I'll

JOSHJ.IINS _S3 I ava a n~LI B RAR I ES^SMITHSONIAN_INST!TUTION%OlinilJLSNI-NVIN<

Z * CO 2L

X
to
O

S ".y > s '\J^ i‘ " ' § ^

•HSONIAN^INSTITUTION </,N0IXIUIi.SN[_NVIN0SHllWSOTS3 I ava ail^LIBRAR I ES^SMITI
^ \ in — ,n

oshiiws ssiavaan libraries Smithsonian institution NoiimiiSNi nvinc

z i~ __ 2 r-

Stc.- O m /452S3X 2 . ,m /tfSSSX 9. ^v«rx ~ „<S

CD

> l-F ^§31 P > 1-F ^331 P f;?i£

msK c \^»>arf/ ? vtSoiw^y i- •tfeSHL f Wiiw?/ - v%^..S9ni »

im

HSONIAN "INSTITUTION NOIlfllllSNI NVINOSRUWS^SS I 8V8 8 IT~U B RAR I ES^SMITI

CO 2 to 2 ..... to

< ^ v.- 2 . < 2 a£x .•» 2

I0SHJUWS"'S3 I ava a 11^ LIB RAR I ES^SMITHSONIAN^INSTITUTION^NOliniliSNI NVIN'

c/> = co = in

o

< ,„j ” z _

rHSONlAN INSTITUTION NOIXflXIXSNI NVINOSHXIWS S31BVH811 LIBRARIES SMIT

5 _ • z

O ><IsTim>v “ O — /floXX O

CO

P* '• ~ _

m \^osv£X ^ Xjvasv^X m '^y 2 m

MOSHXIWS^SS I HVd a Il~LIBRAR I ES $MlTHSONIAN~INSTITUTION%OIXfUIXSNl"NVIN

5 * ^ ^ z n co z

SosSgX >• 2 > '0*22' s ' ^ g •*

rHSONlAN INSTITUTION WNOlirUllSNI NVINOSHilWS^SS I a Va a ll\ I B R AR I ES^SMIT

CO -3» \ CO __ ~ CO

joshiiws ssiavaan libraries Smithsonian institution nouiuilsni nvin

M. z I- _ 2 r* 2 r-

m « v^pius]^ m

co x co ~ co

FHSONIAN INSTITUTION NOIXnXIXSNl NVINOSHXIWS S3IHVHan LIBRARIES SMIT
z .y. co z w z £

m ><VOS»^ 2 Z m V^yw.y z • ■

;W S3 I ava an~ LIBRARIES SMITHSONIAN-INSTITUTIONt/>NOIllUliSNI-NVINOSHimSt/
co z •, c» z , to 2

co

|' ' ' I V^T | | y ! * ' ' g

Institution0’ N0iiniiisNi^NviN0SHiiws”s3 1 ava a n\\ brar i eswsmithsonian

I | i ^ «

^..uw / gj '^jp/ g ffl XjsaiP' 5 “•*’

^ a SS «»fl d

S3 I a Va 311 LIBRARIES SMITHSONIAN^fMSTITUTION NOlinillSNI NVIN0SH1MS
O S - -|p^v ” C

w m xgx ocj^ g <||pF S 's^2*g&X to ^ t

"institution w NoimiiisNrNViNOSHiiws S3 1 ava a n~LI b RAR I es^smithsonian

Z .^-. CO Z — _ CO Z CO _ „

X
co
o
2

> S S >

"S3 I ava a n2 LI B RAR I ES^SMITHSONIAN INSTITUTION NOlinillSNI nvinoshiiws"

CO =; I oo 2 ^ ^ 00 ^

4 ^

o N^voCX ~ O _ X^UixS^X Q

riNSTITUT!0N2M0lini!lSNl‘JNVIN0SHllINS2S3 I ava 8 ll^LI B RAR I ES^SMITHSONIAN”
r* x C " Z r- Z r

o xtSTirl^Sv __ .W O ~ /^av7:\ O

00

>y >- v^viiwfi

m x^osv^X ^ n^vasv^x m '' vw ^ m x^osv^ ^ r

co E co ~ to £ c

; S3 1 ava a n libraries Smithsonian institution NouniiiSNi nvinoshiiws

CO Z CO z v CO z c

< V S X^ilo^s. < ^ s <

A z 5 «o4. Z

X
CO

o

I INSTITUTION </>N0IJtniliSNI NVINOSHIIWS^SS 1 avaa nZLI BRAR I ES^SMITHSONIAN
co ^ \ ^ co

CO

O ‘ O

Z „j Z _J z

S3 1 ava an libraries Smithsonian institution NouniiiSNi nvinoshiiws

| y5 | 5 ^ !

br ^ // tn Lmyrnb&. \ *- -r, ^

<Oulls£

CO ™ CO „ co

J institution NouniiiSNi nvinoshiiins S3 1 ava a n libraries smithsonian"

z *\y\ CO Z CO Z .... CO

^ .AS.

;

QL

541

J866

ENT

E JOURNAL
w. RESEARCH
ON THE LEPIjSlPT ERA

Volurf:;/)

NurnOm

Spring 19:

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

ISSN 0022 4324

Published By:

The Lepidoptera Research Foundation, Inc.
9620 Heather Road
Beverly Hills, California 90210
(213) 274-1052

Founder:

William Hovanitz

Editorial Staff:

Rudolf H. T. Mattoni, Editor
Scott E. Miller, Assistant Editor

Associate Editors:

Emilio Balletto, Italy

Henri Descimon, France
Philip DeVries, U.S.A.

Thomas Emmel, U.S.A.
Lawrence Gall, U.S.A.
Hansjuerg Geiger, Switzerland
Otakar Kudrna, Germany
Arthur Shapiro, U.S.A.

Atuhiro Sibatani, Japan
Karel Spitzer, Czechoslovakia

Manuscripts and Notices Material may be sent to the Editor at:
9620 Heather Road, Beverly Hills, CA 90210 (213) 274-1052

The JOURNAL is sent to all members of the FOUNDATION.

Classes of Membership:

Regular (Individual)
Contributing

Student/Retired-Worldwide
Subscription Rate/Institutions
Life

$ 20.00 year (vol.)

$ 30.00 or more, year (vol.)

$ 18.00 year (vol.)

$ 28.00 year (vol.)
$ 250.00

STATEMENT OF OWNERSHIP AND MANAGEMENT

THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA is published four times a year by THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The office of publication and the general business
office are located at 9620 Heather Road, Beverly Hills, California 90210. The publisher is THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The Editor is R. H. T. Mattoni at the above address.
The Secretary-Treasurer is Leona Mattoni at the general business office. All matters pertaining to
membership, dues, and subscriptions should be addressed to her, including inquiry concerning mailing,
missing issues, and change of address. The owner is THE LEPIDOPTERA RESEARCH FOUNDA-
TION, INC., a non-profit organization incorporated under the laws of the State of California in 1965. The
President is R. H. T. Mattoni, the Vice President is John Emmel, the Secretary-Treasurer is Leona
Mattoni. The Board of Directors is comprised of Barbara Jean Hovanitz, Leona Mattoni, and R. H. T.
Mattoni. There are no bond holders, mortgages, or other security holders.

Printed by Art Printing Works Sdn. Bhd., 29 Jalan Riong, 59100 Kuala Lumpur, Malaysia.

Journal of Research on the Lepidoptera

29(1-2):1-10, 1990(91)

Hilltopping by the Red Admiral Butterfly: Mate
Searching Alongside Congeners

William D. Brown* and John Alcock

Department of Zoology, Arizona State University, Tempe, AZ 85287-1501

Abstract. Males of the red admiral butterfly Vanessa atalanta
(Nymphalidae) establish and defend territories in the late afternoon on
a central Arizona hilltop. Resident males engage conspecific rivals in
lengthy chases but respond less aggressively toward males of 3 conge-
neric species. A comparison between V. atalanta and 2 congeners, V.
annahella and V. cardui, shows that the degree of site tenacity is
correlated with the density of rivals. Males of V. annahella , the least
abundant species, are most site tenacious, as measured by mean
duration of residency and frequency of return to the peak. In contrast,
males of V. cardui experience the highest hilltop densities and are the
most ephemeral. Males of V. atalanta occur in intermediate density
and exhibit intermediate site tenacity. Although male V. atalanta
readily respond to any flying insect, the duration and complexity of
these interactions is reduced. The various species also express differ-
ent perch site preferences, which may aid in reducing the frequency of
non-productive congeneric encounters.

Introduction

Hilltops serve as mate-encounter sites for a striking array of insects
(Alcock, 1987; Shields, 1967). In many cases, including the red admiral
butterfly ( Vanessa atalanta ), males perch along ridges and atop peaks
where there is no concentration of resources useful to females (Alcock,
1984, 1987; Dimock, 1978; Shields 1967). Possibly hilltops act as
orientation guides or dispersal barriers for species in which receptive
females are otherwise evenly or diffusely distributed in the environment.
In some regions, however, hilltops are unavailable and so male Vanessa
butterflies adjust their mate-searching tactics by using alternative
landmarks as territory centers (Alcock & Gwynne, 1988; Bitzer & Shaw,
1979; Palm, 1980). Where hilltops are absent, both V. atalanta and an
Australian congener, V. kershawi, use landscape features such as side-
walks and forest clearings as points at which to search for mates. Thus,
regardless of the local terrain, Vanessa butterflies appear to establish
territories at places that are topographically distinctive.

Although different species of Vanessa have been reported to co-occur at
hilltops (Dimock 1978; Shields 1967), little is known of the influence
congeners may have on their mating systems. Here we report that V.
atalanta interacts with other species of Vanessa {annahella, cardui and
uirginiensis ) at a central Arizona hilltop. We describe the hilltopping
behavior of male V. atalanta at this site with special attention to the

^Present address: Department of Zoology, Erindale College, University of Toronto,
Mississauga, Ontario L5L 1C6, Canada

Fig. 1 . A view of the peaktop study site.

influence of congeners on mating system dynamics and examine if
congeneric males have evolved ways to reduce potentially costly interac-
tions with each other. We also compare the behavior of V. atalanta with
V. annabella and V. cardui, and show a correlation between male density
at the hilltop and the degree of site tenacity.

Methods

The hilltopping behavior of the V. atalanta , V. annabella and V. cardui, was
observed on 69 days between 8 October 1988 and 13 April 1989. The study site
was a dominant peak along Pima Ridge of South Mountain Park, Phoenix,
Maricopa Co., Arizona (Fig. 1). Here the sparse vegetation was dominated by
creosote bush (. Larrea divaricata), brittlebush ( Encelia farinosa), and staghorn
cholla ( Opuntia acanthocarpa ).

At least one day each week, an attempt was made to record the arrival time of
the first Vanessa of the day and the departure time of the last Vanessa , beginning
the week of 20 October 1988. In addition, we analyzed interactions among the
butterflies using the following procedure. Resident V. atalanta males were
captured, given a unique mark on the upper side of the wing with colored liquid
paper, and released within 2 min. They generally appeared undamaged by this
treatment and either returned immediately to perch on the hilltop or flew off the
peak. Once a marked V. atalanta male re-established residency, we recorded the
durations of a single intraspecific interaction and a single congeneric interaction
for 23 individuals. Starting from when the rivals were within ca. 10 cm of each
other, we used a stopwatch to measure either the time one veered off the chase
(i.e. a “drop”, see Results, or a difference of greater than ca. 90° in their flight
paths) or the time until the pair was lost from sight.

November and December 1988 were devoted to a comparative study of the site
tenacity of V. atalanta , V. cardui , and V. annabella. Butterflies of each species
were captured and marked. Sizes of individuals were categorized as either small,
medium or large. Capture frequency gave a rough estimate of the relative
abundance of each species on the hilltop. We also estimated the maximum
number of each species of Vanessa at the hilltop at any one time during the
afternoon. All previously marked butterflies that returned to the hilltop, both

29(1-2):1-10, 1990(91)

3

those marked earlier that day and those marked on previous days, were recorded
along with their time of arrival and duration of stay.

To analyze perch preferences of the 3 species, three 7 x 7 m quadrats were laid
out on the hilltop, along the natural axis of the ridgeline (roughly east-west). The
quadrats were immediately adjacent to each other and varied in altitude.
Quadrat 1 (farthest east) was the peaktop; quadrat 2 was intermediate in
altitude and quadrat 3 (farthest west) was at the lowest elevation, ca. 5 m below
the peaktop. From 17 November to 13 February, we recorded the quadrats in
which marked males perched after periods of flight. In addition to records of
perch sites, the quadrat of capture was noted for each marked individual.

Means are reported ± 1 SE.

Results

SEASONAL TRENDS IN ARRIVAL AND DEPARTURE FROM THE
HILLTOP

Both the arrival and departure of the first and last Vanessa on a given
day are correlated with the time of year (Fig. 2). The butterflies arrived
earlier as days shortened and later as days lengthed. This seasonal effect
explains over 91% of the variance in hilltop departure times, the butter-
flies usually leaving within a few minutes of sunset. Arrival times,
however, were not so predictable with only 50% of variance being
explained by season, perhaps because of climatic effects. On relatively
cool, windy days, arrival was delayed or the butterflies did not appear at
all. On 12 days V. atalanta was absent on days of strong wind and heavy
overcast, but returned on clear days later in the week. Shields ( 1967) also
found a correlation between temperature, relative humidity, and the
arrival time of Vanessa species.

Season also influenced the density of butterflies on the hilltop, mea-
sured both by the number captured and the maximum number observed
at one time (Fig. 3). The density of hilltopping butterflies decreased as
winter approached and increased again in the spring. The number of
butterflies captured on a given day might have been influenced by both
the density of butterflies and the duration of activity. However, because
as days lengthened, butterflies postponed establishing residence on the
hilltop until later in the afternoon, there was no relationship between
date and duration of hilltopping activity (r = 0.14, df = 11, p > 0.05; mean
duration of activity = 2.4 ± 0.2 h).

MALE-MALE ENCOUNTERS

V. atalanta males typically perched on the ground, on either stones or
the bare ground. On especially hot days, males perched on hilltop shrubs.
Resident males flew toward a variety of flying insects such as the wasp
Hemipepsis ustulata and the hairstreak Strymon melinus, returning
quickly to the perch if it was not a fellow Vanessa. Congeners, however,
often elicited a chase, usually a quick, erratic flight across the peaktop.
Chases were typically terminated when one individual veered away from
its rival and returned to its perch. Resident V. atalanta chased
heterospecific males for a mean period of 5.3 ± 1.1 s (N = 23).

4

J. Res. Lepid.

EH

m

5

p

PH

O

W

S

H

2000
1900

1800
1700

1600
1500
1400
1300

OCT NOV DEC JAN FEB MAR

■ ARRIVAL
□ DEPARTURE

D □

□ □
□

□ □

DATE

Fig. 2. The arrival time of the first Vanessa of an afternoon and the departure time
of the last Vanessa of an afternoon, October 1988 to March 1989. For
arrival, r = 0.71 , N = 1 4, p = 0.01 5; for departure, r = 0.96, N = 1 4, p < 0.0001 .
Arrival and departure times were not species specific; results are consistent
for each species (p < 0.001 ).

«o

Ho

Q

'-o

e

HO

$

Q

co

CO

CD

£

o

w

CQ

§

&

£

NOV DEC JAN FEB MAR APR
DATE

Fig. 3. Number of V. atalanta captured per day, November 1988 to April 1989; r
= 0.61 , N = 53, p < 0.0001 . Regression of date versus maximum number
of V. atalanta at any one time of an afternoon yields a similar result (p <
0.0001). Results are also consistent for V. annabella and V. cardui (p <
0.05).

29(1-2):1-10, 1990(91)

5

In contrast, interactions between two V. atalanta males were often
lengthy and complex. V. atalanta males chased conspecific rivals for a
mean of 16.4 ± 3.7 s (N = 23), a significantly longer time than for
heterospecifics (paired t = 2.75, df = 22, p = 0.01). After the initial
approach, rival males often hovered for 1 or 2 s, one directly above the
other, wings occasionally hitting, before entering into a prolonged chase.
Chases took a variety of forms, from circles of the hilltop to sinuous
chases along the ridge-line to long, ascending chases far off the peaktop.
Conspecific chases ended when one of the pair dropped out of the chase,
falling from the sky in a J-shaped dive, and either returned to the hilltop
or continued on its way. Hovering occurred in 9 of 19 intraspecific and
0 of 19 congeneric interactions (%2 = 11.79, df = 1, p < 0.001). Dropping
out occurred in 13 of 19 intraspecific and 0 of 19 congeneric interactions
(X2 = 19.76, df - l,p< 0.001).

MALE ^FEMALE ENCOUNTERS

Male-female interactions differed markedly from interactions involv-
ing two males. Seven encounters between conspecific males and pre-
sumptive females were observed during the study (4 involving V. cardui,
2 with V. atalanta , and 1 with V. annabella ). All male-female encounters
were characterized by a slow, fluttering flight with the male following
closely behind the larger female in a pattern highly reminiscent of
Chlosyne californica, another hilltopping nymphalid (Alcock, 1985).
These encounters began at the hilltop but all pairs subsequently flew
downslope and were eventually lost from view. Unlike male-male
encounters, the males failed to return promptly.

The courtships occurred as early as 1433 MST on 4 December and as
late as 1650 MST on 10 April (Table 1). Moreover, a mating pair of V.
cardui was sighted several hundred feet below the hilltop at 1803 MST
on 29 March. Thus, courtship evidently does occur at the hilltop with
female arrivals occurring throughout the period of male activity at the
peak (see also Palm, 1980). Matings presumably occur downslope.

SITE TENACITY AND

MALE DENSITY

A total of 114 V. atalanta ,

Table 1 .

Times of Vanessa courtship

Species

Date

Time

215 V. cardui and 68 V.
annabella were captured, with

V. annabella

4

December

1547

the numbers reflecting the rela-
tive densities of the three spe-

V. atalanta

28

November

1647

cies. The degree of hilltop site

V. atalanta

4

December

1433

tenacity was negatively asso-
ciated with density of conspe-

V. cardui

4

November

1645

cific males at the hilltop (Table

V. cardui

16

January

1605

2). The lower density species

V. cardui

11

February

1622

(V. atalanta and V. annabella )

V. cardui

10

April

1650

6

J. Res. Lepid.

Table 2. Frequency of return to the hilltop and duration of stay for males of 3

Vanessa species.

