2012 NORTHEASTERN NATURALIST 19(3):421–430
The Effect of Deer Browse on Sundial Lupine:
Implications for Frosted Elfi ns
Jennifer A. Frye*
Abstract - The effect of Odocoileus virginianus (White-tailed Deer) on Lupinus perennis
(Sundial Lupine) was quantifi ed for a site in Worcester County, MD. The reproductive
output of Lupine protected by deer exclosures was compared with Lupine that received
no protection from deer. Lupine in the exclosures had a higher likelihood of producing
seed pods and produced a greater number of seed pods per inflorescence. The implications
of these results on Callophrys irus (Frosted Elfi ns) are discussed.
Callophrys irus Godart (Frosted Elfi ns) are considered to be rare, imperiled,
or extirpated in every state where they have been known to occur (NatureServe
2011). The primary threats to this butterfly species are destruction and alteration
of their dry upland habitat through development, forest succession, and fi re exclusion
(NatureServe 2011, Wagner et al. 2003). The few remaining Frosted Elfi n
populations persist primarily as highly localized and fragmented populations
(Allen 1997, Glassberg 1999, NatureServe 2011, O’Donnell et al. 2007, Pfi tsch
and Williams 2009, Wagner et al. 2003). Odocoileus virginianus Zimmermann
(White-tailed Deer) can have a major impact on Frosted Elfi ns by consuming the
larval host plants, Lupinus perennis (L.) (Sundial Lupine; hereafter also “Lupine”)
and Baptisia tinctoria (L.) Vent. (Wild Indigo), both of which are browsed
by deer (Golden and Pettigrew 2005, NatureServe 2011, Schweitzer 2003, St.
Mary 2007). Deer may also affect Frosted Elfi ns through the direct consumption
of the butterfly’s eggs and larvae. In Maryland, Frosted Elfi ns are present at fewer
than ten sites and are currently listed as state-endangered. A single population
in western Maryland uses Wild Indigo as a host plant, while the other known
Frosted Elfi n populations in the state occur on the Atlantic Coastal Plain and
feed on Lupine. Sundial Lupine is listed as state-threatened and, like the Frosted
Elfi n, persists in small, fragmented populations. The few remaining Lupine sites
in Maryland require active habitat management to maintain, achieved primarily
through mechanical clearing of woody vegetation. Selective mechanical clearing
can maintain early-successional clearings with low to moderate tree cover and
relatively high light intensity, conditions under which Lupine has been shown to
thrive (Pfi tsch and Williams 2009, Smallidge et al. 1996).
The rarity of Lupine in the state is believed to be a limiting factor for
Frosted Elfins. The largest known population of Lupine-feeding Frosted Elfins
in Maryland occurs on a 2.1-ha parcel of private land. The Lupine was
*Maryland Department of Natural Resources, Wildlife and Heritage Service, PO Box 68,
909 Wye Mills Road, Wye Mills, MD 21679; email@example.com.
422 Northeastern Naturalist Vol. 19, No. 3
discovered at this site in 2005 following a timber harvest, and now represents
the largest population in the state with several thousand plants. Along the
road adjacent to the timber harvest site, Lupine has persisted since at least the
1970s, and those plants, although declining, have supported a population of
Frosted Elfins for over 30 years. It was presumably these Frosted Elfins that
quickly colonized the 2.1-ha parcel following the Lupine establishment in
2005. The 2.1-ha parcel is now held as a reserve for Frosted Elfins; the surrounding
habitat is still harvested. Frosted Elfins have been monitored annually
since establishment of the Lupines. While the population exhibits some
variation in the number of individuals observed every year, it has persisted at
the site since 2005.
