Catling 1974; Barrows 1983; Vogt
1990) and repeated with additional references by Argue
(2012). The usual pollinators are leaf-cutter bees (Mega -
chile spp.) and syrphid flies (Syrphus spp.), which enter
the slipper-shaped lip but cannot exit the same way as
a result of its inflexed margin (Argue 2012). They first
have to pass the stigma, where they deposit any pollen
they are carrying and then exit by one of two openings
at the base of the lip where they receive a new load of
pollen on the thorax.
The flowers attract insects by a combination of odour
and colour (white and pink, with pink nectar guides),
but there is no nectar reward, as with other species in
the genus, which are also deceptive (e.g., Argue 2012).
Escape of skipper butterflies entering the lip is prevent-
ed by the inflexed margin; they are also unable to escape
through the basal openings, as they are too delicate and
high, with their wings closed over the top of the body.
Entrapment of skippers may lead to reduced pollina-
tion by disturbance of the normal pollinators; possible
mechanisms include blocking the pathway and/or facil-
itating pollinator escape without pollination. The effect
may be substantial: Vogt (1990) found skippers in about
a third of the flowers at his Vermont site, Barrows
(1983) and Catling (1974) reported skippers in about
half of 100 flowers at several Ontario sites.
The European Skipper was first introduced to North
America in 1910 (Hall et al. 2014), so it has likely not
evolved in the presence of the deceptive orchid flower,
which might have led to avoidance. The European
Skipper is now one of the most abundant butterflies
in North America.
At Purdon Fen (44.99260°N, 76.54596°W) near
Lanark in eastern Ontario, we had an unusual oppor-
tunity to evaluate the impact of the European Skipper
butterfly on capsule (and seed) development of the
Showy Lady’s-slipper orchid. This orchid has an exten-
sive distribution in northeastern North America (Luer
1975), but there are few, if any, locations where it is as
abundant as at the Purdon site where the total popu-
lation of stems within 1.2 ha has fluctuated between
15 735 in 1985 (Mosquin 1986) and 7367 in 2015 (Ross
Fergusson, personal communication). In 2015, the skip-
pers,which generally frequent open country, were much
more abundant than usual in the wooded area surround-
ing the fen and had entered the wooded fen where the
orchids occur.
This area is a semi-open EasternWhite Cedar (Thuja
occidentalis L.) woodland with an understory dominat-
ed by Tussock Sedge (Carex stricta Lam.). During a
visit to the site on 17 June 2015, we saw at least 300
skippers in 30 minutes, flying within the orchid colony,
and some were trapped in the flowers (Figure 1). By
European Skipper Butterfly (Thymelicus lineola) Associated with
Reduced Seed Development of Showy Lady’s-slipper Orchid
(Cypripedium reginae)
PETER W. HALL1, 5, PAUL M. CATLING2, PAUL L. MOSQUIN3, and TEDMOSQUIN4
124 Wendover Avenue, Ottawa, Ontario K1S 4Z7 Canada
2170 Sanford Avenue, Ottawa, Ontario K2C 0E9 Canada
3Research Triangle International, 3040 East Cornwallis Road, P.O. Box 12194, Raleigh, North Carolina 27709-2194 USA
43944 McDonalds Corners Road, Balderson, Ontario K0G 1A0 Canada
5Corresponding author: halljp@rogers.com
Hall, Peter W., Paul M. Catling, Paul L. Mosquin, and Ted Mosquin. 2017. European Skipper butterfly (Thymelicus lineola)
associated with reduced seed development of Showy Lady’s-slipper orchid (Cypripedium reginae). Canadian Field-
Naturalist 131(1): 63–68. https://doi.org/10.22621/cfn.v131i1.1952
It has been suggested that European Skipper butterflies (Thymelicus lineola) trapped in the lips of the Showy Lady’s-slipper
orchid (Cypripedium reginae) may interfere with pollination. This could occur through blockage of the pollinator pathway,
facilitation of pollinator escape without pollination, and/or disturbance of the normal pollinators. A large population of the
orchid at an Ottawa Valley site provided an opportunity to test the interference hypothesis. The number of trapped skippers
was compared in 475 post-blooming flowers with regard to capsule development and thus seed development. The presence of
any skippers within flowers was associated with reduced capsule development (P= 0.0075), and the probability of capsule
development was found to decrease with increasing numbers of skippers (P= 0.0271). The extent of a negative effect will
depend on the abundance of the butterflies and the coincidence of flowering time and other factors. Counts of skippers trapped
in flowers were found to follow closely a negative binomial distribution (P= 0.8656).