Species

Day of Marking1

Male Returns On

Subsequent Days2 Duration of Stay (min)3

V. atalanta

46 of 114(40%)

24 of 114(21%)

26.87 ±4.73 (N = 46)

V. cardui

42 of 215 (20%)

14 of 215 (7%)

15.00 + 3.24 (N = 42)

V. annabella

20 of 68 (29%)

15 of 68 (22%)

41.61 ±6.49 (N = 20)

Y=16.5, df
2%2 = 19.0, df
3ANOVA, F -

= 2, p = 0.0003
= 2, p = 0.0001

4.23, p = 0.007

tended to return to the hilltop after marking to a greater degree than the
high density V. cardui. The pattern holds even when we exclude males
that were marked within an hour of sunset: 24 of 45 (53%) V. atalanta,
15 of 29 (52%) V. annabella, and only 20 of 68 (29%) V. cardui were
sighted again on the afternoon they were marked (%2 = 7.93, df = 2, p =
0.019).

Similarly, the likelihood of return to the hilltop on days after marking
was particularly low for V. cardui (Table 2). Whereas about 21% of V.
atalanta and about 22% of V. annabella were sighted on at least 1 day
after being marked, only 7% V. cardui males were resighted, a statisti-
cally significant result.

If they returned after being marked, males of the different Vanessa
spp. also differed in the duration of stay at the peaktop (Table 2). Male
V. atalanta remained at the peak for up to 152 min on a given afternoon,
with an average of 26.87 ± 4.73 min (N = 46). V. annabella had the longest
average duration of stay (41.61 ± 6.49 min, N = 20, range = 1 - 91), and
V. cardui individuals were the most ephemeral with a mean duration of
15.00 ± 3.24 min (N = 42, range = 1 - 92).

Although male V. atalanta were relatively site tenacious, overthrows
in hilltop possession occurred. Of 49 males that re-established residency
after marking, 27 (55%) were unable to hold the hilltop territory for the
remainder of the afternoon. Similarly, 9 of 21 (43%) V. cardui residents
lost their perch (however, far fewer cases of sole residency occurred for V.
cardui males). Again, V. annabella males were the most site tenacious
with only 4 of 23 ( 17%) being overthrown. Frequency of turnover differed
significantly among species (%2 = 9.07, df = 2, p < 0.025).

Male longevity was much greater than individual tenure at the hilltop
site. Male V. atalanta and V. annabella were re-sighted up to 21 days (N
= 2) after their initial marking (N = 8 sightings of V. atalanta more than
14 days after marking). Thus, the failure of territorial males to hold their
perch site over a period of days is puzzling. Previous residents seemed

29(1-2):1-10, 1990(91)

7

not to have much advantage in holding the site when they did return on
subsequent days. Eleven of 24 (46%) V. atalanta males, 8 of 15 (53%) V.
annabella males, and 7 of 14 (50%) V. cardui males that returned on a
subsequent day remained less than 10 min and failed to establish
territorial residency, perhaps because they were ousted by younger,
more energetic rivals.

No relationship exists between body size and measures of site tenacity.
For neither V. atalanta or V. cardui does the likelihood of return to the
hilltop vary significantly with size class (%2-tests, NS; data on V. annabella
are insufficient for a similar test). Similarly, there is no apparent
relationship between body size and frequency of return after capture for
either of these species; sample sizes, however, are low and further tests
are needed on this point.

PERCH PREFERENCES

The 3 species of Vanessa spatially separated themselves on the hilltop
(Table 3). Both V. atalanta and V. annabella preferentially perched in
quadrat 2, just below the peaktop. The numerically far more abundant
V. cardui preferred the peak (quadrat 1). None of the Vanessas showed
a preference for quadrat 3, the low elevation site. Although V. atalanta
and V. annabella shared residency at quadrat 2, they consistently perched
at the opposite ends of this section, V. atalanta towards the west at
slightly lower elevation and V. annabella along the border of quadrat 1.

Male V. atalanta captured while perching in quadrat 2 were more likely
to be sighted again that afternoon than were conspecific males perched
in other sections of the peaktop (Gadj = 7.84, df = 1, p < 0.01). Males
occupying region 2 tended to be residents that chased off intruders.
These intruders often perched briefly outside the favored area of this
species before departing.

Table 3. Perch preferences and capture sites of Vanessas at the hilltop.

Species Perch Preference1 Capture Site2

Quadrat: 1 2 3 1 2 3

V. atalanta 1 15 0 9 18 2

V. cardui 25 7 4 89 18 5

V. annabella 1 7 0 7 24 0

Quadrat 3 is omitted to fit assumptions required for a G-test;

Gadj = 30.4, df = 2, p < 0.0001

2Quadrat 3 is omitted to fit assumptions required for a %2-test;

X2= 51 , df - 2, p< 0.0001

8

J. Res. Lepid.

Discussion

TEMPORAL PATTERNS OF MATE SEARCHING
In all 3 species, the abundance of males at the peak varies with the
season, with the highest densities occurring at times of moderate tem-
peratures (mid-fall and late winter). Males of all 3 species also exhibit
similar daily patterns of mate-locating activity, with individuals arriving
within 3 h of sunset and departing at dusk. Although receptive females
arrive infrequenty at hilltops, they do so during the period when males
are active at these locations. Perhaps late-day matings work to the
female’s advantage because they allow the protracted copulations to
occur in the evening when they will not interfere with important daytime
activities (Forsberg & Wiklund, 1989). Alternatively, males may limit
mating activity to the late afternoon in order to reduce the likelihood of
immediate remating by the female with consequent sperm dilution
(Svard & Wiklund, 1988).

TERRITORIALITY BY HILLTOPPING MALES
While at the peaktops, males of the various Vanessa species appear to
be territorial, a point that is particularly clear in the case of V. atalanta.
Interactions between conspecifics of V. atalanta are far more complex
and long-lasting than are interactions between heterospecifics, a result
that matches that of Bitzer & Shaw (1979).

Scott ( 1986) has championed the alternative hypothesis that males are
merely inspecting passing butterflies in the attempt to locate potential
mates. This hypothesis has been critiqued elsewhere (Rutowski, 1984).
Here we note that the sex-identification hypothesis produces the predic-
tions that (1) male-male chases of any sort will follow much the same
pattern as male-female pursuits prior to courtship, (2) when two males
perch in the same area and then becomed involved in a chase both are
likely to return to the perching area together, and (3) when only one male
returns from a male-male chase, it is as likely to be one male as the other.
All three predictions fail. (1) Male-female chases are very different in
form from male-male interactions, which sometimes involve elaborate
ascending flights with hovering bouts. (2) On a high proportion of chases,
only one male returns to the perching area, not both. (3) When only one
male returns, it is far more likely to be the established resident male,
rather than a newcomer.

SPECIES DIFFERENCES IN SITE TENACITY
The congeners differed in their duration of residency at the peaktop
and their frequency of return. Interspecific variation in site tenacity and
the degree of territoriality was related to differences in the density of
conspecifics at the peak. V. annabella was the most tenacious of the three
species and was at lowest density on the hilltop; the relative lack of site
tenacity shown by individual V. cardui, in contrast, was associated with
the high density of conspecifics at the peak. V. atalanta exhibited
moderate site tenacity and intermediate densities. Differences in con-

29(1 2)' 1-10, 1990(91)

9

specific density were presumably associated with differences in the
frequency of rival encounters. Thus, the differences in site tenacity and
degree of territoriality of Vanessa butterflies may be caused by differ-
ences in how often individuals must deal with conspecific challenges at
the mate-encounter site.

INTERSPECIFIC CONFLICT

Because a diversity of hilltopping insects have converged on the same
site preferences (Alcock, 1983, 1984, 1987; Shields 1967), there is great
potential for interspecific encoutners at a peaktop. Perched Vanessa
males fly after most objects that pass by and thus spend much time and
(presumably) energy responding to heterospecific insects. A similar
response to moving objects has also been reported for other territorial
butterflies (e.g. Alcock & Gwynne, 1988; Alcock & O'Neill, 1986; Davies,
1978; Lederhouse, 1982; Scott, 1974). It seems likely that this behavior
is maintained because it enhances the rate at which resident males
identify potential mates and conspecific rivals (Bitzer & Shaw, 1979).
Alternatively, Bitzer & Shaw ( 1979) suggest that response to heterospecifics
may function to reduce the costs of interspecific interference at the hilltop
by chasing interspecific competitors from similar territory sites.

Because 4 species of Vanessa ( annahella,atalanta , cardui , virginiensis )
share the same central Arizona hilltops, the energy cost of heterospecific
aggression is potentially great. Have these butterflies evolved any
means to reduce non-productive interaction with congeneric males? Our
observations show that territorial males can quickly distinguish rival
conspecifics from congeneric males and have curbed the expense of
interspecific aggression. Interactions between congeneric males are
much shorter and much less aggressive than chases between conspecific
rivals. Males of V. cardui , the most abundant species, are most often
responsible for initiation and continuation of a congeneric chase and
hence appear to be the least species-discriminating. Because V. cardui
are so numerous at the study site, they interact with congeners less
frequently than do either V. atalanta or V. annabella. Therefore V.
cardui may be under less intense selection to be species-discriminating
than are the other species of Vanessa.

Species-specific perch sites may also have evolved to reduce congeneric
interference on the mating system. Perch site preferences of the three
species were segregated on the peak and were not associated with nectar
sources or oviposition sites. Turner (1990) has recently reported a
similar spatial partitioning of the hilltop among swallowtail butterflies.
However, rather than sorting themselves out horizontally, as in Vanessa,
the swallowtails patrolled species-specific zones above the surface of the
hilltop. Spatial separation on hilltops probably reduce the frequency of
interaction between congeneric males. Further work is needed to deter-
mine if distinctive spatial preferences persist in the absence of congeners.

10

J. Res. Lepid.

Literature Cited

Alcock, J. , 1 983 . Territoriality by hilltopping males of the great purple hair streak,
Atlides halesus (Lepidoptera, Lycaenidae): convergent evolution with a
pompilid wasp. Behav. Ecol. Sociobiol. 13: 57-62.

, 1984. Convergent evolution in perching and patrolling site preferences of

some hilltopping insects of the Sonoran Desert. Southw. Nat. 29: 475-480.

, 1985. Hilltopping in the nymphalid butterfly Chlosyne californica

(Lepidoptera). Amer. Midi. Nat. 113: 69-75.

, 1987. Leks and hilltopping in insects. J. Nat. Hist. 21: 319-328.

Alcock, J. & D. Gwynne, 1988. The mating system of Vanessa kershawi : males
defend landmark territories as mate encounter sites. J. Res. Lepid. 26: 116-
124.

Alcock, J. & K. M. O’Neill, 1986. Density-dependent mating tactics in the grey
hairstreak, Strymon melinus (Lepidoptera: Lycaenidae). J. Zool., Lond. 209:
105-113.

Bitzer, R. J. & K. C. Shaw, 1979. Territorial behavior of the red admiral, Vanessa
atalanta (L.) (Lepidoptera: Nymphalidae). J. Res. Lepid. 18: 36-49.

Davies, N. B., 1978. Territorial defence in the speckled wood butterfly ( Pararge
aegeria ): the resident always wins. Anim. Behav. 26: 138-147.

Dimock, T. E., 1978. Notes on the life cycle and natural history of Vanessa
annabella (Nymphalidae). J. Lepid. Soc. 32: 88-96.

Forsberg, J. & C. Wiklund, 1989. Mating in the afternoon: time-saving in
courtship and remating by females of a polyandrous butterfly Pieris napi L.
Behav. Ecol. Sociobiol. 25: 349-356.

Lederhouse, R. C., 1982. Territorial defense and lek behavior of the black
swallowtail butterfly, Papilio polyxenes. Behav. Ecol. Sociobiol. 10: 109-118.

Palm, C. A., 1980. Territories, leks, and mating in the west coast lady ( Vanessa
annabella). M.S. thesis, University of California, Davis, CA. 70 pp.

Rutowski, R. L., 1984. Sexual selection and the evolution of butterfly mating
behavior. J. Res. Lepid. 23: 125-142.

Scott, J. A., 1974. Mate-locating behavior of butterflies. Am. Midi. Nat. 91: 103-
117.

, 1986. The butterflies of North America. Stanford University Press, Stanford,

CA. 583 pp.

Shields, O., 1967. Hilltopping. J. Res. Lepid. 6: 69-178.

Svard, L. & C. Wiklund, 1988. Prolonged mating in the monarch butterfly
Danaus plexippus and nightfall as a cue for sperm transfer. Oikos 52: 351-354.

Thornhill, R. & J. Alcock, 1983. The evolution of insect mating systems. Harvard
University Press, Cambridge, Mass. 547 pp.

Turner, J. D., 1990. Vertical stratification of hilltopping behavior in swallowtail
butterflies (Papilionidae). J. Lepid. Soc. 44: 174-179.

Journal of Research on the Lepidoptera

29(1-2): 11-20, 1990(92)

Electrophoretic Studies in the Genus Melanargia
Meigen, 1828 (Lepidoptera: Satyridae)

Paola Mensi, Aldo Lattes, Luigi Cassulo, Roberta Cinti, Emilio Balletto*

Istituto di Zoologia, Universita di Genova, V. Balbi 5, 1-16126

Abstract. The phylogenetic relationships of the Euro-Mediterranean
species of Melanargia were studied using enzyme electrophoresis. A
dendrogram is presented representing the degree of relationship among
populations sampled. Results from the biochemical characterization
are only in partial agreement with those obtained from conventional
systematic study.

Systematic problems at the level of the specific differentiation between
M. galathea and M. lachesis are discussed on the basis of biochemical
data obtained by parapatric and allopatric populations. The opinion
that they represent two distinct, although recent, species is supported.

Introduction

The Satyrid genus Melanargia Meigen, 1828, the only genus of the
subfamily Melanargiinae Wheeler, 1903, is characterized by one vein in
the fore wings considerably swollen at the base. The “Marbled Whites”
can be easily recognized within the family, because of distinctive charac-
ters of adult morphology (Higgins, 1975). The wings are of medium size,
with creamy spots and bold black markings. The front legs are small in
both sexes, with hairless femora bearing a central groove. Sexes are
similar; males lack androconial scales. Male genitalia are characterized
by the uncus, short, very thick at basis and swollen on the central part
of ventral surface, and by the valvae, that apart from some apical spines,
lack any other appendages. Female genitalia show a genital plate that is
very sclerotized in front of the opening of the ductus bursae and little
sclerotized posteriorly. Two small sclerotized plates are situated on
either side of the external opening of the ductus bursae. The more
anteriorly placed of these plates is distally swollen in a wing-shaped
protuberance, which comes into contact with the sternite.

Egg morphology and larval stages are also characteristic. The eggs are
oval-shaped, taller than wide, ribbed and reticulate on the chorion and
with a petal-like micropylar area (Wagener, 1983). The ova are laid
during flight. The night-feeding larvae are light green or sandy in colour.
All feed upon Graminaceae (Higgins and Riley, 1980) and all hibernate
in winter as larvae. The pupae lie free among the grass stems.

Previous taxonomic studies have concentrated on morphological rela-
tionship drawn from wing colouration and marking. The genus was
divided into three subgenera: Melanargia Meigen, 1828, Argeformia
Verity, 1955, and Halimede Oberthiir & Houlbert, 1922.

The latter is a Central and East Asian group which will not be dealt
with in the present paper. Of the Euro-Mediterranean species, M.

* Dipartimento di Biologia Animale, Universita di Torino, V. Accademia Albertina 17, 1-10123

12

J. Res. Lepid.

galathea (Linne, 1758), M. russiae (Esper, 1784), M. larissa (Boisduvai,
1828), M. hylata (Menetries, 1832), M. titea (Klug, 1832) and M. syriaca
(Oberthiir, 1894) were arranged in the subgenus Melanargia whereas M.
arge,M. occitanica and M. ines (with its ssp .jahandiezi Oberthiir, 1920)
were arranged in the subgenus Argeformia. Within this same group M.
pherusa (Boisduvai, 1833) is variously considered either as a ssp. of M.
occitanica or as a distinct species.

On the basis of external morphology, species of the subgenus Melanargia
can be subdivided into two groups: one with M. russiae , M. hylata , M.
syriaca and M. larissa and another with M. galathea and M. titea. More
recently, however, Wagener (1983) suggested a different arrangement
based on the ultrastructure of ova. He recognized three species groups:
one including M. galathea and M. syriaca , another one with M. larissa,
M. titea, M. hylata and the Asiatic M. grumi (Standfuss, 1882) and a third
one with M. russiae, the Asiatic M. halimede (Menetries, 1858) and M.
parce (Staudinger, 1882).

An unresolved controversy regards the relationship between M. galathea
and M. lachesis (Hiibner, 1790) and their dubious specific designation.
The mostly parapatric distribution of M. galathea and M. lachesis in the
Western and Eastern sections of the Pyrenees and the morphology of
adult stages, ova and larvae have been extensively studied (Higgins,
1969; Tilley, 1983, 1986; Wagener, 1984; Mazel, 1986). Whereas in the
areas of sympatry in northern Spain the two phenotypes maintain their
distinctive characteristics (Gomez de Aizpurua, 1988), in the contact
zone of the Eastern Pyrenees also somewhat intermediate specimens are
found, equal to one fifth of the whole mixed populations (Mazel, 1986).

Since genitalic and other structural characters are very conservative in
this genus and show little, if any, variation between species, any attempt
to carry out a cladistic reconstruction of its phylogeny could only be based
on a very limited set of characters. The purpose of this work, therefore,
is to evaluate genetic relationship among the Euro-Mediterranean spe-
cies of Melanargia through their respective degree of electrophoretically
detectable enzyme similarity. As a consequence, it also attempts to
provide a yardstick for the rate of speciation processes within this genus
and, at the same time, to supply new reliable taxonomic characters in the
form of electromorph variants obtained by allozyme electrophoresis.

Methods and Materials
PREPARATION OF SAMPLES

Seventeen European and Turkish populations, for a total of 165 specimens,
were scored for enzyme variability. Populations and localities are listed in table
1; geographical distribution of sampled localities is depicted in Fig. 1.