In 2010, Lupine plants in the 2.1-ha site were producing almost no seed
pods, especially in early-successional clearings where the majority of the Lupine
was concentrated (S. Tangren, University of Maryland Arboretum, College
Park, MD, unpubl. data). The only plants producing multiple seed pods were
in, and adjacent to, a wooded buffer between the study site and the adjacent
roadside. This lack of seed pod production is a concern because Frosted Elfin
larvae subsist primarily on Lupine flowers and seed pods (NatureServe 2011,
Pfitsch and Williams 2009, Schweitzer 1992, Swengel 1996). While an obvious
decline in the numbers of Frosted Elfin adults at the site was not noted, the lack
of flowers and seed pods raised concerns of a future population crash. The lack
of seed pods was suspected to be the result of deer browsing. Accordingly, an
experiment was initiated to study the effect of deer browsing on Lupine and to
determine whether deer exclusion might be necessary to sustain theLupine and
Frosted Elfin populations at the site.
The study was conducted in Worcester County, MD, on the Atlantic Coastal
Plain east of the Chesapeake Bay. The area is predominantly made up of oakpine
forests underlain by relatively young (Tertiary and Quarternary) sediments
of uniformly low-relief (Schmidt 1993). Warm, humid summers and mild winters
are characteristic of the climate.
The study site was located on a 2.1-ha parcel of private property formerly
harvested for timber. Much of the area is located on a sand dune of the Parsonsburg
formation (Denny et al. 1979). The 2.1-ha study site is currently managed
to slow forest succession by clearing regenerating Pinus taeda (L.) (Loblolly
Pine) and limiting the growth of hardwoods (Quercus spp., Carya spp.) through
selective, mechanical removal. The surrounding area is actively harvested for
commercial timber production. Dominant shrubs and herbaceous vegetation include
Lupine, Vaccinium pallidum Ait. (Blueridge Blueberry), Gaylussacia baccata
(Wangenh.) K. Koch (Black Huckleberry), Nuttallanthus canadensis (L.)
D.A. Sutton (Canada Toadflax), Tephrosia virginiana (L.) Pers. (Virginia Tephrosia),
Piptochaetium avenaceum (L.) Parodi (Blackseed Needlegrass) and various
sedges including Carex tonsa var. tonsa (Fernald) E.P. Bicknell (Shaved Sedge),
2012 J.A. Frye 423
C. pensylvanica Lam. (Pennsylvania Sedge), C. albicans var. emmonsii (Dewey
ex Torr.) Rettig (Emmon’s Sedge), and C. nigromarginata Schwein. (Black Edge
Sedge). While Lupine is scattered throughout the study area, there are two main
concentrations, one on the eastern side of the dune and one on the western side.
In early April 2011, 41 clusters of Lupine were randomly selected for the deerexclosure
treatment. Deer exclosures were circular, constructed of vinyl green lawn
fence, and were approximately 76 cm high and 91 cm in diameter (fig. 1). The exclosures
allowed access to butterflies and other insects but completely protected the
plants within from deer browsing. Lupine clusters in 28 exclosures were randomly
chosen to be “paired” with Lupine clusters that were not protected by a deer exclosure
(fig. 2). Because it is impossible to differentiate between individual Lupine
plants in the fi eld, I used individual inflorescences ( Lupine flowering stems) as the
sample unit. Every Lupine inflorescence within each exclosure was monitored.
Unexclosed Lupine inflorescences were marked with wooden skewers pushed deep
into the ground with only the upper 2” visible. In all cases, the original number of
inflorescences inside the exclosures was comparable to those marked with skewers
outside the exclosures in early April. However, as the study progressed, the number
of inflorescences within the exclosure greatly outnumbered those that were marked
outside the exclosure. As a result, a greater number of exclosed inflorescences were
monitored as compared to the number of unexclosed inflorescences.
In a separate experiment, ten exclosures were modifi ed so that the bottom 6–8
inches of the fencing was removed to allow access to rabbits and other smaller
mammals, which also feed on Lupine (Zaremba and Pickering 1994, Zaremba et
al. 1991). The sharp edges of the fencing were then bent upwards to discourage
deer from reaching under the fencing to feed on Lupine flowers and seed pods.