Key Words: Cypripedium reginae;Thymelicus lineola; Showy Lady’s-slipper; orchid; European Skipper; butterfly; pollination;
pollinator interference; negative binomial distribution; Ottawa Valley
63
©The Ottawa Field-Naturalists’ Club (2017)
A contribution towards the cost of this publication has been provided by the Thomas Manning Memorial Fund of the Ottawa
Field-Naturalists’ Club.
2017 HALL ET AL.: REDUCED SEED DEVELOPMENT OF LADYS-SLIPPER ORCHID 64
10 July, the flowers had turned brown with development
of ovaries into capsules or had more or less withered.
The former had been pollinated but the latter had not
(Figure 2). Because most flowers had not completely
deteriorated, it was possible to open the brown lips and
note the presence of skippers and count them (Figure
3). This we did to test the hypothesis that the presence
of any skippers was deleterious, and the more skippers
there were in a flower, the greater the likelihood of
interference and the lower the likelihood of capsule
development.
FIGURE 1. Fresh Showy Lady’s-slipper (Cypripedium reginae) flower with one European Skipper (Thymelicus lineola) trapped
inside and one trying to enter the flower, Purdon Fen, eastern Ontario, 2015. Photo: Peter W. Hall.
FIGURE 2. Shrivelled Showy Lady’s-slipper (Cypripedium reginae) flowers, one with developed (right) and one with undeveloped
capsule, Purdon Fen, eastern Ontario, 2015. Photo: Peter W. Hall.
Methods
Purdon Fen is protected and managed by the Mis-
sissippi Valley Conservation Authority (MVCA). We
obtained permission to gather data; MVCA has shown
much interest over two decades in accumulating infor-
mation that would help to inform management (Mos -
quin and Brown 2006).
We sampled 475 flowers in an extensive area acces-
sible from the boardwalk and considered to be repre-
sentative of the entire colony. Within this area, we
selected flower stalks approximately 1 m apart. This
area included flowers that opened before the peak abun-
dance of skippers as well as those that opened after.
We avoided a small area where hand pollinations had
been done to increase fecundity. We examined flowers
and scored capsules as developed, partly developed, or
completely undeveloped (with empty brown shrivelled
capsules). For statistical purposes, we compared num-
bers of undeveloped capsules with the combined total
of fully and partly developed capsules.
Each flower was opened and the number of skippers
inside counted. We also examined 100 undamaged indi-
vidual skippers to determine their sex and whether they
carried any lady’s-slipper pollen. Males were distin-
guished by a characteristic (but sometimes inconspic-
uous) horizontal black stigma on the forewing.
Our main study hypothesis — that there was less
capsule development with skippers present and less
with more skippers was tested statistically using
tests of proportions and logit modeling. Also of interest
was the possibility of gender bias among counted skip-
pers, which was tested via a one-sample test of propor-
tions. We also considered the distribution of numbers of
skippers within flowers and whether it followed either
a Poisson or negative binomial model.
All statistical analyses were performed using the R
statistical computing language (R Core Team 2015).
The packages MASS (Venables and Ripley 2002) and
binom (Dorai-Raj 2014) were used for maximum like-
lihood estimation of the negative binomial distribution
and to calculate Bayesian credible intervals, respec-
tively.
Results
Of 475 flowers examined, 398 (83.8%) had no cap-
sule development and 77 (16.2%) had partly to fully
65 THECANADIAN FIELD-NATURALIST Vol. 131
FIGURE 3. Opened shrivelled Showy Lady’s-slipper (Cypripedium reginae) flower with seven trapped European Skippers
(Thymelicus lineola), Purdon Fen, eastern Ontario, 2015. Photo: Peter W. Hall.
developed capsules. Members of each group with var-
ious numbers of trapped skippers are indicated in Table
1. Of the plants with no skippers trapped in flowers,
24.3% had developed capsules, compared with plants
where skippers were present, of which 13.9% had
devel oped capsules. A two-sample test of equality of
proportion of developed capsules for plants both with
and without skippers present was rejected in favour
of the alternative that this proportion was less when
skippers were present (P = 0.0075). A logit model of
the probability of capsule development, with number
of skippers as the covariate, showed a significant effect
for number of skippers (P= 0.0271; one-sided because
our alternative hypothesis was directional; we tested
the null hypothesis of no effect of skipper number ver-
sus the alternative of a detrimental effect). The proba-
bility of capsule development decreased with number
of skippers (estimated logit model parameters: intercept
−1.37, slope −0.152). Model-predicted probabilities are
plotted in Figure 4, together with sample estimates and
their associated 95% Bayesian credible intervals (Jef-
frey’s prior) as a measure of their uncertainty.