Only field-collected adult males were employed for enzyme analysis. Their
wings were immediately removed with sharp scissors and the whole bodies were
frozen in liquid nitrogen while still alive. Specimens were kept in the same
medium for several days until their preparation for electrophoresis could take

29(1-2): 11 20. 1990(92)

13

Table 1. Populations sampled

Locality

Country, Region

n°

symbol

Paola

Italy, Calabria

12

A1

(arge)

Fuscaldo

Italy, Calabria

3

A2

( arge )

Capo Cervo

Italy, Liguria

14

01

{occitan ica)

Mont Leuze

France, Alpes Mar.

4

02

(occitanica)

Ficuzza

Italy, Sicily

10

Ph

(, pherusa )

Melilla

Morocco

4

In

lines)

Camlibel Dag.

Turkey, Sivas

10

LI

( iarissa )

Mazikiran Gee.

Turkey, Sivas

10

L2

( larissa )

Onagil

Turkey, Van

7

HI

( hylata ) *

Kocak

Turkey, Van

10

H2

(hylata) *

Asagi Kolbasi

Turkey, Bitlis

12

Sy

( syriaca )

Capo Cervo

Italy, Liguria

14

G1

(galathea)

Monte Amaro

Italy, Abruzzo

9

G2

(, galathea )

Kastamonu

Turkey, Bolu

10

G3

(galathea)

Campo

Spain, Huesca

18

G4

(galathea)

Tragacete

Spain, Cuenca

10

GL1

(lachesis)

Campo

Spain, Huesca

8

GL2

(laches is)

* These populations, once considered identical with ssp. karabagi, will soon be described as
a distinct subspecies by S. Wagener.

Figure 1 - Geographic distribution of sampling localities of Melanargia populations
studied. 1 : Melilla; 2: Tragacete ; 3: Campo; 4: Mont Leuze; 5: Capo Cervo;
6: Monte Amaro; 7: Fuscaldo; 8: Paola; 9: Ficuzza; 10: Kastamonu; 11:
Camlibel Daglari; 12: Mazikiran Gecidi; 13: Asagi Kolbasi; 14: Kogak; 14:
OnagiL

14

J. Res. Lepid.

place. Samples were prepared as follows. Individual butterflies were macerated,
using a tissue grinder, in 300 pi of a homogenizing solution containing 0.12 rnM
NADP and lOmM 2-mercaptoethanol. The homogenate was centrifuged at 13000
g for 15 min to obtain a clear supernatant. Samples were stored at -80°C until
electrophoresed.

ELECTROPHORESIS

Allozyme electrophoresis was carried out on Cellogel sheets at 4°C using buffer
systems and stains as described by Meera Khan (1971) and Richardson et al.
(1986). Thirteen gene-enzyme systems were studied: glycerol-3-phosphate dehy-
drogenase (E.C.1.1.1.8) ( GPD ), malate dehydrogenase (E.C. 1.1. 1.37) ( MDh ),
malic enzyme (E.C. 1.1. 1.40) (ME), isocitrate dehydrogenases (E.C. 1.1. 1.42), 2
loci ( IDh-1 , 2), 6-phosphogluconate dehydrogenase (E.C. 1.1. 1.44) (6PGD), aspar-
tate aminotransferase (E.C. 2. 6. 1.1) (AAt), hexokinase (E.C. 2. 7. 1.1) (HK), adeny-
late kinase (E.C. 2. 7. 4.3) (AK), phosphoglucomutases (E.C. 2. 7. 5.1), 2 loci (PGM-
1, 2), mannose phosphate isomerase (E.C. 5. 3. 1.8) ( MPI ), phosphoglucose isomer-
ase (E.C. 5. 3. 1.9) ( PGI ). Isozymes and alleles were designated numerically ac-
cording to their decreasing mobility rate.

STATISTICAL ANALYSIS

Average probability of interpopulation genetic identity and distance were
calculated by Nei’s I and D related indexes (jackknifed according to Mueller and
Ayala, 1982). A number of other more or less similar indexes were also computed:
Rogers’ distance, Wright’s modification of Rogers’ distance, Cavalli-Sforza’s &
Edwards’ arc and chord distances (for these indexes see Wright, 1978) and Hillis’
modification of Nei’s distance (1984). Several dendrograms resulting from the

Table 2. Allele frequencies

Al

A2

01

02

Ph

In

LI

L2

HI

H2

Sy

G1

G2

G3

G4

GL1

GL2

aGPD

120

0.05

100

1.00

1.00

1.00

1.00

1.00

1.00

1.00

0.95

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

MDH

120

0.40

0.35

100

0.07

0.60

0.65

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

90

1.00

0.83

0.93

1.00

0.94

1.00

65

0.17

0.06

ME

120

1.00

0.20

0.25

110

0.21

0.10

0.05

100

1.00

1.00

1.00

1.00

0.61

0.75

0.15

0.43

1.00

0.90

1.00

0.78

0.95

1.00

0.75

0.75

90

0.39

0.25

0.85

0.36

0.22

0.05

IDH-1

110

0.14

0.28

0.25

100

1.00

1.00

0.32

0.25

0.17

0.25

0.15

0.05

0.08

1.00

0.94

1.00

0.86

0.72

0.75

90

0.68

0.75

0.83

0.62

1.00

0.70

1.00

0.95

0.92

0.06

80

0.12

0.15

IDH-2 130 1.00 1.00
100

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

29( 1-2): 1 1-20, 1990(92)

15

A1

A2

01

02

Ph

In

LI

L2

HI

H2

Sy

G1

G2

G3

G4

GL1

GL2

6PGD

120

0.29

0.07

0.17

0.25

110

0.06

0.07

0.08

0.25

0.14

0.25

0.06

0.58

0.31

100

0.92

1.00

0.33

0.05

0.37

0.05

80

0.39

0.70

0.28

0.10

0.08

0.64

0.86

0.70

0.94

0.25

0.44

70

0.08

1.00

1.00

1.00

1.00

0.11

0.30

0.50

0.85

0.17

0.03

50

0.11

0.14

AAt

150

0.04

140

0.03

0.12

0.55

0.70

0.07

0.15

0.42

0.11

120

0.11

0.25

0.43

0.65

0.33

0.28

0.11

0.10

110

0.05

0.10

0.14

100

0.32

0.37

0.33

0.40

0.05

0.28

0.10

0.17

0.53

0.55

1.00

0.60

0.44

0.44

80

1.00

1.00

0.53

0.25

0.67

1.00

0.15

0.07

0.05

0.04

0.18

0.17

0.25

0.56

0.56

70

0.05

0.05

0.05

HK

120

0.04

110

1.00

1.00

1.00

1.00

0.10

0.08

100

0.83

1.00

1.00

0.90

1.00

1.00

0.87

1.00

1.00

1.00

1.00

1.00

1.00

90

0.17

AK

150

0.07

0.17

120

—

—

—

0.33

—

—

—

100

1.00

1.00

1.00

—

—

—

1.00

0.95

1.00

0.95

0.82

0.93

0.50

1.00

—

—

—

75

—

—

—

0.05

0.05

0.18

—

—

—

PGM-1 110

0.07

0.05

0.17

0.11

0.35

0.28

0.19

0.36

100

0.30

0.15

0.93

0.90

0.75

1.00

0.83

0.65

0.71

0.81

0.62

90

0.70

0.85

0.05

0.08

0.06

0.02

PGM-2100

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

80

0.12

1.00

0.87

75

1.00

1.00

60

0.88

1.00

0.12

MPI

150

—

0.17

0.10

145

—

—

—

0.30

0.17

0.08

0.20

0.22

0.19

0.05

—

—

—

135

—

—

—

0.20

0.33

0.08

0.10

0.50

0.28

0.44

0.11

—

—

—

130

—

—

—

0.10

0.08

0.05

0.50

0.37

0.83

—

—

—

120

0.18

—

—

—

0.40

0.28

0.33

0.30

0.15

—

—

—

110

0.32

—

—

—

—

—

—

100

0.96

0.83

0.14

—

-

0.05

0.33

0.10

0.15

—

—

—

90

0.04

0.17

0.25

—

—

—

0.08

0.35

—

—

—

80

0.11

—

—

—

—

—

PGI

175

0.03

160

0.12

0.03

0.33

0.05

145

0.73

0.70

0.45

0.14

0.54

0.11

0.39

0.25

0.37

0.62

100

1.00

1.00

0.10

0.20

1.00

0.45

0.28

0.10

0.12

0.17

0.14

0.11

0.70

0.56

0.44

0.31

75

0.10

0.10

0.57

0.60

0.75

0.17

0.53

0.11

0.12

0.19

0.06

60

1.00

0.31

55

0.07

40

0.30

0.12

0.14

0.06

— indicates that in the population scoring was not possible.

For some enzymes not all individuals of all populations were scored.

Table 3 Jackknifed average Nei Distances between populations

A1

A2

01

02

Ph

In

LI

L2

HI

H2

Sy

G1

G2

G3

G4

GL1

GL2

A1

—

0.00

0.07

0.11

0.09

0.21

0.10

0.13

0.11

0.12

0.08

0.10

0.10

0.09

0.09

0.07

0.08

A2

0.01

—

0.07

0.12

0.10

0.22

0.10

0.13

0.12

0.12

0.08

0.09

0.10

0.09

0.09

0.07

0.08

01

0.68

0.69

—

0.00

0.02

0.04

0.05

0.06

0.04

0.04

0.05

0.08

0.07

0.08

0.11

0.08

0.08

02

0.81

0.84

0.02

—

0.02

0.05

0.06

0.09

0.07

0.07

0.05

0.11

0.09

0.10

0.11

0.09

0.09

Ph

0.64

0.66

0.20

0.23

—

0.03

0.08

0.07

0.08

0.09

0.09

0.12

0.13

0.11

0.10

0.10

0.11

In

0.96

0.99

0.33

0.39

0.23

0.26

0.16

0.16

0.17

0.20

0.29

0.29

0.43

0.18

0.18

0.17

LI

0.77

0.77

0.58

0.55

0.63

1 .04

—

0.00

0.01

0.01

0.00

0.01

0.01

0.01

0.02

0.02

0.02

L2

1.00

1.00

0.72

0.78

0.69

0.92

0.13

—

0.01

0.02

0.01

0.02

0.02

0.03

0.03

0.03

0.03

HI

0.88

0.88

0.55

0.58

0.61

0.87

0.26

0.24

—

0.00

0.01

0.01

0.02

0.02

0.02

0.02

0.03

H2

0.89

0.88

0.49

0.53

0.62

0.83

0.29

0.31

0.07

—

0.01

0.02

0.02

0.03

0.03

0.03

0.04

Sy

0.76

0.76

0.53

0.48

0.71

0.97

0.15

0.28

0.20

0.18

—

0.01

0.01

0.02

0.03

0.02

0.02

G1

0.73

0.73

0.71

0.73

0.86

1.18

0.32

0.38

0.23

0.26

0.27

—

0.00

0.00

0.00

0.00

0.00

G2

0.78

0.78

0.69

0.65

0.83

1.18

0.28

0.38

0.31

0.37

0.23

0.07

—

0.00

0.00

0.00

0.00

G3

0.65

0.65

0.74

0.72

0.76

1.25

0.29

0.48

0.36

0r41

0.33

0.09

0.10

—

0.00

0.00

0.00

G4

0.61

0.61

0.75

0.75

0.72

1.05

0.30

0.43

0.30

0.33

0.32

0.07

0.08

0.03

—

0.00

0.00

GL1 0.58

0.58

0.61

0.67

0.70

1.00

0.32

0.47

0.33

0.35

0.25

0.10

0.11

0.09

0.10

—

0.00

GL2 0.63

0.62

0.58

0.64

0.76

0.98

0.31

0.47

0.36

0.39

0.24

0.12

0.08

0.08

0.10

0.03

—

Below the diagonal values of distances, associated standard errors above.

UPGMA and Fitch-Margoliash (1967) clustering of all such indexes were also
constructed.

Results

Allele frequencies are shown in Table 2. The overall number of alleles
detected at the 13 loci of all Melanargia species studied amounts to 61
(mean per locus 4.69, range 2-9). No locus proved monomorphic across
the whole sample.

The jackknifed Nei’s indexes of genetic similarity, I, and genetic
distance, D, were calculated on the basis of the 13 shared loci for pairwise
combinations of all populations studied (table 3). A dendrogram repre-
senting the degree of relationships between populations sampled is
depicted in fig. 2 for Nei’s indexes.

Rogers’ index, Wright’s modification of the same, Cavalli-Sforza’s arch
and chord measures, and Hillis’ average distances were also computed
(data not shown) and employed to generate clusters, that proved almost
identical with the dendrogram shown in fig. 2.

29(1-2): 11-20, 1990(92)

17

Figure 2 - Dendrogram based on jackknifed Nets identities

Species fall clearly into four groups, as follows:

1) Melanargia arge

2) M. occitanica , M. pherusa and M. ines

3) M. larissa , M. hylata and M. syriaca

4) M. galathea and M. lachesis

The populations of M. arge are the most differentiated with respect to
all other populations studied (D = 0.753, with respect to others). Rather
surprisingly, M. arge shows complete fixation at 8 loci (61% of total).
Remarkably, 2 of these (PGM-2 and IDh-2) proved alternative with
respect to the other Melanargia studied.

In contrast, no private locus was found to characterizes group (2).
Within this group M. occitanica and M. pherusa are more closely related
to each other (D = 0.215) than they are with respect toM. ines (D = 0.314).
This relatively higher value of D agrees with the traditional separation
between M. occitanica and M. ines : in this case genetic divergence
correlates with morphological and biological differences (in fact in Spain
M. occitanica and M. ines fly together without interbreeding).

Groups (3) and (4), corresponding to subgenus Melanargia , cluster at
a value of D = 0.334 and share a diagnostic locus (at PGM2 they fix allele
100 , absent in other groups).

Within group (3) the closest electrophoretic similarity is shown by the
two populations of M. hylata.

18

J. Res . Lepid.

Interestingly enough, electromorph variation ranks M. syriaca within
species of group 3, whereas on the basis of its egg morphology (Wagener,
1983), it should be expected to be more similar to M. galathea.

A high degree of genetic similarity is demonstrated by the analysis of
group 4, even between geographically very distant populations: for
example D values between the Turkish and the Spanish sample of M.
galathea (0.032) is even lower than that found between the Italian
populations of Capo Cervo (N.W. Italy) and Monte Amaro (C. Italy) (D =
0.074). Even though these values appear very low, it may be relevant to
observe that both were proved to differ significantly from zero by the
jackknife procedure (Tab. 3 and Fig. 2).

Discussion

Results from the biochemical characterization of entities belonging to
the genus Melanargia is only in partial agreement with those obtained
from other more conventional systematic studies. The main branching
point of the dendrogram shown in Fig. 2 is between M. arge and the other
groups of Melanargia studied (D - 0.753). This accounts for the isolated
position of the populations of M. arge in the dendrogram. However, the
group including M. occitanica, M. pherusa and M. ines branches at D =
0.739, which is only slightly below the highest level of genetic distance.
Therefore these biochemical-genetic criteria support a taxonomical ar-
rangement different from the current subgeneric differentiation of
Melanargia and Arge formia.

Since M. arge proved highly monomorphic over the whole sample of
enzymes investigated (H = 0.039), values for average heterozygosity vary
depending on whether or not data for M. arge are included (H = 0.124
with, or H - 0.210 without M. arge). The first of these values is in better
agreement with average values found for a number of invertebrate
species (H = 0.134, Ayala et ah, 1972). The exceedingly low heterozygos-
ity of M. arge could be the effect of a recent bottle-neck. This species, in
fact, may well have survived glaciations even in a single refuge area in
the South of the Italian peninsula (or perhaps in Sicily), where it
remained restricted at least since the last ice age.

Genetic differences between M. galathea and M. lachesis are relatively
small, but within the range previously reported for closely related species
of Lepidoptera (Geiger, 1988).

External (phenetic and morphological) characteristics of the two forms
are notably different, M. lachesis being larger and with reduced black
markings on the dorsal surface of the wings. Their distribution is broadly
parapatric and, apart from a few small areas of sympatry , in the Iberian
Peninsula their biotopes are sharply delineated by altitude and expo-
sure. M. galathea is a sub-nemoral species (Balletto and Kudrna, 1985)
diffused throughout Europe. Its typical biotope is in the outskirts, or the
clearings of, mesophilous woods, at elevations generally not exceeding
800-900 m, where the vegetation includes an abundant growth of Phleum

29(1-2): 11-20, 1990(92)

19

pratense , the main foodplant of its larvae. In SE France and the southern
belt of the Pyrenees M. galathea is replaced by M. lachesis. The latter
species flies generally in dry and warm biotopes with a vegetation of a
Mediterranean or sub-Mediterranean type, where its larval food plant,
Lamarkia aurea , grows.

Notwithstanding their somewhat different environmental and cli-
matic requirements, both species are known to live in sympatry in some
fifteen biotopes of the antipyrenaic Spanish districts of Huesca, Lerida
and Gerona and in only one biotope in Southern France (Col de Gres, in
the Aude; Mazel, 1988), between the mouth of the River Rhone and the
Pyrenees. In either case, the areas of sympatry are extremely small, in
relation at least to the extension of the parapatric populations.

Interestingly, in the contact area of southern France some roughly
intermediate individuals have been described, which up until now
remain unreported from the Spanish side of the Pyrenees. Even though
this was not experimentally demonstrated, such specimens are often
supposed to be hybrids (Higgins, 1969; Mazel, 1986).

We studied two sympatric populations at a single Spanish biotope,
whereas the French zone of sympatry between the two species could not
yet be investigated. We found only typical individuals, easily classifiable
either as M. galathea or M. lachesis. Data reveal that these sympatric
populations are not significantly more similar to each other (D = 0.098)
than to aliopatric samples (D = 0. 103), although in any case no diagnostic
locus could be found. This finding suggests that little, if any, introgres-
sion between the two species may exist, and is in agreement with the
hypothesis that there is a separation at species level between M. lachesis
and M. galathea.

A rough dating of reproductive isolation between the two taxa is
provided by time divergence estimates based on D values. Nef s estimates
(Nei, 1972) to match electrophoretic data to the molecular clock hypoth-
esis have been much questioned (see Mindell et ah, 1990); we tentatively
apply Nef s calibration, as it has often been shown to match very well with
current geological and biogeographical knowledge (Mensi et ah. 1987).