The number of inflorescences within these ten exclosures and the number of seed
pods per inflorescence were compared with those of the 28 exclosures that were
not modifi ed to allow small-mammal access.
In both experiments, Lupine clusters were monitored approximately once a
week from April 27 to June 14. I quantifi ed (1) the number of inflorescences per
cluster, (2) inflorescence state (intact or browsed), and, as the main measure of reproductive
output, (3) the number of seed pods per inflorescence. Only pods that
were dehisced (having produced seed) were considered in the fi nal analysis. The
number of pods that had failed to develop and had not dehisced was negligible.
The average number of seed pods per inflorescence for each treatment in
both experiments (exclosed vs. unexclosed, small-mammal accessible vs. inaccessible)
was compared. The data did not follow a normal distribution and
were analyzed using Mann-Whitney U-Tests (Sokal and Rohlf 1969). In the
main experiment comparing exclosed and unexclosed Lupine clusters, the data
from the unexclosed plants was highly skewed (average number of seed pods,
skewness = 1.5). In the second experiment, which compared exclosed clusters
of Lupine accessible to small mammals with clusters that were inaccessible,
424 Northeastern Naturalist Vol. 19, No. 3
figure 2. Photograph of uncaged Sundial Lupine plants taken on 2 May 2011. This uncaged
cluster of plants was “paired” for photographic comparison with the caged plants
in figure 1.
figure 1. Photograph of caged Sundial Lupine plants taken on 2 May 2011.
2012 J.A. Frye 425
figure 3. Partially browsed Sundial Lupine inflorescence.
figure 4. Cluster of Sundial Lupine plants sheltered by a low-lying tree limb.
426 Northeastern Naturalist Vol. 19, No. 3
the smaller and unequal sample sizes also warranted use of the Mann-Whitney
U-Test. All analyses used upper-bound tables for critical values of the Mann-
Whitney U statistic. An online calculator (Avery 2007) was employed for all
Of the 41 original exclosures, three were excluded from the analysis because
the Lupine never fully developed and did not produce inflorescences.
Exclosed vs. unexclosed clusters
Reproductive output was measured as the number of seed pods produced by
each inflorescence. Twenty-eight exclosed and unexclosed clusters were compared
(this number excludes those exclosures accessible to small mammals).
The 28 clusters of unexclosed plants contained 144 inflorescences. Of these, 61
(42%) produced at least one pod. The 28 clusters of exclosed plants contained
753 inflorescences. Of these, 713 (95%) produced at least one pod.
Of the original 144 inflorescences that were not exclosed, the average number
of pods per inflorescence was 1.78 (SD = 1.97; Table 1). Of the original 753
inflorescences that were protected by an exclosure, the average number of pods
per inflorescence was 4.85 (SD = 1.56; Table 1). Pods per inflorescence was signifi
cantly different between exclosed and unexclosed clusters (Mann-Whitney U
= 696, n1 = n2= 28, P < 0.0001, r = 0.67).
Exclosures with and without access to small mammals
Ten exclosures were accessible to small mammals. These were compared
with the 28 exclosures that did not allow access to small mammals. There were
no significant differences in either the number of inflorescences per exclosure
(Mann-Whitney U = 164, n1 = 10, n2 = 28, P = 0.44, r = 0.13) or in the average
number of pods produced per inflorescence (Mann-Whitney U = 173, n1=
10, n2= 28, P = 0.29, r = 0.18). The data were also permuted three times to
conduct the test using equal sample sizes (comparing the ten exclosures that
allowed access to small mammals with three different sets of ten randomly selected
exclosures that did not allow access to small mammals). These results
also showed that there was no difference in the number of inflorescences per
exclosure (Mann-Whitney 59 < U < 67, n1 = n2 = 10, 0.21 < P < 0.53, 0.15 < r
< 0.28) or in the number of pods produced per inflorescence (Mann-Whitney
56 < U < 64, n1 = n2 = 10, 0.32 < P < 0.68, 0.09 < r < 0.23).