Considering only the counts of skippers trapped in
flowers, we tested the hypothesis that the total count
(Table 1) follows a Poisson distribution. This would
happen if the rate of entrapment was constant and did
not depend on the number of skippers already trapped
and if all flowers were exposed to skippers for equal
amounts of time. After collapsing skipper counts of sev-
en or larger because of small expected values, a good-
ness-of-fit test (P< 0.0001; likelihood ratio) rejected
this simple model.
In our sample, the mean number of skippers per
flower was 1.94 with a variance of 3.28; thus, skippers
were over-dispersed relative to the Poisson distribution,
which requires mean and variance to be equal. A com-
mon model for over-dispersed count data is the two-
parameter negative binomial distribution. Maximum
likelihood estimation of its parameters gave a mean of
1.94 and a dispersion parameter of 2.87. After collaps-
ing skipper counts of nine or more because of small
expected values, goodness-of-fit testing did not reject
this model (P= 0.8656; likelihood ratio), but instead
resulted in a very close fit to these data. A frequency
distribution of the observed skipper counts together
with Poisson and negative binomial fitted values is
provided in Figure 5.
2017 HALL ET AL.: REDUCED SEED DEVELOPMENT OF LADYS-SLIPPER ORCHID 66
Table 1. Number of European Skippers (Thymelicus lineola)
in developed and non-developed capsules of Showy Lady’s-
slippers (Cypripedium reginae) at Purdon Fen near Lanark,
eastern Ontario in 2015.
No. Capsule not Capsule
skippers developed developed
0 81 26
1 104 24
2 85 9
3 57 10
4 33 1
5 23 3
6 7 2
7 4 0
8 2 1
9 1 0
10 0 1
11 0 0
12 1 0
FIGURE 4. Model-predicted probability of capsule development in Showy Lady’s-slipper (Cypripedium reginae) depending on
the number of entrapped European Skippers (Thymelicus lineola).
Of the sample of 100 trapped skippers with an un -
damaged upper forewing, 92 were males and the dif-
ference in proportion of sexes was highly significant
(P< 0.0001). Only three of 100 butterflies carried any-
thing that looked like a pollen smear. In each case, it
was on the smooth surface (by complete loss of scales)
of the dorsal thorax, and microscopic examination re -
vealed it to be stigmatic tissue. Only one monad char-
acteristic of Cypripedium pollen was seen.
Discussion
In this paper, we have shown that capsule develop-
ment in Showy Lady’s-slipper orchids at Purdon Fen
in spring 2015 was decreased when European Skipper
butterflies were trapped inside the flowers and, also,
that the probability of development decreased with the
number of skippers trapped. We have also shown that
the distribution of trapped skippers closely followed a
negative binomial distribution and that there was strong
gender bias in entrapment, males being predominant.
The 16.2% of orchids with capsule development in
2015 is smaller than the 22.7% with capsule develop-
ment recorded at the study site by Mosquin in 1985
(Mosquin 1986: 42, Table 3). However, this decline
should not necessarily be attributed to interference by
skippers, because many factors can account for differ-
ences in fecundity from year to year, such as late frosts
killing pollinators and/or flowers and population sizes
of pollinators. We note, however, that the percentage
of orchids with developed capsules, which had no en -
trapped skippers, was 24%; thus, the 1985 percentage,
being slightly smaller than this, might be achievable in
a year with very few skippers. In 2015, there was an
un usually high population of European Skippers in the
area, as determined by qualitative observations, so, they
might have had a higher than normal impact on capsule
development. Comparisons with other years could be
an interesting follow-up study.
The idea that productivity in 2015 might be com-
pared with that of past years when skippers were like-
ly less abundant is of interest, but constrained by a lack
of historical data and by a lack of information on many
factors that might affect seed development from year to
year. If additional risk factors are proposed, then the
logit modeling in this paper could be extended.