Rased on these estimates, in the evolutionary history of genus
Meianargia.M. galathea and M. lachesis split roughly 0.5 Myr ago. In the
light of such a recent separation, it is not surprising that in nature they
may perhaps be still capable of producing hybrids, although exception-
ally.

Acknowledgements . We are indebted to Dr. P. S. Wagener for identifying our
Turkish material of Melanargia and for having supplied many useful comments.
This work has been financially supported by the Italian Ministry for Scientific
Research under *40%’ and '6G%? funding programs, as well as by the Italian
National Research Council (CNR) under “Special project on microspeciation
mechanisms”.

20

J. Res. Lepid.

Literature Cited

Ayala, F. J., J. R. Powell, M. L. Tracey, C. A. Mourao and S. Perez-Salas, 1972.
Enzyme variability in the Drosophila willistonii group. IV. Genic variation in
natural populations of Drosophila willistonii. Genetics 70: 113-139.

Balletto, E. and O. Kudrna, 1985. Some aspects of the conservation of the
butterflies (Lepidoptera: Papilionoidea) in Italy, with recommendations for
future strategy. Boll. Soc. ent. ital. 117:39-59.

Fitch, W. M. and E. Margoliash, 1967. Construction of phylogenetic trees. Science
155: 279-284.

Geiger, H. , 1988. Enzyme electrophoresis and interspecific hybridization in Pieridae
(Lepidoptera) - The case for enzyme electrophoresis. J. Res. Lepid. 26: 64-72.
Gomez de Aizpurua, C., 1988. Atlas provisional de los Lepidopteros de la zona norte.

III. Lepidoptera rhopalocera. Serv. centr. Publ. Gov. Vasco, Vitoria.

Higgins, L. G., 1969. Observations sur les Melanargia dans le Midi de la France.
Alexanor 6: 85-90.

— — ■, 1975. The classification of European butterflies. W. Collins & S., London.
Higgins, L. G. and N. D. Riley, 1980. A field guide to the butterflies of Britain and
Europe. 4th edition. Collins, London.

Hillis, D. M., 1984. Misuse and Modification ofNei’s Genetic Distance. Syst. Zool.
33: 238-240.

Mazel, R., 1986. Contacts parapatriques entre Melanargia galathea L. et M.

lachesis Hiibner - Lepidoptera - Satyridae. Nota Lepid. 9: 81-91.

Meera Khan, P., 1971. Enzyme electrophoresis on cellulose acetate gel: zimogram
patterns in man-mouse and man-chinese hamster somatic cell hybrids. Arch.
Biochem. Biophys. 145: 470-483.

Mensi, P., A. Lattes, S. Salvidio and E. Balletto, 1987. Taxonomy, evolutionary

biology and biogeography of SW European Polyommatus coridon (Lepidoptera,
Lycaenidae). Zool. J. Linn. Soc. 93: 259-271.

Mindell, D. P., J. W. Sites and D. Graur, 1990. Mode of allozyme evolution:
Increased genetic distance associated with speciation events. J. Evol. Biol. 3:
125-131.

Mueller, L. D. and F. J. Ayala, 1982. Estimation and interpretation of genetic
distance in empirical studies. Genet. Res., Cambridge, 40: 127-137.

Nei, M., 1972. Genetic distance between populations. Amer. Nat. 106: 283-292.
Richardson, B. J., P. R. BAVERSTOCKand M. Adams, 1986. Allozyme Electrophoresis.
Academic Press, Sydney.

Tilley, R. J. D., 1983. Rearing Melanargia galathea (L.) and M. lachesis (Hiibner)
(Lepidoptera: Satyridae). Entomol. Gaz. 34: 9-11.

, 1986. Melanargia galathea galathea (L.) and M. galathea lachesis (Hiibner)

in the south of France (Lepidoptera: Satyridae). Entomol. Gaz. 37: 1-5.
Wagener, S., 1983. Struktur und Skulptur der Eihiillen einiger Melanargia-Arten
(Lepidoptera, Satyridae). Andrias, 3: 73-96.

, 1984. Melanargia lachesis Hiibner 1790 est-elle une espece different de

Melanargia galathea Linnaeus 1758, oui ou non ? Nota lepid. 7: 375-386.
Wright, S., 1978. Evolution and the Genetics of Populations. Vol. 4: Variability
within and among Natural Populations. Univ. Chicago Press, Chicago.

Journal of Research on the Lepidoptera

29(l-2):21-32, 1990(91)

Hybridization of the Brazilian Papilio
(Pyrrhosticta) (Section V) with the North American
Papilio (Pterourus) (Section III)

J. Mark Seriber1

Robert C. Lederhouse1

and

Keith S. Brown Jr.2

Abstract. Male Papilio (. Pyrrhosticta ) scamander Boisd, from
Campinas, Brazil were hand-paired to virgin P. (Pterourus) glaucus L.
and P. (. Pterourus ) palamedes Drury females. Egg viability was less
than 10% for two P. glaucus x scamander crosses and 18% for two
palamedes x scamander crosses, glaucus x scamander larvae
developed to pupation on sweetbay, tuliptree and cucumbertree
(Magnoliaceae) and black cherry (Rosaceae) but died on redbay
(Lauraceae). P. palamedes x scamander larvae developed to the last
instar on redbay and camphortree (Lauraceae), but none pupated
successfully. One female and 3 male glaucus x scamander adults
emerged. Backcrosses were unsuccessful in producing viable eggs.
Similarities of P. scamander adults, larvae and food plant use with the
Mexican P. garamus Geyer are intriguing. Crosses of Brazilian P.

( Pyrrhosticta ) cleotas Gray with P. glaucus gave a higher egg fertility
(68%) and the larvae fed well on Talauma ovata and Michelia champaca
(Magnoliaceae), though none molted to the second instar. These results
suggest a closer biological relationship between the five species men-
tioned than their current placement in two (sub) genera and four
species groups would indicate.

Introduction

Of the 560+ species of Papilionidae (Munroe 1961, Hancock 1983,
Collins and Morris 1985, Miller 1987), only 62 species are reported to use
hosts from more than one plant family, and only 23 species use more than
3 food plant families (Seriber 1984). The most polyphagous of these
species are from the North American Papilio glaucus group of Section
III and the South American Papilio scamander group of Section V
(Munroe 1961, Seriber 1988). The behavioral and physiological mecha-
nisms by which this atypically broad feeding capability is achieved in P.
glaucus and P. scamander species groups is the object of general studies
in our laboratories. We are also interested in the systematic relation-
ships of Section III and Section V species (Seriber et al. 1991).

It has been long recognized that the North American glaucus species

^ept. of Entomology, Michigan State University, East Lansing, MI 48824.

2 Departmento de Zoologia, Institute de Biologia, Universidade Estadual de Campinas, C.P.
6109, Campinas, Sao Paulo, Brazil 13.081.

22

J, Res. Lepid.

group (P. glaucus L .,P. alexiares Hoppfer?P. rutulus Lucas, P. eurymedon
Lucas, and P . multicaudatus Kirby) and. the troilus species group (. P .
troilus L .,P.pilumnus BoisduvaL and P, pulamedes Drury) are closely
related (Forbes 1951, Munroe 1981, Brower 1959a). A number of strong
arguments can be made for considering thes eglaucus and troilus groups
as sister taxa (Hancock 1983, Scriber et at 1991, Hagen and Scriber
1991), and both are included in Section III. Numerous interspecific hand-
pairings (within Section III, Munroe 1981) have produced viable hybrids
between many taxa of both groups (Scriber 1982, 1988, Scriber et al,
1988 [90], 1990a, Scriber and Lederhou.se 1988[89], West and Clarke
1987188]), Hancock (1983) identified 3 Central and South American
species groups (. scamander , zagreus , and homerus ) comprising Section ¥
as sister taxa to the glaucus and troilus groups. Among criteria for
linking the Section III and Section V as sister taxa were the similarities
in larval host families. In addition, the Eutaceae, Lauraceae, and
Magnoliaceae (used by both Sections III and ¥) share secondary chemi-
cals with the South American Hernandiaceae and Berberidaceae used by
Section ¥ (Scriber et ah 1991). Other characters linking the two sections
include a “mature, solitary, smooth, green larva with metathoracic
eyespots and, usually, a transverse dark band on the first abdominal
segment” (Hancock, 1983). A character differentiating the two sections
is an X- shaped saddle in Section ¥ (= subgenus Pyrrhosticta of Hancock's
classification) versus the smooth green mature larvae without the saddle
in Section III (= subgenus Pterourus of Hancock’s classification). Hancock
(1983) suggested that Section ¥ represents a South American offshoot of
Section III.

In this study we produced viable hybrids between the South American
P. scamander and P, cleotas (of Section ¥) and species of the North
American Section III. We illustrate the adult and larval color patterns
of resulting hybrids and report food plant utilization abilities of the
hybrids.

Methods

Adults of Papilio scamander were collected near Campinas, Brazil in late
March 1988 and brought in envelopes to East Lansing, Michigan for oviposition
or hand-pairing with 1937 lab-reared virgin females of Papilio glaucus and
Papilio palamedes and P. troilus . Larvae of P. scamander , collected in
Campinas, Brazil from the Asian ornamental Michelia champaca (Magnoliaceae)
used in street arborization, were also brought to Michigan for pupation. Two
subsequent female P. scamander adults were hand paired to P.g. glaucus males.
Hybrid adult males (P. glaucus x P. scamander ) were backcrossed to virgin P.
glaucus . Laboratory-reared males were fed a mixture of 1 part honey to 4 parts
water supplemented with amino acids and electrolytes for at least 3 days prior
to handpairing (Lederhouse et ah 1990). Field-collected and laboratory-mated
females were set up in plastic boxes (10 cm x 20 cm x 27 cm) with a sprig of
tuliptree, Liriodendron tulipifera ; sweetbay, Magnolia virginiana ; Michelia
champaca ; black cherry, Primus serotina ; and/or redbay, Per sea borhonia under

29(l-2):21-32, 1990(91)

23

saturated humidity. The boxes were placed 0.7 - 1.0 m from 100 watt incandes-
cent bulbs under a 6 hr photoperiod followed by 6 hr of darkness. Females were
fed a mixture of 1 part honey to 4 parts water at least once daily. Females were
allowed to oviposit until death. Eggs were collected and counted at 2-day
intervals except on weekends. Larvae were removed as they hatched, and the
remaining eggs were monitored for 10 days after the last larva hatched. Egg
viability was the proportion of the total eggs laid that hatched. Eggs that changed
color but did not hatch were scored as fertile. Using fine camel-hair brushes, first
instar larvae (neonates) were gently placed on fresh leaves of various potential
host plants for bioassays of consumption and survival. Leaf moisture was
maintained using aquapics, and fresh leaves were provided 3 times per week
throughout larval development.

Results

PAIRINGS AND FOODPLANT USE

Table 1 . Lab pairings of female Papilio glaucus glaucus (North American,
Section III) with male Papilio scamandar (South American, Section V).

Mother

number

Morph

Total eggs

Number of
larvae

Additional
fertile eggs

% fertile

% viable

5181

D

74

2

3

6.8

2.7

5182

D

191

9

3

6.3

4.7

5187

D

84

0

0

0.0

0.0

5188

D

10

0

0

0.0

0.0

5210

D

9

0

0

0.0

0.0

5939

Y

0

0

0

—

■ — ■

6226

D

1

0

0

0.0

0.0

6835

D

24

0

0

0.0

0.0

6840

D

0

0

0

—

—

6841

Y

0

0

0

—

—

6866

D

0

0

0

—

6868

Y

0

0

0

—

—

Twelve hand-pairings of Papilio glaucus glaucus females with P.
scamander males were made. Although females from seven pairings
produced some eggs, larvae were obtained from only two of the females
(Table 1), plus one of two P. glaucus female x P. cleotas male pairings.
The fertility and viability of the eggs were very low (less than 10%) in P.
scamander crosses but higher (68%) in one P. cleotas cross. Zero fertility
was observed in two additional hybrid pairings (P. scamander female x
P. g. glaucus male, Fig la) and in two backcrosses of P.g. glaucus females
with hybrid males (Table 2), as well as all six pairings between scamander
or glaucus and Brazilian species in Section IV ( astyalus , anchisiades,
torquatus, and hectorides).

Of the two hybrid larvae produced in brood 5181 ( P. glaucus x P.
scamander ), both survived the first instar and to pupation on Magnolia

24

J. Res. Lepid.

Table 2. Hybrid and Backcross Pairings with Papilio glaucus and Papilio

scamander.

Code

P. scamander

X

P. g. glaucus

= no larvae

5244

(female)

(male)

(hybrid)

Code

P. scamander

X

P. g. glaucus

= no spermatophore

5230

(female)

(male)

(hybrid)

Code

P. g. glaucus

X

P. g. glaucus x P. scamander

= no spermatophore

5458

(female)

(hybrid male from 5181)

(backcross)

Code

P. g. glaucus

X

P. g. glaucus x P. scamander

= no spermatophore

5599

(female)

(hybrid male from 5181)

(backcross)

virginiana. The neonate larval survival of nine individuals from brood
5182 was: 100% (n=l) on Prunus serotina (Rosaceae), 100% (n=l) on
Liriodendron tulipifera (Magnoliaceae), 50% (n=2) on Magnolia
acuminata and 67% (n=3) on Magnolia virginiana , and 0% (n=2) on
Per sea horbonia (Lauraceae). Larvae from 5182 were reared to pupation
on each of the first three plants (Table 3).

Neonate larvae of cleotas x glaucus 6210 (n=26) fed and grew very well
on Talauma ovata , the only native Brazilian Magnoliaceae, and moder-
ately well on the introduced Michelia champaca, in the same family.
They rested on, but hardly fed on, Cryptocarya aschersoniana (Lauraceae,

Table 3. Hybrid ( Papilio glaucus x P. scamander from brood 5181 and 5182)
larval survival/development on five plant species. All food plants tested except
Persea horbonia are satisfactory hosts of P. glaucus. None are encountered
naturally by P. scamander.

Foodplant

Genus Species

Family

n

Survival

through 1st
stadia

Survival to

pupation

Magnolia virginiana
(sweetbay)

MAGNOLIACEAE

5

80%

Yes

Magnolia acuminata
(cucumbertree)

MAGNOLIACEAE

2

50%

Yes

Liriodendron tulipfera
(tuliptree)

MAGNOLIACEAE

1

100%

Yes

Prunus serotina
(black cherry)

ROSACEAE

1

100%

Yes

Persea horbonia
(redbay)

LAURACEAE

2

0%

No

29(l-2):21-32, 1990(91)

25

used by both scamander and cleotas in the field). None reached the
second instar.

Two pairings of Papilio palamedes virgin females with P. scamander
males were made. Both of these pairings produced viable hybrids. It is
noteworthy that fertility and viability of the eggs were greater than
observed for the glaucus x scamander hybrids (Table 4). Of 30' hybrid
larvae set up from brood 5194, first instar survival was observed as: 70%
(n=10) on Persea borbonia ; 57% (n=7) on Cinnamomum camphora ; 25%
(n=8) on M, virginiana , 0% (n=4) on L. tulipifera ; and 0% (n=l) on M.
acuminata. Hybrid survival for four larvae from brood 5198 was 33%
(n=3) on Persea borbonia and 0% (n=l) on M. acuminata . The two species
of Lauraceae (redbay and camphorlree) supported vigorous hybrid larval
growth until the final instar, but none of these large larvae were able to
successfully pupate (Table 5, Fig. 2). A pairing of a P. palamedes male
with a P. astyalus female (Section IV), gave only 7 eggs, none of which
hatched.

Table 4. Lab pairings of female Papiiio palamedes (North American, Section III)
with male Papilio scamandar (South American, Section V).

Mother

number

Total eggs

Number of
larvae

Number of
additional
fertile eggs

% fertile

% viable

5194

107

31

14

26.9

18.6

5198

22

4

4

36.4

18.2

Table 5. Hybrid (P. palamedes x P. scamander from brood 51 94 and 51 98)
larval survival/development on five plant species. Pure F. palamedes develop
successfully on the two lauraceous host but die on the Magnoiiaceae.

Foodplant

Genus Species

Family

n

Survival
through 1st
stadia

Survival to
pupation

Magnolia virginiana

MAGNOLIACEAE

8

25%

No

(sweetbay)

Magnolia acuminata

MAGNOLIACEAE

2

0%

No

(cucumbertree)

Liriodendron tulipfera

MAGNOLIACEAE

4

0%

No

(tuliptree)

Persea borbonia
(redbay)

LAURACEAE

13

62% Huge larvae

(but unsuccessful)

Cinnamomum camphora
(camphortree)

LAURACEAE

7

57% Huge larvae

(but unsuccessful)

Fig. 1 a) A Papilio scamander female (top, from Campinas, Brazil) in copulation
with a P. g. glaucus male (bottom, from Clinton County, Michigan, USA),
b) A Papilio troilus female (top, from Allegan County, Michigan, USA) in
copulation with a P. scamander male (bottom, from Campinas, Brazil), c)
Dorsal view of a hybrid male adult reared on Prunes serotina from a hand-
pairing of a dark morph Papilio glaucus glaucus female with a P.
scamander male (from Campinas, Brazil), d) Ventral view of another
hybrid male adult reared on Liriodendron tulipifera of the same pairing
(#5182) as the hybrid figured in 1 c. e) Dorsal view of a female hybrid adult
reared on Magnolia virginiana from pairing (#51 81 ). f) Ventral view of the
same hybrid female adult.

4:wr ,■

mmm

— * - *„****

28

J. Res. Lepid.

Table 6. Larval survival and development of Papiiio gammas * on four plant

species.

Foodplant

Genus Species

Family

n

Survival
through 1st
stadia

Survival to
pupation

Magnolia virginiana
(sweetbay)

MAGNOLIACEAE

2

50%

Yes

Liriodendron tulipfera
(tuliptree)

MAGNOLIACEAE

4

50%

Yes

Persea borbonia
(redbay)

LAURACEAE

4

75%

Yes

Sassafras aibidum
(sassafras)

LAURACEAE

5

80%

Yes

A virgin Papiiio troilus female was successfully hand-paired with a P.
scarnander male (Fig. lb). However, no eggs were obtained from this
female. Two additional pairings of P. g. glaucus females from Michigan
pupae with field-captured P. scarnander males were made in Brazil in
January, 1989. Only one infertile egg was obtained.