Table 1. Seed pod data for exclosed and unexclosed Sundial Lupine clusters.
# stems Mean
Original producing # pods
Treatment # stems seed pods per stem (SD) Mode Median Range Variance Skewness
Exclosed 753 713 4.85 (± 1.56) 5.7 4.95 2.3–7.9 2.43 0.12
Unexclosed 144 61 1.79 (± 1.97) 0.0 1.10 0.0–8.0 3.88 1.52
2012 J.A. Frye 427
Deer have a signifi cant effect on Lupine at the site. Clusters protected from
deer browsing exhibited both more inflorescences producing seed pods (95% vs.
42%) and a greater number of seed pods per inflorescence (4.85 vs. 1.78); photographs
of an exclosed and unexclosed Lupine cluster are provided in figures
1 and 2. In addition, deer were frequently photographed (motion-sensor camera)
feeding in the Lupine clusters at night, and deer sign (tracks and pellets) were
evident at the site. While rabbits and small mammals may have some effect on
Lupine, the effect was not statistically signifi cant.
Despite the presence of Frosted Elfi ns at the site every year, their success is
almost certainly hampered by deer herbivory. In 2011, Frosted Elfi n females were
frequently seen ovipositing on the new inflorescences before the flower buds
opened. As the buds started opening up, a large percentage of these were consumed
by deer. Based on the data from this study, at least 58% of inflorescences
were browsed entirely and 84% were partially browsed. This result suggests that
not only are deer consuming the host plant, but they are likely consuming many
Frosted Elfi n eggs (and possibly larvae) as well.
Occasionally, but much less frequently, ovipositing Frosted Elfi n females
were observed laying eggs on the Lupine leaves, which did not appear to be affected
by deer at this study site. This oviposition behavior was atypical, and it
was far more common to see females ovipositing on new inflorescences, which
coincides with what other researchers have reported (NatureServe 2011, Pfi tsch
and Williams 2009, Swengel 1996). Even if eggs that were laid on Lupine leaves
were to hatch and the larvae spared direct mortality from deer browsing, it seems
likely that many larvae would not survive to maturity, given the impact of browsing
and the resulting likelihood that few or no flowers or seed pods would be
available for their consumption. While there are no published studies on Frosted
Elfi n larval dispersal, it is generally accepted that due to their small size and
relative immobility, Frosted Elfi n larvae likely complete their development on a
single plant or small cluster of plants and may not be capable of moving freely
between Lupine clusters separated by large areas of open habitat.
Unprotected inflorescences produced few seed pods. In many cases, a blooming
inflorescence was entirely consumed, but in other instances it was common
for the lower one or two flowers to remain. figure 3 shows an example of a partially
browsed inflorescence in which several of the lower buds remained intact.
Many times these flowers developed into seed pods that did reach maturity and
eventually dehisced. It is not known if the presence of one or two single flower
buds on an inflorescence provides enough food for a Frosted Elfi n larva to complete
Throughout the study site, unprotected inflorescences (including those that
were monitored for this study and those that were not) that successfully set seed
frequently met one of two conditions: (1) they were protected by low-lying tree
limbs (fig. 4) or large dead tree braches, or (2) they were relatively isolated from
the large Lupine concentrations. It is worth noting that most of the ovipositing
Frosted Elfi n females that I encountered were found in areas with the high Lupine
concentrations. It would be useful to know if female Frosted Elfi ns are less likely
428 Northeastern Naturalist Vol. 19, No. 3
to encounter or oviposit on smaller, isolated Lupine clusters or Lupine inflorescences
that are partially obscured by trees.
finally, it is also noteworthy that Frosted Elfins on the adjacent roadsides
have persisted for over 30 years, despite the fact that the concentration of Lupine
along the road is much reduced compared to the concentration of Lupine
at our study site. However, Lupine plants along the roadside are successful
in producing large numbers of seed pods (S. Tangren, unpubl. data), probably
because deer are less likely to feed on Lupine that occupy smaller and more
isolated clusters along a narrow roadside with no shoulder. It is possible that
many of the Frosted Elfin adults observed at the study site every year actually
developed along the adjacent roadsides on Lupine with abundant flowers and
seed pods. These roadside plants are also quite shaded, which may provide
favorable microhabitat conditions for Frosted Elfin larvae. Albanese et al.