We considered the possibility that an entrapped skip-
per could encounter an anther in its attempts to escape
and then transfer pollen to the stigmatic surface of the
same flower. This could lead to an expanded ovary with
seeds. However, we have no evidence of this happening
based on examination of the surfaces of trapped skip-
pers for a pollen smear. The single monad found may
have been left on the stigma by an earlier pollinating
bee and then rubbed by the skipper onto its thorax. The
results of the data analysis also argue against regular
within-flower pollen transfer because the deleterious
ef fect on pollination increases with the number of
trapped skippers, whereas it would likely decrease with
increased trapped skippers available to transfer pollen
if there was within-flower pollen transfer.
Because skipper counts were negatively associated
with capsule development, we examined their distribu-
tion and found the number of trapped skippers to follow
a negative binomial distribution very closely. A well-
67 THECANADIAN FIELD-NATURALIST Vol. 131
FIGURE 5. Distribution of the observed number of European Skippers (Thymelicus lineola) trapped in Showy
Lady’s-slippers (Cypripedium reginae), showing both fitted-model Poisson and negative binomial model
counts.
known property of this distribution is that it is equiv-
alent to the counts being Poisson distributed with rate
parameter randomly distributed according to a gamma
distribution. Thus, if this decomposition is valid for the
Showy Lady’s-slipper,skipper counts are Poisson dis-
tributed at a flower-varying rate. Future analyses of
counts may be amenable to Poisson or negative bino-
mial regression analysis, if potential explanatory fac-
tors can be identified.
We found an extreme gender bias toward males
among entrapped skippers. The likely explanation, al -
though not sustainable by evidence in this case, is that
males of most butterflies emerge before females and
more likely overlap with the peak lady’s-slipper flow-
ering period.
This work is the first to associate European Skipper
butterflies with a negative impact on capsule and, there-
fore, seed production in the Showy Lady’s-slipper
orchid. It is also the first time that the potential impor-
tance of this phenomenon has been suggested because
of the widespread occurrence of trapped skippers.
Acknowledgements
We thank Jeremy Aldworth for helpful comments on a
draft of this paper. We also thank Ross Fergusson, Mis-
sissippi Valley Conservation Authority, for his estimate
of the number of orchids at Purdon Fen in 2015.
Literature Cited
Argue, C. L. 2012. The Pollination Biology of North Amer-
ican Orchids. Volume 1. Springer, New York, New York,
USA. https://doi.org/10.1007/978-1-4614-0592-4
Barrows, E. M. 1983. Cypripedium flowers entrap adult
Thymelicus lineola (Lepidoptera: Hesperiidae) in northern
Michigan. Journal of the Lepidopterists’ Society 37: 265–
268.
Catling, P. M. 1974. A butterfly-trapping orchid. Michigan
Entomological Society Newsletter 19: 1–3.
Dorai-Raj, S. 2014. binom: binomial confidence intervals for
several parameterizations. R package version 1.1-1. Ac -
cessed 30 January 2017. http://CRAN.R-project.org/pack
age=binom.
Hall, P. W., C. D. Jones, A. E. Guidotti, and B. Hubley.
2014. The ROM Field Guide to Butterflies of Ontario.
Royal Ontario Museum Press, Toronto, Ontario, Canada.
R Core Team. 2015. R: A language and environment for
statistical computing. Version 3.2.2. R Foundation for Sta-
tistical Computing, Vienna, Austria. Accessed 10 Decem-
ber 2015. https://www.R-project.org/.
Mosquin, T. 1986. A management plan for the Showy Lady-
slipper Cypripedium reginae, in the Purdon Conservation
Area, Lanark County, Ontario. Part I: final report. Mis-
sissippi Valley Conservation Authority, Lanark, Ontario,
Canada.
Mosquin, T., and L. Brown. 2006. A new management plan
for the Showy Lady-slipper, Cypripedium reginae, in Pur-
don Conservation Area, Lanark County, Ontario. Final re -
port. Mississippi Valley Conservation Foundation. Lanark,
Ontario, Canada.
Venables, W. N., and B. D. Ripley. 2002 Modern Applied
Statistics with S. Fourth edition. Springer, New York, USA.
https://doi.org/10.1007/978-0-387-21706-2
Vogt, C. A. 1990. Pollination in Cypripedium reginae. Lind-
leyana 5: 145–150.