Papiiio garamas from Mexico is another species from Section ¥ with
larvae strikingly similar to P. scarnander and P. cleotas and with the
ability to feed on both the Magnoliaceae and Lauraceae (Table 6). Five
hand-pairings of virgin P. g. glaucus and P. garamas males were made
(6228-8232). However, no eggs were obtained from these females.

Fig. 2 a) A second instar hybrid larva reared on Magnolia acuminata of a female
Papiiio glaucus glaucus and male P. scarnander pairing (#5182). b) A
fourth instar hybrid larva reared on Magnolia virginiana of a female P. g.
glaucus x P. scarnander pairing (#5182). c) A fifth instar hybrid larva of
a P. g. glaucus x P. scarnander pairing (#5181). Note the intermediate
thoracic “eyes” and the intermediate mid-abdominal X-saddle compared to
the parental types below (Fig. 2e, 2f). d) A prepupal hybrid of a P. g. glaucus
x P. scarnander pairing (#5182). e) A fifth instar P. scarnander larva
collected in Campinas, Brazil in late March, 1988. Note the abdominal X-
saddle and the thoracic band of “eyes”, f) A fifth instar P. glaucus larva.
Note the smooth green body and cohesive thoracic “eye”.

Fig. 3 a) A first instar hybrid larva of a Papiiio palamedes female x P. scarnander
male pairing (#5198). b) A second instar hybrid larva of a P. palamedes
female x P. scarnander male pairing (#51 94). c) A fourth instar hybrid larva
of a P. palamedes x P. scarnander pairing (#51 94). d) A fifth instar hybrid
larva of a P. palamedes x F. scarnander pairing (#5194). e) A fifth instar
P. palamedes larva, f) A fifth instar F. garamas larva (#5988).

29(l-2):21-32, 1990(91)

29

LARVAL AND ADULT COLOR COMPARISONS

An adult hybrid (. P . g. glaucus x P. scamander ) male is illustrated in
dorsal (Fig. 1c) and ventral view (Fig Id). The hybrid larvae of female P.
g. glaucus with maleP, scamander are illustrated for the second, fourth,
and fifth instar (Fig. 2a, b, c). The hybrid prepupal stage (Fig. 2d) show
remnants of the P. scamander cross saddle (Fig. 2e), which is lacking in
P glaucus (Fig. 2f).

The hybrid larvae of female P. palamedes with male P. scamander are
illustrated for the first (Fig. 3a) second (Fig. 3b), fourth (Fig 3c), and fifth
instars. A fifth instar P . palamedes is shown for comparison (Fig. 3e).
The fifth instar of the Mexican P. garamas (also of Section V) is
illustrated for comparison (Fig. 3f, see also Fig 2e).

Discussion

Interpretation of the phylogenetic significance of hybrid incompatibili-
ties is difficult (Ae 1979, Collins 1984, Lorkovic 1988, Geiger 1988, Coyne
and Orr 1989). Since no clear standard of reference exists despite Ae’s
(1979) fine work, our evaluation must be largely descriptive. Incompat-
ibility in presumed regulatory genes occurs predominantly at develop-
mental stages where new groups of genes begin to interact. Our experi-
ence with Papilio and the literature on intergeneric crosses (Peigler
1978, Carr 1984) suggests that the genes for producing a larva remain
relatively unmodified. However, genes affecting developmental stages
such as pupation or adult emergence are more likely modified between
species. The hybrid compatibility between species believed to be closely
related, such as Papilio troilus and P. palamedes , may be quite low
because of relatively small changes in key genes. This further compli-
cates the evaluation of hybridization data.

Nevertheless, the degree of genetic compatibility of Papilio scamander
and P . cleotas of the subgenus Pyrrhosticta (Section V) with P. glaucus
and P. palamedes of the subgenus Pterourus (Section III) was surprising.
These insect species are currently separated in distribution by about
5000 miles and by the Central American isthmus. Although these two
sections are considered to be sister taxa (Munroe 1961, Hancock 1983),
it is still significant that both adult males and females were obtained for
P g« glaucus xP. scamander pairings. This is remarkable since relatively
few hybrid females of interspecific pairings within the P. glaucus species
group itself are able to survive to adulthood (with the exception of P.
glaucus xP. alexiares pairings; West and Clarke 1 98 7 [88], Scriber et al.
1988, 1990). Pairings of P. glaucus and P. pilumnus have similarly
produced only male hybrid offspring (Scriber and Lederhouse 1988[89]).

The inabilitj^ for any hybrids of P. palamedes x P. scamander to
successfully pupate suggests a developmental incompatibility. The
fertility and viability of palamedes x scamander eggs were considerably
better than glaucus x scamander pairings (Tables 1 and 4). Larval
growth was excellent until the final instar, although none of these
heal thy -looking hybrids could successfully pupate (see Figs. 3c, 3d).

30

J. Res. Lepid.

Hybridization of glaucus with cleotas gave high initial juvenile
viability; this needs to be repeated under more favorable conditions,
perhaps on small growing plants to avoid the likely phenolic oxidation
suspected in excised Talauma ovata leaves. The relatively high biologi-
cal compatibility between Section III and Section V Papilio species does
not support their maintenance in separate genera or even subgenera
(. Pterourus , Pyrrhosticta ), though they can be conveniently separated
from the species in Section IV ( Heraclides ) with which they share very
low fertility and viability upon hybridization (this work and Ae, 1979).

The foodplant utilization abilities of hybrid larvae were not especially
surprising since P. scamander feeds commonly on both the Magnoliaceae
and Lauraceae (Jordan 1907, Scriber 1984, Ruszczyk 1986). The P.
troilus species group tend to be specialized on plants of the Lauraceae,
and hybrids of palamedes x scamander survived and grew very well on
both redbay, Persea borbonia, and camphortree, Cinnamomum camphora
(both of the Lauraceae). In contrast, the three Magnoliaceae foodplants
(. Magnolia virginiana, M. acuminata , and Liriodendron tulipifera ) ap-
peared to be poorer foods with no first instar survival on the latter two
plant species (see Table 5).

In contrast, the glaucus x scamander hybrids survived and grew well
on the three Magnoliaceae species and black cherry, Prunus serotina
(Rosaceae), but none survived on the Lauraceae (redbay, see Table 3). It
is interesting that hybrids of P. glaucus x P. scamander would initiate
feeding and survive on black cherry whereas none of the hybrid larvae
(n=27) of P. glaucus x P. pilumnus could do so (Scriber and Lederhouse
1988[89]). Although hybrids of glaucus x pilumnus could use all
Magnoliaceae tested (Scriber and Lederhouse 1988[89]), hybrids of
palamedes x scamander could use neither tuliptree nor cucumbertree of
the Magnoliaceae and did poorly on sweetbay (Table 5). The scamander
population in Campinas feeds on at least 4 genera of Lauraceae ( Persea
gratissima , Ocotea corymbosa , Chrytocarya aschersoniana, Cinnamomum
zeylanicum , and two Magnoliaceae ( Michelia champaca and Talauma
ovata ) and accepts and may regularly use an introduced Sterculiaceae,
Brachychiton (see Ruszczyk, 1986). In southern Brazil, cleotas uses T.
ovata , C. aschersoniana , Persea rigida , and Ocotea species, thus also
combining both families in its diet.

These differential abilities support the concept that major phytochemi-
cal differences exist between the Magnoliaceae and the Lauraceae
(Scriber 1986). Even within these plant families differences in feeding
behavior and detoxification ability exist within certain Papilio taxa for
different plant species (Hagen 1986, Scriber et al. 1991, Nitao e£aZ. 1991).
The fact that any North American hosts were acceptable and suitable for
the P. scamander hybrids is intriguing. Additional studies to assess the
extent of the feeding and growth capabilities of the generalized P.
glaucus larvae on an array of South American hosts, and the reciprocal
study of the South American generalist (P. scamander ) on North
American plants are planned.

29(l-2):21-32, 1990(91)

31

Additional interspecific, intergroup, and inter-sectional pairings will
help resolve the systematic and phylogenetic relationships of the various
taxa in Sections III and ¥ of the genus Papilio . A generally agreed upon
phylogeny would greatly improve our understanding of the evolution of
generalized feeding habits (e.g. upon several families of phytochemically
diverse plant species) (Miller 1987b). For example, it will be interesting
to know if the same general (e.g. mixed function oxidases) or specific
detoxification enzymes are used by both P. glaucus and P. searnander or
cleotas on Magnoliaceae, and similarly for P. palamedes and P.
scamander or cleotas on the Lauraceae. The same could be said for the
other Section III and Section ¥ Lauraceae/Magnoliaceae feeders in
North, South and Central America (e.g. Papilio gammas , Table 8).

Acknowledgements. This research was supported by the Michigan State Univer-
sity College of Natural Sciences and the Agricultural Experiment Station (Project
8051 and 8072), the National Science Foundation (BSR 8718448), and USDA
grants #85CRCR- 1-1598 and #87-CRGR- 1-2851. We are also thankful for support
of the National Science Foundation and the Brazilian CNPq, FAPESP, and
UNICAMP for travel support and a research fellowship (CNPq to KB). We would
like to thank Ana Beatriz Barrais de Morais, Robert Dowell, Bruce Giebink, Robert
Hagen, James Nitao, David Robacker, Alexandre Ruszczyk and William Warfield
for their discussion and/or assistance. Comments by two anonymous reviewers
were particularly helpful in improving this manuscript.

Literature Cited

Ae, S.A. 1979. The phylogeny of some Papilio species based on interspecific
hybridization data. Syst. Ent. 4: 1-18.

Brower, UP. 1959. Speciation in butterflies of the P. glaucus group. I.

Morphological relationships and hybridization. Evolution 13:40-03.

Collins, M.M. 1984. Genetics and ecology of a hybrid zone in Hyalophora
(Lepidoptera: Satumiidae). Univ. of Cal. Publ. Entomol. 104: 1-93.

Collins, M.M. and M.G. Morris. 1985. Threatened Swallowtail Butterflies of the
World. The IUCN Red Data Book. Gland, Switzerland and Cambridge, UK.
401 pp.

Coyne, J.A. and H.A. Orr. 1989. Two rules of speciation. pp. 180-207. IN:D. Otte
and J.A. Endler, eds., Speciation and its consequences. Sinauer, Sunderland,
MA.

Forbes, W.T.M. 1951. Footnote on Papilio . Lep. News 5:16.

Geiger, H. 1087[88], Enzyme electrophoresis and interspecific hybridization in
Pieridae (Lepidoptera) - The case for enzyme electrophoresis. J. Res. Lepid. 26:
64-72.

Hagen, R.H. 1986. The evolution of host plant use by the tiger swallowtail
butterfly, Papilio glaucus . Phi). Thesis, Cornell University, Ithaca, NY. 293

pp.

Hagen, R.H. and J.M. Scriber. 1991. Systematics of the Papilio glaucus andP.
troilus species groups (Lepidoptera: Papilionidae): inferences from allozym.es .
Ann. Ent. Soc. Amer. (in press).

Hancock, D.L. 1983. Classification ofthe Papilionidae (Lepidoptera): a phylogenetic
approach. Smithersia 2:1-48.

32

J. Res. Lepid.

Jordan, K 1907. Papilio IN: A. Seitz, ed., Macrolepidoptera of the World. The
American Rhopalocera 5:11-45. Alfred Kernan, Stuttgart.

Lederhouse, R.C. , M.P. Ayres, and J.M. Scriber. 1990. Adult nutrition affects male
virility in Papilio glaucus . Functional Ecology 4:743-751.

Lorkovic, Z. 1986. Enzyme electrophoresis and interspecific hybridization in
Pieridae (Lepidoptera). J. Res. Lepid. 24: 334-358.

Miller, J.S. 1987a. Phylogenetic studies in the Papilioninae (Lepidoptera:
Papilionidae). Bull. Amer. Mus. Nat. Hist. 186: 365-512.

. 1987b. Host-plant relationships in the Papilionidae (Lepidoptera): parallel

cladogenesis or colonization? Cladistics 3:105-120.

Munroe, E. 1961. The generic classification of the Papilionidae. Can. Ent. Suppl.
17:1-51.

Nitao, J.K., M.P. Ayres, R.C. Lederhouse, and J.M. Scriber. 1991. Larval
adaptation to lauraceous hosts: geographic divergence in the spicebush
swallowtail butterfly. Ecology (in press).

Ruszczyk, A. 1986. Mortality of Papilio scamander (Lep. Papilionidae) pupae in
four districts of Porto Alegre (S. Brazil) and the causes of superabundance of
some butterflies in urban areas. Rev. Brasil. Biol. 46(3): 567-579.

Scriber, J.M. 1982. Foodplants and speciation in the Papilio glaucus group. IN:
J.H. Visser and A.K. Minks, eds. , Proceedings 5th International Symposium on
insect-plant relationships. Pudoc. Wageningen, Netherlands, pp. 307-314.

. 1984. Larval foodplant utilization by the world Papilionidae (Lep.):

latitudinal gradients reappraised. Tokurana (Acta Rhopalocerologica) 6/7:1-
50.

. 1986. Origins of the regional feeding abilities in the tiger swallowtail
butterfly: Ecological monophagy and the Papilio glaucus australis subspecies
in Florida. Oecologia 71: 94-103.

. 1988. Tale of the tiger: Beringial biogeography, bionomial classification, and

breakfast choices in the Papilio glaucus complex of butterflies, pp. 241-301. IN:
Chemical Mediation of Coevolution (K.C. Spenser, ed.) Academic Press, NY.
609 pp.

Scriber, J.M., R.V. Dowell, R.C. Lederhouse, and R.H. Hagen. 1991. Female color
and sex ratio in hybrids between Papilio glaucus glaucus and P. eurymedon,
P. rutulus and P. multicaudatus. J. Lepid. Soc. 44:229-244.

Scriber, J.M., M.H. Evans, and R.C. Lederhouse. 1988 (1990). Hybridization of the
Mexican swallowtail Papilio alexiares garcia with other glaucus group species
and survival of pure and hybrid larvae on potential host plants. J. Res. Lepid.
27: 222-232.

Scriber, J.M. and R.C. Lederhouse. 1988 (1989). Handpairing of Papilio glaucus

glaucus and Papilio pilumnus and hybrid survival on various foodplants. J.
Res. Lepid. 27: 96-103.

Scriber, J.M., R.C. Lederhouse and R.H. Hagen. 1991. Foodplants and evolution
withi nPapilio glaucus and Papilio troilus species groups (Lepidoptera:
Papilionidae), IN: P.W. Price, T.M. Lewinsohn, G.W. Fernandes, and W.W.
Benson, eds. , Plant- Animal Interactions: Evolutionary Ecology in Tropical and
Temperate Regions. John Wiley, New York.

Journal of Research on the Lepidoptera

29(l-2):33-36, 1990(91)

Significant Additions to the Butterflies of the
Trinity Alps and Mount Eddy, Northern California

Arthur M. Shapiro

Department of Zoology, University of California, Davis, California 95615

The Trinity Alps and Trinity Divide, including Mount Eddy, are
situated in northwestern California and provide high-montane, sub-
alpine and alpine environments of granitic, metamorphic, and ultrabasic
substrates far removed from the Sierra Nevada and thus of considerable
biogeographic interest. The butterfly faunistics and ecology of the region
were investigated by Shapiro, Palm and Wcislo (1981), ( Shapiro, A. M.,
C. A. Palm & K. L. Wcislo. 1981. The ecology and biogeography of the
Trinity Alps and Mount Eddy, northern California, J. Res. Lepid. 18: 69-
152), who based on five years5 field work reported 115 species. Despite
frequent visits in the subsequent eight field seasons, only one additional
species has been recorded. However, of the hundreds of additional
records and the twenty-five or thirty new collection sites, a number are
of special interest and bear on our understanding of the biogeography of
the region. This paper reports the most important of the new information
gathered since 1980. It is intended to be used with Shapiro, Palm and
Wcislo (1981). New locality and capture data are presented by species;
page numbers refer to the original text.

The following are the most significant of the new localities, presented
in the format of Appendix I (pp. 145-147). The 15' topographical maps on
which they can be found are BK - — Bonanza King and D — Dunsmuir.
Since 1981 a large number of new 7.5' quadrangles have appeared, so
most of the study area is now mapped on a much finer scale. The 7.5'
quadrangles are also identified for the new localities.

Asbestos Gulch: 4 km SW Horse Heaven Meadows; hillside seep with tufa
deposits, unusual vegetation (discussed below); Trinity Divide: 1603 m. (Seven
Lakes Basin 7.5')

Crow Creek: Darlingtonia bog near jet. Crow Creek and E Fork Trinity River,
1220 m. (Mumbo Basin 7.5') BK

Horse Flat Camp: Lower montane mixed forest, 5.3 km up Eagle Creek from
trailhead, 1450 m. BK

Horse Heaven Meadows: Ledum bog, Castle Creek Road at E Fork Trinity
River below Grey Rocks, 1.6 km W Whalen Summit, 1767 m. (Seven Lakes Basin
7.5') D

Additions to Annotated List of Species

2. Oeneis nevadensis Felder and Felder, (p. 92)

RECORDS: Horse Heaven Meadows, 21.VI.86.

4. Cercyonis pegala boopis Behr. (p. 93)

RECORDS: Deadfall Meadow, 22.VI.86.

34

J. Res. Lepid.

21. Phyciodes orseis orseis Edwards, (p. 99)

RECORDS: Horse Heaven Meadows, 2 1.VL86; Deadfall Meadow (3), 22. VI. 1986.

33. Precis coenia Hbn. (p. 103)

RECORDS: Horse Heaven Meadows, 21.VI.86.

34. Limenitis lorquini Bdv. (p. 103)

RECORDS: Horse Heaven Meadows, 22.VI.86.

37. Lycaena arota Bdv. (p. 104)

RECORDS: Horse Heaven Meadows, 5.DC.87.

41. Lycaena mariposa Reak. (p. 106)

RECORDS: Horse Heaven Meadows, 21.VI.86, 5.IX.87.

49A. Mitoura johnsoni Skinner. Johnson’s Hairstreak.