(2008) suggests that in populations of Wild Indigo-feeding Frosted Elfins, adult
females will oviposit on Wild Indigo plants across a range of microhabitats.
Despite this seemingly indiscriminate microhabitat selection by adult females,
Frosted Elfin larvae experienced increased larval success on plants that occurred
in shaded areas; very few larvae reached maturity on plants that were
in open areas exposed to full sun. Similar results have been reported for the
Lupine-feeding Lycaeides melissa samuelis Nabokov (Karner Blue Butterfly)
(Grundel et al. 1998, Lane and Andow 2003) and for two species of swallowtail
butterflies (Rausher 1979). Other researchers have reported an increase in the
density of Frosted Elfin butterflies in areas with some degree of canopy cover
(Albanese et al. 2007, Pfitsch and Williams 2009, Swengel 1996), furthering
the idea that some degree of shade may be required for the success of Frosted
Elfin development. There is certainly a great need for data that assesses larval
development and survivorship as a function of microhabitat differences for the
Frosted Elfin as well as other imperiled Lupine feeders.
This project was made possible through State Wildlife Grant Funding administered
through the US fish and Wildlife Service, and through a generous grant from The
Forestland Group LLC. The Forestland Group removed the 2.1-ha site from timber production
to promote the conservation of Sundial Lupine and Frosted Elfi ns. Larry Walton
and Vision Forestry were instrumental in assisting with site management and with the
coordination and oversight of research activities at the study site. Sara Tangren’s observations
of Lupine reproduction (or lack thereof) spearheaded this study. She, along with
staff and volunteers from both the Maryland Department of Natural Resources and The
Nature Conservancy, were instrumental in assisting with construction and deployment of
lupine cages. I am also grateful to Chris Frye for his assistance with fi eld work and plant
identifi cation, to Robert Gano and John Moulis for their assistance with exclosure design,
and to Dale Schweitzer, Dave Wagner, Mike Nelson, and Richard Smith, all of whom
provided information, references, and guidance that allowed me to design and execute
the study. Ralph Grundel, Glen Mittelhauser, Chris Frye, Wes Knapp, Jim McCann, and
two anonymous reviewers provided comments that greatly improved this manuscript.
2012 J.A. Frye 429
Albanese, G., P.D. Vickery, and P.R. Sievert. 2007. Habitat characteristics of adult Frosted
Elfi ns (Callophrys irus) in sandplain communities of southeastern Massachusetts,
USA. Biological Conservation 136:53–64.
Albanese, G., P.D. Vickery, and P.R. Sievert. 2008. Microhabitat use by larvae and females
of a rare barrens butterfly, Frosted Elfi n (Callophrys irus). Journal of Insect
Allen, T.J. 1997. The Butterflies of West Virginia and their Caterpillars. University of
Pittsburgh Press, Pittsburgh, PA. 388 pp.
Avery, L. 2007. Mann-Whitney U-Test online calculator. Available online at http://elegans.
swmed.edu/~leon/stats/utest.html. Accessed June 2011.
Denny, C.S., J.P. Owens, L.A. Sirkin, and M. Rubin. 1979. The Parsonsburg sand in
the central Delmarva Peninsula, Maryland, and Delaware: Surface and shallow subsurface
geologic studies in the emerged coastal plain of the Middle Atlantic States.
Geological Survey professional paper; 1067-B. US Department of the Interior, Washington,
DC. 18 pp.
Glassberg, J. 1999. Butterflies through Binoculars—the East: A field Guide to the Butterfl
ies of Eastern North America. Oxford University Press, New York, NY. 242 pp.