Received 10 October 2015
Accepted 3 February 2017
2017 HALL ET AL.: REDUCED SEED DEVELOPMENT OF LADYS-SLIPPER ORCHID 68
ResearchGate has not been able to resolve any citations for this publication.
Book
A guide to using S environments to perform statistical analyses providing both an introduction to the use of S and a course in modern statistical methods. The emphasis is on presenting practical problems and full analyses of real data sets.
Book
Recent studies have revealed remarkable complexity and diversity in orchid-pollinator relationships. These studies comprise a vast literature currently scattered in numerous, often obscure, journals and books. The Pollination Biology of North American Orchids brings together, for the first time, a comprehensive treatment of this information for all native and introduced North American orchids found north of Mexico and Florida. It provides detailed information on genetic compatibility, breeding systems, pollinators, pollination mechanisms, fruiting success, and limiting factors for each species. Distribution, habitat, and floral morphology are also summarized. In addition, detailed line drawings emphasize orchid reproductive organs and their adaptation to known pollinators. This, the first of two volumes, furnishes a brief introduction to the general morphology of the orchid flower and the terminology used to describe orchid breeding systems and reproductive strategies. It treats the lady’s-slippers of genus Cypripedium, subfamily Cypripedioideae, and nine genera of the subfamily Orchidoideae, including the diverse rein orchids of genus Platanthera. The Pollination Biology of North American Orchids will be of interest to both regional and international audiences including: • Researchers and students in this field of study who are currently required to search through the scattered literature to obtain the information gathered here. • Researchers and students in related fields with an interest in the co-evolution of plants and insects. • Conservation specialists who need to understand both the details of orchid reproduction and the identity of primary pollinators in order to properly manage the land for both. • Orchid breeders who require accurate and current information on orchid breeding systems. • General readers with an interest in orchid biology. © Springer Science+Business Media, LLC 2012. All rights reserved.
Article
Adults of the introduced skipper Thymelicus lineola were attracted to the nectarless flowers of the native orchids Cypripedium reginae and C. calceolus. No doubt in search of food, they crawled into orchid labella. Up to 24 skippers and other insects became entrapped and died in a single labellum. . (Burns, 1966, pers. comm.; Preston & Westwood, 1981) and has become a pest of hay fields in Canada (McNeil & Duchsne, 1975; McNeil et aI., 1976). Adults seek nectar from many species of flowers including the lady's-slipper orchids Cypripedium reginae Walter and C. calceo-Ius L. This report increases knowledge about the peculiar entrapment of adult T. lineola by the labella of these orchids in northern Michigan.
A butterfly-trapping orchid
  • P M Catling
Catling, P. M. 1974. A butterfly-trapping orchid. Michigan Entomological Society Newsletter 19: 1-3.
R: A language and environment for statistical computing. Version 3.2.2. R Foundation for Statistical Computing
  • P W Hall
  • C D Jones
  • A E Guidotti
  • B Hubley
Hall, P. W., C. D. Jones, A. E. Guidotti, and B. Hubley. 2014. The ROM Field Guide to Butterflies of Ontario. Royal Ontario Museum Press, Toronto, Ontario, Canada. R Core Team. 2015. R: A language and environment for statistical computing. Version 3.2.2. R Foundation for Statistical Computing, Vienna, Austria. Accessed 10 December 2015. https://www.R-project.org/.
A management plan for the Showy Ladyslipper Cypripedium reginae, in the Purdon Conservation Area
  • T Mosquin
Mosquin, T. 1986. A management plan for the Showy Ladyslipper Cypripedium reginae, in the Purdon Conservation Area, Lanark County, Ontario. Part I: final report. Mississippi Valley Conservation Authority, Lanark, Ontario, Canada.
A new management plan for the Showy Lady-slipper, Cypripedium reginae
  • T Mosquin
  • L Brown
Mosquin, T., and L. Brown. 2006. A new management plan for the Showy Lady-slipper, Cypripedium reginae, in Purdon Conservation Area, Lanark County, Ontario. Final report. Mississippi Valley Conservation Foundation. Lanark, Ontario, Canada.
Pollination in Cypripedium reginae
  • C A Vogt
Vogt, C. A. 1990. Pollination in Cypripedium reginae. Lindleyana 5: 145-150.