The only additional species from the region, a large female collected from
flowers of Ledum glandulosum Nutt. var. californicum (Kell.) Hitchc.
RECORDS: Horse Heaven Meadows, 21.VI.86.

54. Callophrys lemberti Tilden. (p. Ill)

RECORDS: Mount Eddy summit, 18.VI.89.

57. Glaucopsyche piasus Bdv. (p. 112)

RECORDS: Coffee Creek, 1000m, 20.VI.89; Horse Flat Camp, 21.VI.89; Deadfall
Meadow (4), 22. VI. 1986.

66. Agriades glandon Prunner. (p. 115)

RECORDS: Horse Heaven Meadows, 21.VI.86; Crow Creek, 21.VI.86.

71. Neophasia menapia Feld. & Feld. (p. 117)

RECORDS: Horse Heaven Meadows, 5.DC.87.

88A. Euphyes vestris Bdv. (p. 123)

RECORDS: Crow Creek (4), 21.VI.1986.

94. Hesperia harpalus Edwards (. H . “comma complex”) (p. 124)

As noted in the original text, Trinity-Eddy populations are extraordinarily
variable. On 5. IX. 1987 a series of 70 specimens (29 ? , 41 / ) was taken on the tufa
exposure near Asbestos Gulch in the Trinity Divide. This represented a small
fraction of the number of animals flying in the densest population of this complex
I have ever seen outside high desert. The variability was extreme, even for this
region. A sample is illustrated in figs. 1 and 2. The most extensively-marked
males were strongly reminiscent of H. lindseyi Holland. All of the animals, along
with Polites sabuleti Bdv. and Hesperia juba Scudder, were visiting flowers of a
dwarfed ecotype of rabbitbrush, Chrysothamnus nauseosus (Pall.) Brit, which
was the aspect dominant on the site. Associated with it were Cirsium breweri
(Gray) Jeps. (mostly already setting seed) and Triglochin maritima L., both
strongly calciphilic and the latter very local inland.

100. Carterocephalus palaemon Pallas, (p. 126)

RECORDS: Horse Heaven Meadows, 21.VI.86.

104. Erynnis icelus Scud. & Burg. (p. 127)

RECORDS: Horse Flat Camp, 21.VI.89; Horse Heaven Meadows, 21.VI.86.
107. Erynnis persius Scudder. (p. 128)

RECORDS: Horse Heaven Meadows, 5.IX.87.

Dorsal surfaces of some Hesperia " comma complex" col- Fig. 2. Ventral Surfaces of some Hesperia "comma complex" col-
lected near Asbestos Gulch, Trinity Co., 5.SX.1987. lecfed near Asbestos Gulch, Trinity Co., 5JX.1987 (not the

same individuals as fig. 1).

36

J. Res . Lepid.

Discussion

Most of the Horse Heaven Meadows records represent either first
collections or significant range extensions within the Trinity Divide,
which has been rather poorly documented except for Mount Eddy and
Deadfall Meadow. Most of the Trinity Divide is relatively low and heavily
forested, and until very recently 7.5' topographic quadrangles were not
available for it. It is crisscrossed by old logging and mining roads and
Jeep trails, and by relatively random wandering on them it is possible to
encounter unusual habitats, including the Ledum bog at Horse Heaven,
the Darlingtonia bog at Crow Creek, and the tufa exposure at Asbestos
Gulch, all of which produced important records. The Crow Creek
Euphyes vestris connect the previously-known and isolated colonies at
Scott Mountain Summit and Mount Shasta City, and suggest that many
other colonies of this very local skipper may exist in the region.
Carterocephalus palaemon was previously unrecorded in the Trinity
Divide. The many species collected at Horse Heaven and elsewhere in the
vicinity document the ecological breadth which allows them to transcend
the edaphic and vegetational complexity of northern California; the
montane fauna is much more eurytopic than the sub-alpine and alpine.
At least one species enters the eastern fringe of the Trinity Divide, but
fails to extend more than about 6 km west of Interstate 5. This is the
apparent Lycaena editha Mead — L. xanthoides Bdv. intergrade, which
is common at Dunsmuir and extends locally westward along the Callahan-
Gazelle Road, but is otherwise absent form the Trinities and Eddies. Its
distribution does not correlate very well with climate or vegetation.

The Asbestos Gulch series of Hesperia, embracing as much variation as
occurs in the entire region, includes specimens which if taken out of
context could be called three different subspecies. One wonders how
many taxonomic decisions have been predicated on such small and
unrepresentative samples!

Acknowledgments. This research has been supported by California Agricultural
Experiment Station project CA-D*-AZO-3994-H, “Climatic Range Limitation of
Phytophagous Lepidopterans.” I thank David Olson, Linda Farley, Bill Overton,
John Wagoner, and Hansjiirg Geiger for field assistance. The photographs are by
Samuel W. Woo.

Journal of Research on the Lepidoptera

29(l-2):37-53, 1990(91)

Heritable Color Variants in Automeris io
(Saturniidae)

Thomas R. Manley*

Research Affiliate in Entomology, Peabody Museum of Natural History, Yale University

Abstract: Automeris io populations east of the Mississippi River and
north of the Gulf States show extreme sexual dimorphism in forewing
dorsal coloration. The yellow males are relatively uniform and the red-
brown females are more polymorphic in appearance. A large number
of inbred experimental lines has yielded several genetic color forms not
known from wild sampling. The Mendelian heterozygote “broken-eye”/
’’claw”, the more polygenic “large” and “small” eyespot and the poly-
genic “broad” to “narrow” black intermarginal hindwing band genes
were reported in 1978. To these are added three simple Mendelian
recessives: dorsal hindwing “teardrop” with variable expressability;
“brown” forewing dorsal ground color; and “rose” fore and hindwing
ventral ground color, plus a recessive that produces “yellow” larvae
when homozygous. Variability of wild males in Louisiana and wild
females in Georgia is discussed.

Introduction

In 1964 a series of crosses were made within Automeris io (Fabricius)
to determine the genetics of this species hindwing eyespot. Eyespot-like
markings have evolved independently many times among insects, fishes,
reptiles and birds. Very large eyespots may function as an escape
mechanism, the possessor eluding capture by creating a “startling effect”
on potential predators, enhanced by a variety of behavioral activities
associated with the “eyes”. Darwin (1859) was an early commentator on
its survival value. Blest (1957) and Brower ( 1960) are among more recent
investigators of eyespots on lepidoptera.

Automeris io, like many other members of its genus, has a large eyespot
on each hindwing dorsum. The “eyes” of wild-caught moths are slightly
variable in size. My initial experimentation was two phased: to increase
the size of the hindwing eyespot by repeatedly crossing moths with the
largest eyespots; and the second phase to reduce the eyespot by crossing
moths with the smallest eyespots in successive generations. The results
of these experiments not only revealed developmental genetics of the
eyespot but by serendipity the several inbred lines exposed various
remarkable Mendelian recessives altering the conspicuous markings of
the moth (Manley 1978). The “broken-eye” breeding program was
terminated in 1986 with the loss of the several lines, due to adverse
weather conditions that summer. During the twenty years of continuous
selective inbreeding, several additional variations of the conspicuous

^Correspondence: Route #1, Box 269, Port Trevorton, Pennsylvania 17864

38

J. Res. Lepid.

markings and ground color were produced. The genetics of eyespot size
and of a hindwing/forewing pair of characters, “broken-eye” and “claw”
was described and illustrated earlier (Manley 1978). In the present paper
one more pattern variant “teardrop”, two forewing ground color genes
“brown” and “rose” and a larval color form are discussed and figured.

Materials and Methods

Breeding stock was derived from wild Pennsylvania Automeris io females
taken in 1963 in the vicinity of Klingers town, Schuylkill Co. First instar larvae
were started in sleeves on wild Black Cherry ( Prunus serotina Ehrh). Final instar
larvae were placed in screened cages, with leaves available for cocoon spinning.
Cocoons were refrigerated at 5°C from October to May. Pupae were placed in
screened cages at room temperature in early May; adults emerged in early June.
Adult behavior and breeding techniques have been described in detail elsewhere
(Manley 1991). Crosses were made from selected individuals, and specimens
involved in experimental crosses and their progeny were killed and spread for
permanent reference. Virtually all specimens are deposited in the Entomology
Division, Peabody Museum of Natural History, Yale University (YPM).

Description and Modifications of Conspicuous Wing Markings

The conspicuous markings (Fig. 1) are: dorsal hindwing eyespot, nor-

A - Dorsal Surface B - Ventral Surface

Dorsal surface Ventral surface

Fig. 1. Venation, color patterns, and conspicuous markings of Automeris io io. Dorsal
surface: A - chevron line; B - forewing discal patch; C - outer marginal band;
D - inner marginal band; E - focus or pupil of eyespot; F - eyespot; G - basal
hair band. Ventral surface: A - forewing bar; B - discal patch; C - hindwing
bar; D - hindwing discal spot; E - basal hair band.

29(l-2):37-53, 1990(91)

39

mally round or oval, always black, with a gray or bluish iris surrounding
the small white pupil or “focus” (Nijhout 1978, 1980, 1981) located in the
center of the hindwing; the dorsal forewing discal patch, normally a
kidney shaped mark slightly posterior to the costal region on the areolar
area of the forewing, may extend along the subcostal and radial veins
forming four blunt finger-like projections toward the outer margin of the
wing; the ventral fore wing discal spot, an oval patch of black scales with
a small spherical white pupil or patch, commonly called the forewing
ventral eyespot. The ventral hindwing discal spot is located beneath the
white pupil or “focus” of the dorsal hindwing eyespot in the form of a white
dot or “focus”, it does not appear to be associated with the dorsal hindwing
eyespot (Manley 1978). The other conspicuous character is the black
intermarginal band which is genetically independent of the other con-
spicuous markings, its width controlled by a single gene which broadens
the band.

Variant Imaginal Phenotypes
A. HINDWING DORSAL “TEARDROP”

In 1973 two inbred lines 11-70 and 13-70 expressing the “broken-eye”
phenotype, produced a male and two females with a new variant eyespot
having an anterior black satellite spot. Its emergence from the eyespot
was reminiscent of a brimming tear, and was ultimately named “tear-
drop”. The spot is usually connected to the eyespot but is sometimes
entirely separated, especially if very small. There is a pronounced
asymmetry in size and shape of the “teardrop” between the left and right
wings. When it is present in an individual showing “broken-eye”, it can
easily be mistaken for another outreaching lobe, thus the “teardrop”
variant was initially overlooked (Plate 1, Figures 1-6) .

Again in 1974, six “teardrop” forms were noted in the 2100 adult io
spread for study. Three more from a cross involving the normal eye,
recessive to “broken-eye”, were observed as small black spot separations
from the black outer ring of the eyespot, suggesting the possible forma-
tion of a line breeding true for a normal eyespot plus a “satellite”. The
years 1975-1976 produced no variant eyespots in the “broken-eye” inbred
lines, indicating an unstable developmental pathway due to inbreeding
rather than a discrete “gene” controlling trait. In 1977 the 13-70 “broken-
eye” inbred line produced a female with an eyespot variant which we then
finally designated as “teardrop”, and two other females had the “tear-
drop” variant superimposed over the “broken-eye”, so that only a small
portion of the “teardrop” extended beyond the “broken-eye” area. The
1978 “broken-eye”, 13-70 series, produced 12 females and 6 males with
evidence of the “teardrop” eyespot modification; five matings resulted in
no fertile ova.

The “teardrop” variant was not observed again until 1982 when two
crosses of the 13-70 inbred “broken-eye” line produced “teardrop” in 12
females and three males. A successful mating of a pair with “teardrop”
eye initiated the “teardrop” line. A single cross of “teardrop” parents,

40

J. Res. Lepid.

0 10 20 30 40 50 60 70 80 90 1 00

, MILLIMETERS ,

Plate I. Varying degrees of expression of “teardrop” eye in females of Automeris io.

Fig. 1 . Cross 27-85 “teardrop” x teardrop Typical “teardrop” expression.

Fig. 2. Cross 4-73 “broken-eye” x “broken-eye” “Teardrop” expression from the
recessive normal eye in the “broken-eye” line.

Fig. 3. Cross 5-74 “broken-eye” x “broken-eye” “Teardrop” expression with a
“satellite” spot to the recessive normal eye in the “broken-eye” line.

Fig. 4. Normal eyespot, Wild Colorado female. Control.

Fig. 5. Cross 7-74 “broken-eye” x “broken-eye” “Teardrop” superimposed on
“broken-eye”.

Fig. 6. Cross 28-85 “teardrop” x “teardrop” Expression of incomplete penetration,
note ellipsoid shape of eyespot.

obtained in 1983 produced in 1984 two successful matings out of 15
attempts; 1985 provided enough “teardrop” adults to set up the entire
range of experimental crosses, resulting in eight successful crosses out of
50 matings, which provided the necessary data to analyze phenotypes
expressed by this condition. In 1986 two “teardrop” matings were
successful, thus maintaining the genetic strain for further study. Crosses
in 1987 produced four successful matings, but all larvae died due to
adverse weather and line was lost.

The “teardrop” eyespot is controlled as a recessive (Fig. 2); as crosses
to normal wild A. io produce no “teardrop” eyespots in the F 1 generation.
In backcrosses, “teardrop” appears only if the normal-eye parent is
heterozygous for the “teardrop” gene. Crosses of “teardrop” x “teardrop”
all have “teardrop” or if no satellite spot is present, the form is an ellipsoid
instead of a round eyespot. No precise frequencies of ellipsoid eyespots to

29(l-2):37-53, 1990(91)

41

1977

1978

1982

P

13-70 13-70

120 84 2 9 0 2 86 46 32 25 0 3 0 0 2 0 27 32

P

13-70 13-70

0 N from 0

m f

m fm fm fm fm f
66 46 2 10 1 1 0 1 80 44

Figure 2. Isolation crosses from inbred 1 3-70 broken-eye line to establish pure lines
of the “teardrop” phenotype in Automeris io. Numbers indicate the success-
ful matings. Abbreviations: 0 Broken-eye; TD Teardrop; IP Incomplete
Penetration; N Normal Eyespot.

“teardrop” have been determined, due to the difficulty of evaluating the
wide variation in expression of the ellipsoid eyespot. There seems to be
a slightly greater ellipsoidal distortion in males than in females.

The “teardrop” gene has highly variable expressibility (Hartl 1980). It
demonstrates a reduced or incomplete penetrance (Herskowitz 1980),
some moths appearing normal for eyespot yet being homozygotes for the
gene to express the “teardrop” phenotype. Crosses involving parents with
normal appearing eyespots from the “teardrop” line produced the same
phenotypic ratios as did parents possessing the “teardrop” eyespot.

42

J. Res . Lepid.

Unlike the “broken-eye” gene, the “teardrop” gene appears to be indepen-
dent of the forewing discal patch, as none of the 650 “teardrop” specimens
studied show any distortion of the forewing discal patch. The fitness of
“teardrop” broods were low due to their inability to mate, and to mortality
of first instar larvae due mainly to their poor acceptance of a suitable food
plant. When matings were successful, and first instar larvae fed well on
Prunus serotina, then maturation and survival rate of pupae was normal
for A. io.

Over much of its range, Automeris io has spectacular sexual dimor-
phism, due to the bright yellow forewings of males and the deep brown-
red of many shades of the fore wings of the females. It is interesting that
the “teardrop,” “broken-eye”/“claw,” and eyespot size genes show no
pronounced sexual differences.

B. GROUND COLORS

1. Basic description of ground color of Automeris io io wild type

Many authors have referred to the extreme variability of Automeris io
io within any portion of its range. Numerous infraspecific forms, with
many that are basically localized genetic variants, have been named
(Packard 1914, Ferguson 1972).

A basic pattern emerges from controlled breeding experiments, the
study of many wild specimens across its range, and mass samples from
critical areas such as Florida, Georgia, Louisiana, Texas, New Mexico,
Colorado, and northeastern United States. Deviations from this pattern
have been isolated and genetically analyzed. The principal ones and
some rare aberrations are discussed in this paper. The color descriptions
of each sex, including noteworthy regional variations to the basic color
pattern, are discussed. Specific regionally expressed genes changing the
basic color pattern are genetically analyzed.

Emphasis is given only to ground color patterns as the genetics of the
conspicuous markings typical to all A. io io and its subspecies is known
(Manley 1978). Variation in ground color of the unique geographical
subspecies of A. io io will be discussed in separate publications. Maerz
and Paul 1930, color plate designations and descriptive terminology is
used to describe color throughout the manuscript.

The ground color of males is jasmine or bright yellow. The dorsal
forewing contains the kidney shaped forewing discal patch; the chevron
line, parafocal elements (Schwanwitch, 1924; and Suffert, 1927) are
located approximately two-thirds of the distance, beginning in the anal
2 cell and extends upward to the subcosta cell at the margin. The line is
frequently broken in the radial cell area, in some cases absent or reduced
to a small patch in the anal 2 - cubitus 2 area. The color of these markings
is determined by a series of complementary genes whose expressions
range from dull rosy red to opal grey.

The dorsal hindwing of the male is consistently yellow with a dense
area of long rosy red scales, the basal hair band, extending from the point

20(l-2):37-53, 1990(91)

43

of attachment of the wing to the metathoracic body segment. These
hairlike scales cover the surface of the anal 2 and cubitus 2 cells and fuse
with the outermarginal band. The outermarginal band is generally roeey
red parallels the contour of the margin of wing beginning in the anal 2 cell
extending to the subcostal + radial 1 cell of the wing. Its width acts
independently of the highly conspicuous black intermarginal band,
whose width is controlled by a single gene (Manley 1978).

The forewing ventral surface of the male is yellow. The forewing bar, ’
a rosy red line of scales, separates the outer one-third of the wing. This
line begins at the anal 2 cell and extends forward to the outer margin
terminating in radial 1 or frequently in the radial 4+5 cell area. The inner
two-thirds of the ventral forewing may be rosy red, the amount varies
from a limited expression, resulting in a generally yellow forewing, to full
expression with the rose color extending from anal 2 area to radial 4+5
vein, A yellow band along the outer margin covers the subcosta, El and
112-4 3 cells, its presence is enhanced when a full extension of the rosy
coloration of the inner two-thirds of the wing is present.