Golden, D.M., and L. Pettigrew. 2005. Frosted Elfi n Management Plan. New Jersey
Department of Environmental Protection, Division of fish and Wildlife, Trenton, NJ.
Grundel, R., N.B. Pavlovic, and C.L. Sulzman. 1998. The effect of canopy cover and
seasonal change on host plant quality for the endangered Karner Blue Butterfly (Lycaeides
melissa samuelis). Oecologia 114:243–250.
Lane, C.P., and D.A. Andow. 2003. Oak savanna subhabitat variation and the population
biology of Lycaeides melissa samuelis (Lepidoptera: Lycaenidae). Annals of the Entomological
Society of America 96:799–809.
NatureServe Explorer. 2011. An online encyclopedia of life [web application]. Version
4.6. Arlington, Virginia, USA: NatureServe. Available online at http://www.natureserve.
org/explorer. Accessed July 2011.
O’Donnell, J.E., L.F. Gall, and D.L. Wagner. 2007. The Connecticut Butterfly Atlas. State
Geological and Natural History Survey Bulletin No. 118, Department of Environmental
Protection, Harford, CT. 376 pp.
Pfi tsch, W.A., and E.H. Williams. 2009. Habitat restoration for lupine and specialist butterfl
ies. Restoration Ecology 17:226–233.
Rausher, M.D. 1979. Larval habitat suitability and oviposition preference in three related
butterflies. Ecology 60:503–511.
Schmidt, M.F., Jr. 1993. Maryland’s Geology. Tidewater Publishers, Centreville, MD.
Schweitzer, D.F. 1992. Comments regarding Erynnis persius persius and Incisalia irus.
The Ohio Lepidopterist 14:21–23.
Schweitzer, D.F. 2003. Panel Discussion Paper: Mitigating effects of intense deer herbivory.
Pp. 16-17, In S. Damais, E. Stiles, and M. Dufort (Eds.). Proceedings of the
State of the Forest Symposium: Ecological issues regarding Highlands Forest degradation
and restoration. New Jersey Audubon Society, Bernardsville, NJ. 22 pp.
Smallidge, P.J., D.J. Leopold and C.M. Allen. 1996. Community characteristics and
vegetation management of Karner Blue Butterfly (Lycaeides melissa samuelis)
habitats on rights-of-way in east-central New York, USA. Journal of Applied Ecology
Sokal, R.R., and F.J. Rohlf. 1969. Biometry: The Principles and Practice of Statistics in
Biological Research. W.H. Freeman and Company, San Francisco, CA.
430 Northeastern Naturalist Vol. 19, No. 3
St. Mary, M.K. 2007. A study on the effectiveness of transplanting vs. seeding of Lupinus
perennis in an oak savanna regeneration site. M.Sc. Thesis. Bowling Green State
University, Bowling Green, OH. 40 pp.
Swengel, A.B. 1996. Observations of Incisalia irus (Lepidoptera: Lycaenidae) in central
Wisconsin 1988–1995. The Great Lakes Entomologist 29:47–62.
Wagner, D.L., M.W. Nelson, and D.F. Schweitzer. 2003. Shrubland Lepidoptera of southern
New England and southeastern New York: Ecology, conservation, and management.
Forest Ecology and Management 185:95–112.
Zaremba, R.E., and M. Pickering. 1994. Lupine ecology and management in New York
State. Pp. 87–93, In D.A. Andow, R.J. Baker, and C.P. Lane (Eds.). Karner Blue Butterfl
y: A Symbol of a Vanishing Landscape. Minnesota Agricultural Experimental
Station (Misc. Publication Series), St. Paul, MN. 220 pp.
Zaremba, R.E., M. Pickering, and B. Hunt. 1991. Preliminary results of the 1990 lupine
(L. perennis) propagation project in the Albany Pine Bush. Report to the NYSDEC,
Endangered Species Unit, Albany, The Nature Conservancy, NY field Offi ce, Albany,
NY. 24 pp.