The ventral hindwing is light yellow and is semi-transparent in that
the conspicuous markings on the dorsal surface are visible, especially the
focus of the eyespot which appears as a white dot. The size of the dot is
dependent upon the size of the focus of the eyespot. The ventral hindwing
bar is a line of rosy red scales extending from the marginal terminus of
the radial 1 vein diagonally across the wing to the anal 2 cell separating
the outer third of the wing. The bar may be a fine line or quite broad and
distinct. A fine band of rosy red scales extends along the outer margin of
the hindwing.

2. Regional differences in ground color in Georgia and Lousiana

Regional modifications of the basic ground color are most easily
recognized in the males which have yellow fore wings, less observable in
females which have dark forewings. Along the southern boundary of A.
io io where it is bivoltine (Manley 1991), the ground color is subject to a
variety of regional gene modifications. Those in northern Georgia are
particularly dramatic, as a high degree of uniformity is expressed in
individual broods reared from wild females. A, seasonal polyphenism
(Shapiro 1984) appears to exist in that several broods may differ pheno-
typically from each other, a situation not observable in the wild due to the
natural dispersal of the brood. Dr. Hermann Flaschka of Decatur,
DeKalb County, Georgia, has over the past seven years, supplied ova
from wild females, which have produced the “yellow larva51 phenotype
and some large broods with uniform ground color expressions not typical
of northern A io io. In several of these broods the males were uniformly
orange-yellow ground color, Plate 2, Figure 3; females of these broods
were predominantly copper brown, suggesting a high degree of homozy-
gosity or independent of a sexual dimorphism mechanism for ground
color. The male progeny of one wild female were predominantly honey

44

J. Res. Lepid.

yellow, Plate 2, Figure 10, with some typical yellow males. These copper
or yellow brown males show no close similarity to the tawny orange
brown males of the Automeris io lilith from along the Georgia coast and
the Florida peninsula. The status of A. io lilith in Florida will be discussed
in a later paper.

In Louisiana there must be a wide range of genes which make A. io io
males diverse in their color patterns. Phenotypes for some of these
modifications are present in northern A. io io but not expressed to the
degree observed in Louisiana, except in controlled inbred lines described
later in this paper. Vernon A. Brou collected over an eight year period
(1978-1985) more than eight hundred Automeris males in Abita Springs
(St. Tammany Parish), Edgard (St. John’s Parish), and Weyanoke (West
Filiciana Parish). A ground color phenotype not observed in our 25 years
of breeding northern A. io io has rosy or brown scales on the ventral wing
surface of these males, and on the limited number of females we have
studied. It was present in 92% of the Louisiana sample. Unique to this
phenotype is its intensity of expression, which appears to be influence
other genes producing a “brown wing” phenotype. Manley (1978) demon-
strated that there was no relationship of the dorsal surface genes for
conspicuous markings to those on the ventral surface. Enhancement of
the dorsal surface “brown wing” phenotype intensifies the expression of
rosy or brown scales on the ventral wing surface.

Another ground color phenotype observed in the Louisiana sample
is rose-brown or cinnamon, Plate 2, Figure 9, fore wing ground color gene,
observed in six percent of the males in the March- June diapausing
generation. This color pattern was not expressed in the non-diapausing
generation flying July-September. The mechanism of expression of this
gene appears to be similar to the one controlling the tawny-brown males
of the diapausing generation of Florida A. io lilith. This color pattern has
not appeared in our 25 years of inbreeding Pennsylvania A. io io. Pupae
from southern Louisiana, finally obtained fall 1989 should allow us to
further evaluate this phenotype and others unique to that region,
especially since there should be females for study.

The female ground color of A. io is difficult to describe due to the wide
array of potential hues that range from red to opal grey. The ground color
pattern is sex limited (Remington, 1954, 1976) and is further complicated
by the expressions of specific genes, conspicuous in males but masked by
ground color or absent in expression in the female. Plate 3, Figures 19-
30 show females representing the potential range of ground colors
observed in A. io. Wild specimens were used to demonstrate the
predominant color of the female displayed in various regions of its range.
Pennsylvania A. io has in its genome the ability to produce any of the
basic ground colors found in the United States, with the exception of the
tawny brown male coloration of the diapausing generation of Florida A.
io lilith. The source of genes for this color pattern will be presented in a
separate paper on the status of the Florida populations.

29(l-2):37-53, 1990(91)

45

The color range of the dorsal forewing, based on examination of over
5,000 female specimens, is red tones in the range of Persimmon or copper
brown, Plate 3, Figure 19; the darkest tone is opal grey, Plate 3, Figure
30. The dorsal forewing discal patch may be slightly darker than the basic
ground color or it may blend into the ground color pattern; it is not as
conspicuous as it is in males. The chevron line separating the outer third
of the wing may be prominent or it may blend into the ground color.
Occasionally the outer margin of the wing will be lighter in color, making
these females more conspicuous.

Figures 20 through 28 represent the most frequently expressed forew-
ing ground colors of A. io io in the United States. In controlled crosses a
majority of females have similar color patterns with a strong tendency
toward dull plum red tones suggesting the expression of a heterozygous
complementary gene complex for color. Many predominantly rosy red to
plum red female crosses produce a small number of darker forms,
suggesting that opal gray is recessive.

The subdued color tones of the dorsal forewing of a wild female may
make it inconspicuous when resting in the shadows of leaves near the
trunk of a tree.

The dorsal hindwing color pattern is uniform throughout the species
range; it forms the background for the display of the conspicuous “eye”
markings. With the exception of the three narrow marginal bands the
central portion of the hindwing is always a shade of yellow. The black
eyespot and black intermarginal band are displayed in the yellow area.
The band along the margin of the wing may be a fine line, or a narrow

Plate 2. Color range of Automeris io males and digress of expression of the rose
underwing phenotype.

Fig. 7. Wild Liverpool Pennsylvania, representing typical ground color for north-
eastern Io males.

Fig. 8. Progeny of Wild Georgia female — orange red ground color phenotype
assumed to be homozygous as all siblings were same color.

Fig. 9. Louisiana male, phenotype observed only in the diapausing generation.

Fig. 10. Wild northern Georgia honey-brown phenotype.

Fig. 1 1 . Cross 9-74. Full expression of “brown” wing gene.

Fig. 1 2. Wild Louisiana, partial expression of “brown” wing gene, usually present in
varying degrees in Louisiana males.

Fig. 13. Cross 7-73. Tawny pink northern male, color occasionally appears in
Northern Io.

Fig. 14. Cross 30-85. “teardrop” x “teardrop”. Typical “teardrop” male.

Fig. 15. Cross 9-74. Full expression (YyRR) of rose underwing phenotype on fore
and hindwings.

Fig. 16. Cross 10-85. Partial expression (yyRr) of rose underwing phenotype
hindwing rose, forewing normal.

Fig. 17. Cross 10-85. Limited hindwing expression (YyRr) of the rose underwing
phenotype rose dusting on hindwing.

Fig. 18. Cross 10-85. Normal (YYrr) hindwing ground color.

48

J. Res. Lepid.

band of color identical to the basic ground color of the dorsal forewing.
The middle band is always lighter in color, a suffusion of the ground color
scales on yellow, and its width varies. The outer marginal band bordering
the yellow area of the dorsal hindwing is generally the widest and most
pronounced. It may be the same color as the forewing blending the
hindwing profile with the forewing, or it may be brighter colored and
quite conspicuous. Extending from the base of the wing along the inner
marginal surface, area Cu2 and 2A, is the basal hairband, a dense patch
of long hairlike scales ranging from rosy red to opal grey. The color of
these scales blends with the ground color of the forewing regardless of the
depth of color of the forewing.

The ventral surface is uniformly colored, both forewing and hindwing.
The colors are slightly lighter and duller than those of the dorsal surface.
As in the male, a line of rosy red or plum red, rarely opal grey, scales
forming the dorsal and ventral hindwing bars separate the outer one-
third of the wing. The bar is a line of deeper colored scales extending from
the middle of the anal 2 upward across the wing to the terminal point of
the radial 1 or radial 4-5 vein on the margin of the ventral forewing
surface. The bar extends from the anal 2 cell upward to the point where
the subcosta + radial 1 or radial 5 vein terminate on the margin of ventral
hindwing. Occasionally bars are missing or masked by certain genes,
namely the red underwing gene, making the ventral surface a single
color.

The ventral forewing discal spot, an oval or egg shaped patch of varying
size of black scales with a small white pupil or focus, is the conspicuous
marking on the ventral surface. The gene controlling its size and
intensity is independent of genes controlling conspicuous markings on
the dorsal surface (Manley 1978). The white pupil of the eyespot on the
dorsal hindwing is visible ventrally as a white dot in the center of the
hindwing.

Plate 3. Color range of Automeris io females.

Fig. 19. Progeny of Wild Georgia Female, sibling Figure 8 male, homozygous
dominant red.

Fig. 20. Wild New Jersey.

Fig. 21. Cross 9-72. Female shows “broken-eye”.

Fig. 22. Cross 14-73. Female shows full expression of “broken-eye”.

Fig. 23. Wild Louisiana.

Fig. 24. Wild Colorado.

Fig. 25. Wild Pennsylvania.

Fig. 26. Wild New Jersey.

Fig. 27. Wild Kansas.

Fig. 28. Wild Georgia.

Fig. 29. Cross 9-73. With full expression of “broken-eye”.

Fig. 30. Cross 3-74. Homozygous recessive for plum gene, a rare expression.

29(l-2):37-53, 1990(91)

49

3. “Brown” fore wing dorsum

Any mass sample of eastern A. io io will have a high frequency of males
with brown scales of varying intensity on the dorsal forewing. Our
Liverpool, Snyder County, Pennsylvania sample (N=123) taken over a 25
year period shows 69% of the males with some degree of brown suffusion.
A. io in the Peabody Museum Collection at Yale University and other
large mass samples show a similar percentage of expression of this
phenotype. In most cases the genes for “brown” wing are minimally
expressed in wild males; however, strong expression of the genome can
quickly be produced by selective breeding. With the maximum expres-
sion of this gene complex, the color of the basal two-thirds of the dorsal
forewing may be rosy red to brownish opal grey depending on the basic
ground color geny complex being expressed. “Brown” wing is not observ-
able in females as it may be masked by the normal dark ground color (or
perhaps it is not present in females).

The initial full expression of “brown wing” gene complex was first
observed in cross 15-71, F5 generation of the inbred line for “broken-eye”
(Plate 2, Figure 11). Crosses were made yearly 1972-1978 in an attempt
to isolate true breeding lines of “rosy red” and “opal grey”. Color isolation
was abandoned in 1978 due to the inability to diagnose the phenotypes
of the females, and the relatively high percentage of uniform brownish
males present in every cross. Females in these crosses were uniform in
ground color.

Males, assumed to be homozygous, having full expression of maximum
forewing coloration, were mated with sibling females with background
colors, predominantly “rosy red” or “opal grey”, produced a relatively
uniform distribution of male color patterns regardless of the ground color
of the male, suggesting a sex-limited polygenic autosome controlling its
expression. Frequency of the full expression of the gene in these crosses
averaged 88.8%.

The Louisiana A. io population differs from others in that random mass
samples show the full range of expression of the “brown” wing phenotype,
making it distinct from other A. io. A random mass sample (N=805)
segregated: “evidence” 21.1%; “medium expression” 42.3%; “strong ex-
pression” 33.2%; “no expression”, yellow wings .03% for the “brown” wing
gene. Although the full expression of the gene was .07% by wild random
mating it provides evidence this gene plays a major role in the unique
color patterns of Louisiana A. io. The similarity in male color patterns
between controlled crosses of northern A. io and random wild matings in
Louisiana involving the “brown” wing gene make their separate identi-
ties difficult.

4. “Rose” underside phenotype

The rosy underside phenotype, Plate 2, Figures 15-18-17, conspicuous
in the yellow male, is extremely difficult to observe in females whose
normal rose ground color masks its degree of expression. Analysis of this

50

J. Res. Lepid.

quantitatively expressed recessive is further complicated by the (diffi-
culty in selecting females with a recognizable degree of expression of the
gene to mate with “rosy” underside males. Evidence of the phenotype first
appeared in cross 18-69 and F 3 inbred line for “broken-eye” gene when
17 males possessed “rosy” scales of varying density superimposed among
the normal lemon yellow scales that form the basic underwing ground
color.

This phenotype was expressed regularly in the 13-70 inbred line for
“broken-eye” from 1971-1975. A 1975 cross produced individuals with the
entire ventral surface deep rose, Plate 2, Figure 15; this phenotype was
present in varying degrees of expression, on approximately 50% of the
males, suggesting a 1:1 ratio of yellow to “rosy” underside for that cross.
Serious attempts to isolate the underlying gene or genes began in 1981
and continued through 1985. To measure the degree of expression, the
rosy phenotype was designated as; “normal yellow” (YYrr); “trace”, a
faint dusting of rosy scales along the outer margin of the hindwing and
faint dusting on the hindwing venter (YYRr); “rose dusting”, a light to
medium rose dusting over the entire hindwing surface (YyRr); and “deep
rose”, heavy rose scales on the ventral hindwing (yyRr); and in extreme
instances heavy rose scales extending over the entire ventral forewing
(yyRR).

By combining the phenotypes of six crosses for “rose” underside using
the above segregation criteria, phenotypes of the offspring were 229
yellow, 77 “rose trace”, 89 “faint to moderate dusting” and 28 “heavy rose
dusting”, a close fit to a 9:3:3: 1 ratio (x2=1.6, df=2, p>.50). On rare
occasions when the recessive opal gray ground color is expressed, the
“rose” underside genome is expressed as grayish brown (N=4:423). I have
never observed this phenotype in wild males in northeastern United
States; however evidence of this phenotype was observed in 0.5% of the
Louisiana males (N=805).

5. “Yellow larvae” gene

Ova from a female taken in DeKalb County, Georgia produced two
distinctly different larval colors: the normal green and lemon yellow in
a 123 :97 ratio. During my 25 years of A. io breeding no “yellow larvae” had
been observed.

The occurrence of mutations affecting the color of the hemolymph of
Lepidoptera is well documented. Certain rare variations in larval color
were first reported by Gerould (1921) blue-green vs. green in Colias
philodice philodice Latreille in New Hampshire; Hoffman and Watt
( 1953) described blue-green vs. green in Colias philodice eriphyle Edwards
in Colorado; Gray (1953) reported yellow vs. green in Pieris rapae L.;
Stehr (1953) recorded yellow vs. green larvae in moths of the genus
Chorestoneura ; and Collins and Weast (1961) bluish vs. green larvae
(1:20) in Automeris io texana Barnes and Benjamin.

“Yellow larvae” were separated from “green larvae” into screen cages

29(l-2):37-53, 1990(91)

51

during final instar. Rearing continued on native wild cherry (Primus
serotina Ehrh.) and daily observations were made to note any differences
in behavior and growth. None was observed. Unique to this brood was
that the “yellow larvae” attracted large numbers of Arilus cristatus L. the
reduviid Wheel Rug.

These large insects could project their long beak through openings in
the screen into the body of larvae crawling along the surface. Although
some “green larvae” were killed by Wheel Bugs, they seem to, reason
unknown, concentrate on the more conspicuous “yellow larvae”. A mass
sample of 87 Arilus was taken in the vicinity of the cages, few have been
observed in the area since that time. Gerould (1921-1926) observed that
English Sparrows ( Passer domesticus L.) could locate and feed on the
highly visible “blue-green” larvae of Colias philodice while missing the
normal green larvae.

Surviving larvae began spinning cocoons on 1 September, and pupation
was complete on 12 September 1985. There was no observable difference
in shape or color of cocoons, pupae were stored at 5°C in plastic containers
from October to 1 May 1986. Pupae were placed in hatching cages and a
temperature of approximately 22 °C was maintained until adults emerged.
Ninety-seven “yellow larvae” produced 49 pupae; their emergence period
was 1 June - 23 June; sex ratio of adults, males 15/females 0. One
hundred twenty three “green larvae” produced 71 pupae; their emer-
gence period 1 June - 29 June; sex ratio of adults, males 30/females 5.
Sixty-five pupae eventually died, sexed by pupal case size; “yellow
larvae”, males 25/females 9; “green larvae”, males 18/females 13. The
high loss of pupae could be attributed to early September pupal forma-
tion resulting in many pupae lacking the ability to diapause. Non-
diapausing pupae normally hatch in October, thus are incapable of
enduring extended periods of storage at 5°C Manley (1991). The adult
males are identical in color regardless of larval color. Three matings of
“yellow larvae” males to “green larvae” females produced no fertile ova
and the brood was lost.

Fall 1988 Dr. Flaschka sent pupae reared from a wild female taken 24
June 1988 in the vicinity of Lake Allatoona, Bartow County, Georgia
which had both “yellow” and “green larvae”. Larvae were separated by
color during the final instar to enable one to isolate adults for future
study. From this mating only “yellow larvae” males survived. Females
from “yellow larvae” apparently were not able to develop into pupae; as
shrivelled, spine-covered larval bodies were found in the cocoons they
had spun. “Green larvae” developed normally, permitting a successful
mating of a “yellow larva” male and “green larva” female. Larvae of this
cross were poor feeders and were small sized in the final instar in
comparison with other crosses. Pupae from this mating produced ten
green larva males, one yellow larva male, two green larva females and no
yellow larva females. This brood produced no successful matings.

The apparent 1:1 ratio (123 “green” to 97 “yellow” larvae) of the DeKalb

52

J. Res. Lepid.

County, Georgia female in 1985 suggests that “yellow larvae” is ex-
pressed as a recessive. A female, heterozygous for “green larvae” mated
with a homozygous recessive “yellow larvae” male. The Barton County,
Georgia female (1988) apparently had the same genotype suggesting the
necessity of the presence of at least one dominant allele for “green larva”
color for females to survive, as no homozygous “yellow larva” females
have survived to date in this study.

The Au tomeris io research team (Manley 1991) was alerted to watch for
“yellow larvae” in wild populations in the Gulf Coast states. David
Ritland found a wild brood of Io on wild cherry with “yellow larvae” in
Chattahoochee National Forest, Union County, Georgia, which he reared
to adults. His comment: “Yellow larvae were very yellow, with no hint of
green. There were no intermediate colored larvae” (Ritland 1986). It
would appear the “yellow larva” gene is widely distributed across north-
ern Georgia. The other report of wild “yellow larvae” comes from Terhune
S. Dickel, who collects extensively in the Florida Keys. He wrote in 1986:
“All Io larvae that I have seen in southern Florida and the Keys thus far
have been bright lemon-green.” Lemon-green larval color in southern
Florida was initially observed by Annie T. Slossonin 1887 (Beutenmuller,
1895).

To date these are the only areas in the southeastern United States
where “yellow larvae” of A. io have been observed.

Acknowledgments. My principle thanks go to my long time friend and research
colleague, Professor Charles L. Remington, with whom I have discussed this
work throughout its duration - planning the crosses, speculating on the meanings
of data as they emerged, and not least in editorial advice with drafts of the
manuscript. I thank V. A. Brou and H. A. Flaschka who provided mass samples,
ova and pupae along with regional data. Thanks also to L. J . Kopp of Klinger stown,
Pennsylvania who reared larvae over a 25 year period, W. K. Sacco of Peabody
Museum, Yale, for the photography, and L. F. Gall, Yale for review and assistance
in finalizing the manuscript.

Literature Cited

Beutenmuller, W., 1895. Note on Hyperchiria io variety lilith. J. N.Y. Ent. Soc.
3:138-39.

Blest, A.D., 1957. The function of the eyespot pattern in Lepidoptera. Behavior
11:209-256.

Borror, D.J., D.M. Delong, & C.A. Triplehorn, 1977. An Introduction to the
Study of Insects, 4th ed., Holt, Rinehart. Winston, N.Y. 485.

Brower, L.P., 1960. Experimental studies of mimicry. Am. Nat. 44:271-282.
Collins, M.M. & R.D. Weast, 1961. Wild Silk Moths of the United States
Saturniinae. Collins Radio Company, Cedar Rapids, Iowa. 63-65.

Darwin, C., 1859. The Origin of Species: New America Library of World Literature.
N.Y. 115-116.

Ferguson, D.C., 1972. The Moths of America North of Mexico, Bombycoidea,
Saturniidae. Fascile 20. 2B 158-161, E.W. Classey Limited, London.

29(l-2):37-53, 1990(91)

53

Gerould, J.H., 1921. Blue-green caterpillars: the origin and ecology of a mutation
in hemolymph color in Colias (Eurymus) philodice. J. Exp. Zool., 34:385-415.

1926. Inheritance of olive-green and blue-green variations appearing in the
life cycle of a butterfly, Colias philodice. J. Exp. Zool., 43:413-427.

— — , 1927. Studies in the general physiology and genetics of butterflies. Quart.
Rev. Biol. 11:58-78.

Gray, P.H.H., 1953. A note on colors of pupae of Pieris rapae developed under
artificial conditions. Lepid. News 7:5-6.

Hartl, D.L., 1980. Principles of Population Genetics. Sinauer Associates Inc.,
Sunderland, Mass. 488 pp.

Hersokwitz, I.H., 1980. Genetics, 2nd ed. Little Brown and Co., Boston. 554 pp.
Hoffman, R. J. and W.B, Watt, 1974. Naturally occurring variation in larval color
of Colias butterflies: isolation from two Colorado populations. Evol. 28:326-
328.

Maerz, A. andM.R. Paul, 1930. A Dictionary of Color. McGraw-Hill Co., Inc., N.Y.
207 pp.

Manley, T.R., 1978. Genetics of conspicuous markings of the Io Moth. J. Heredity
69 :11- 18.

, 1991. Diapause, voltinism and food plants of Automeris io in the southeastern

states. J. Lepid Soc., in press.

Nijhout, F.H., 1978. Wing pattern formation in Lepidoptera: A model. J. Exp.
Zool. 206:119-136.

, 1980. Pattern formation on Lepidopteran wings: Determination of an

eyespot. Develop Bio. 80:267-274.

, 1981. The color patterns of butterflies and moths, sci. Amer. Nov. 1981:140-

148.

, 1986. Pattern and pattern diversity on Lepidopteran wings. Bio. Sci.

36:527-533.

Packard, A.S., 1914. Monograph of Bombycine Moths of North America. Part 3.
Mem. Natl. Acad. Sci. 12:1:276.

Remington, C.L., 1954. The genetics of Colias (Lepidoptera). Adv. Genetics.
6:404-407.

, 1976. Part male, part female: The expression of mixed sexuality in insects.

Discovery 11:51-59.

Schwantisch , B.N., 1924. On the ground plan of wing pattern of nymphalids and
certain other families of rhopalocerous Lepidoptera. Proc. Zool. Soc. Lond.
34:509-528.

Shapiro, A.M., 1984. Polyphenism, phyletic evolution and the structure of the
pierid genome. J. Res. Lepid. 23(3)177-195.

Stehr, G., 1959. Hemolymph polymorphism in a moth and the nature of sex
controlled inheritance, evol. 13:537-560.

Suffert, F., 1927. Zur Vergleichende Analyse der Schmetterlingszeichnung.
Biol. Zentralbl. 47:385-413.

Journal of Research on the Lepidoptera

29(l-2):54-76, 1990(91)

Patterns of geographic variation and evolution in
polytypic butterflies

Paul C. Hammond

2435 E. Applegate, Philomath, Oregon 97370

Abstract. Patterns of geographic variation and evolution are exam-
ined in polytypic butterflies. It is concluded that genetic and evolution-
ary cohesion at the full species level is limited to monotypic organisms
that are strongly vagile or migratory. Polytypic species with a frag-
mented population structure lack such cohesion, and each isolated
population tends to function as an independent evolutionary unit.
Taxonomic subspecies are usually the products of geographic isolation
and differentiation, and represent separate phylogenetic lineages.
Secondary intergradation between differentiated populations takes
place in hybrid suture zones that often correlate with past geographic
isolating barriers. Vicariance is a common phenomenon at the subspe-
cific level in polytypic butterflies. Older populations survive as relicts
in disjunct refugia, and are frequently surrounded by newer popula-
tions with more recently expanded distributions. There is no clear-cut
distinction between a subspecies and a full species. A complete con-
tinuum is observed among differentiated populations with regard to
reproductive isolating mechanisms and genetic or ecological compat-
ibility. Speciation is strictly a secondary process that may or may not
result from the primary process of macroevolution, the geographic
differentiation of populations adapting to diverse environmental con-
ditions. It is also suggested that macroevolution is usually character-
ized by peripatric cladogenesis and punctuated equilibria. As a conse-
quence, extant insect populations have the dimension of time and
ancestor-descendant relationships.

Introduction

One of the most controversial and perplexing questions in biology
concerns the fundamental nature and evolutionary significance of geo-
graphic variation within species (see Wilson & Brown, 1953; Gillham,
1956; and Pimentel, 1959 for earlier reviews). This controversy still
persists, and Arnold (1985) and Hammond (1986) have recently pre-
sented conflicting points of view. However, much has been learned about
such variation in polytypic butterflies during the past thirty years, and
it may be useful to review these questions in light of our current
knowledge. In addition, basic theories of macroevolution have also
advanced during this time.

Five different phenomena are included under the general subject of
geographic variation. These include (1.) genetic gradients within cohe-
sive populations called primary dines, (2.) differentiation between popu-
lations resulting from geographic isolation, (3.) secondary intergradation
between previously isolated populations along hybrid suture zones, (4.)

29(l-2):54-76, 1990(91)

55

vicariance and polytopic variation, and (5.) incipient speciation. In the
present paper, I will consider lepidopteran examples of these phenomena
and the taxonomic problems surrounding them. Current theories of
macroevolution and punctuated equilibria are also of interest to this
discussion.

Population Structure of Species

Much of the controversy regarding the nature of geographic variation
is rooted in differing individual views of population structure within
biological species. Mayr ( 1963) has presented one of the more comprehen-
sive treatments of geographic variation in animal species. He believed
that biological species are characterized by an internal genetic cohesion
and homeostasis maintained by gene flow throughout all of the reproduc
tively linked populations of the species. As a consequence, Mayr (1963,
p. 348) concluded that geographic races or subspecies could not function
as independent evolutionary units, and that geographic variation within
species is of limited significance. He also suggested that speciation is the
most basic process of evolutionary change. This process apparently takes
place in complete geographic isolation by a so-called “genetic revolution,"
when the newly evolving population would be protected from the disrupt-
ing effects of external gene flow.

Many authors have embraced Mayr’s point of view. For example,
Rindge (1987) has recently expressed the opinion that geographic varia-
tion is of little or no scientific interest unless complete isolation exists
between populations. Gould & Johnston (1972) have favorably reviewed
multivariate studies that portray geographic variation as continuous
with no spatial disjunctions, consisting of simple genetic perturbations
within cohesive and panmictic gene pools. Futuyma (1979, p. 198) stated
that “Infraspecific categories are simply constructs of our imagination,
erected for the sake of convenience; we can recognize as few or as many
races as we find convenient, for they have no independent biological
reality". Likewise, Eldredge & Cracraft (1980, p. 16) have stated “We
define species in such a way as to stress their internal cohesion, their
identity as discrete, real entities, and their unique position as phyloge-
netic units. No taxon other than species serves as ancestors and descen-
dants (i.e. as phylogenetic units) in evolution."

Most recently, some theorists have suggested that biological species
may be viewed as phylogenetic “individuals” (Eldredge, 1989). Such
entities experience birth (speciation), death (extinction), and selection
(group selection) in a manner analogous to that of individual organisms.
Of course, such theory depends upon the purported genetic cohesion of
the biological species.

In sharp contrast, other authors have seriously questioned this concept
of the cohesive species (Epling & Gatlin, 1950). Ehrlich & Raven (1969)
argued that many species are actually quite sedentary and display a
highly fragmented population structure, with little or no gene flow

56

J. Res. Lepid.

among the isolates. They also suggested that the real unit of evolution in
such organisms is the locally isolated population, rather than the
collective species. In the complete absence of gene flow, the only thing
that local populations of a fragmented species really share with each
other is a common phylogenetic ancestry, perhaps combined with similar
stabilizing selection for the ancestral characteristics (Ehrlich & White,
1980).

The idea that geographic isolates constitute discrete evolutionary units
and independent phylogenetic lineages is not new among lepidopterists.
Rothschild & Jordan (1906, p. 429) regarded subspecies as incipient new
species and basic units of evolutionary change, a view also shared by
Fruhstorfer (H. Descimon, per. comm.) . Indeed, Karl Jordan was one of
the first advocates of the modern subspecies concept (Mayr, 1976 p. 303).
Grey & Moeck (1962) considered this with their discussion of geographic
variation in Speyeria and other polytypic butterflies. Grey, Moeck, &
Evans (1963) suggested that “the largest factor controlling population
structure in butterflies is the residual inertia of genetic heritage, infer-
ring that wing facies reflect earlier dispersal and isolation, relating only
secondarily to present situations”. Similar views were expressed by
Bowden (1979) regarding variation in Pieris. Moreover, the philosophy
behind the taxonomic classification of subspecies is based upon the
existence and function of geographic isolates as independent evolution-
ary units and phylogenetic lineages.

Geographic variation within a cohesive and panmictic gene pool is of
limited evolutionary and taxonomic significance as Mayr and other
authors have rightly suggested. However, the genetically cohesive spe-
cies is probably a relatively rare phenomenon in nature, at least among
butterflies. Some species do appear to conform to the Mayrian model of
cohesive population structure. These are usually migratory or highly
vagile organisms, and they are generally monotypic with little or no
recognizable geographic variation. Among North American butterflies,
possible examples of cohesive species include Danaus plexippus (L.),
Vanessa cardui (L.), Colias eury theme Bdv., Speyeria idalia (Drury), and
S. edwardsi (Reakirt). The last two species occupy the native tail-grass
and short-grass prairies of the Great Plains, and their strong vagility
appears to be an adaptation for quickly recolonizing large areas of
habitat burned by periodic prairie fires. In sharp contrast, polytypic
species in genera such as Argynnis, Speyeria , Euphydryas , Melitaea,
Limenitis, Papilio, Colias , Cercyonis , Erehia , Coenonympha, Lycaena,
Gallop hrys, and most polyommatine blues display a highly fragmented
population structure that conforms to the Ehrlich-Raven model.

Differentiation of Populations

It is very important to distinguish between a primary genetic gradient
within a cohesive population and differentiation between populations
resulting from geographic isolation. These are entirely different phenom-

29(l-2):54-76, 1990(91)

57

ena. Much discussion has appeared in the literature concerning this
distinction, although many authors have felt that the two processes are
difficult or impossible to distinguish in actual situations {i.e. Mayr, 1963
p. 380; Endler, 1977).

I tend to disagree with this view. Primary genetic gradients usually
involve a simple polymorphism at a single locus or a few loci of a polygenic
complex. It is not difficult to recognize such polymorphisms in actual
populations. Good examples in Lepidoptera include the phenomenon of
industrial melanism in moths such as Bistort betularia L., the black and
yellow morphs of Papilio glaucus L. andP. bairdi Edwards, and the alba
female morph of Colias species. In such cases, the genetic basis of the
polymorphism is usually quite simple (Robinson, 1971). However, most
polymorphisms distributed along primary dines are probably not evi-
dent in the external morphology of the organism, but are internally
“hidden” as with allozyme frequencies. The enzyme polymorphisms
studied by Johnson (1976) in Colias populations are an example.

By contrast, the genetic differentiation that takes place between
isolated populations commonly involves a major restructuring of the
over-all genome, affecting many different loci on different chromosomes
controlling completely unrelated characteristics. This involves not only
changes in external morphology such as wing color pattern, but also
changes in larval and pupal characters, ecological adaptations to differ-
ent foodplants and temperature regimes, changes in allozyme systems
and developmental rates, and changes in behavior. This is the so-called
“genetic revolution” emphasized by Mayr (1963), characterized by a
major reorganization of polygenic balances (Carson, 1982).

I believe that Mayr is correct in his view that geographic isolation is the
primary factor behind major evolutionary divergence between popula-
tions. There are alternative theories for significant differentiation within
a cohesive gene pool along a primary clinal gradient, including models for
parapatric and sympatric speciation (Endler, 1977; White, 1978). How-
ever, such differentiation at dozens or even hundreds of independent loci
on different chromosomes would appear to be very difficult to attain
within a cohesive gene pool due to the disruption of gene flow and sexual
recombination.

I also believe that most of the butterfly subspecies listed by Miller &
Brown (1981) are in fact the products of geographic isolation. Arnold
(1985) has been critical of “minor” subspecies that differ by only one or
two characters of wing color pattern, implying that such variation may
be a single allelic substitution along a primary clinal gradient. However,
in most cases this view is not correct. “Minor” subspecies that only differ
by a single morphological character are often found to differ by many
other “hidden” biological characters as well. For example, McCorkle &
Hammond (1988) found a number of biological differences between
similar subspecies of Speyeria zerene (Bdv.). A particularly good example
of “minor” subspecies are the tiger swallowtails of eastern North America,

58

J. Res. Lepid.

Papilio glaucus glaucus, P. g. canadensis Roths. & Jordan, and P. g.
australis Maynard. These three races are difficult to consistently distin-
guish on a morphological basis, but exhibit major biological differences
in foodplant adaptations and pupal diapause characteristics (Scriber,
1986; Rockey, Hainze, & Scriber, 1987). Thus, even “minor” subspecies
have evolutionary (adaptive) significance.

Microgeographic variation within a taxonomic subspecies is also of
interest. For highly sedentary organisms, the subspecies has no more
genetic or evolutionary cohesion than the full species. Indeed, many
subspecies are fragmented into locally isolated populations that exhibit
evidence of independent differentiation, and such populations only share
a common phylogenetic ancestry. In western North America, each
mountain range may have an endemic microrace, and individual moun-
tains within a range may have local colonies or demes that are partially
or completely isolated from other colonies. The evolutionary divergence
within a complex polytypic species often shows a hierarchical arrange-
ment consisting of local demes, microraces, subspecies,, and major sub-
species groups.

F or example, Speyeria callippe elaine dos Passos & Grey is a distinctive
subspecies belonging to the callippe subspecies group along the West
Coast. However, S. c. elaine is not homogeneous in western Oregon, and
consists of five differentiated microraces distributed in (1.) high eleva-
tions of the Siskiyou Mountains, (2.) low elevations of the Siskiyous, (3.)
low elevations north of the Siskiyous, (4.) high elevations in the Cascade
Range, and (5.) low elevations in the Willamette Valley. Moreover, the
microraces are comprised of local colonies that may be separated by five
to ten miles. Such demes often exhibit minor differentiation, particularly
if habitat conditions vary in different areas. In sedentary organisms, the
local deme is probably the basic evolutionary unit, rather than the
microrace, subspecies, or full species. In more vagile organisms such as
Speyeria coronis (Behr), the taxonomic subspecies probably has some
evolutionary cohesion.

Hybrid Suture Zones

If most taxonomic subspecies of polytypic butterflies are the result of
evolutionary divergence during geographic isolation, clinal intergrada-
tion among such races is the result of secondary contact along hybrid
suture zones. The characteristics of suture zones have received consider-
able attention in the literature (Mayr, 1963; Remington, 1968), and
include three types of secondary intergradation. Many subspecies join
together along an abrupt step-cline. This phenomenon may result from
considerable genetic incompatibility between the subspecies or reduced
ecological fitness of hybrid individuals. Indeed, Barton & Hewitt (1983)
suggest that hybrid zones may serve as strong barriers to gene flow if
multiple loci and characters are simultaneously involved in the geo-
graphic divergence. Conversely, the suture zone between many other

29(l-2):54-76, 1990(91)

59

subspecies forms a long, gradual cline, indicating good genetic compat-
ibility and ecological fitness of hybrids.

In the third type of intergradation, hybrid individuals exhibit superior
ecological fitness over both of the parental genotypes, resulting in a fully
developed reticulate fusion between previously differentiated subspe-
cies. In such populations, the original parental phenotypes only appear
as rare recombinants, and most individuals are of the intermediate
hybrid phenotype. This last type of intergradation may be an uncommon
occurrence in nature, however. Of the some 110 subspecies within the
genus Speyeria, I have identified only six that appear to be of reticulate
origin.

One example are the highly variable populations of Speyeria mormonia
(Bdv.) in the northern Rocky Mountain and Canadian prairie regions.
These may have originated from a hybrid fusion between S. mormonia
eurynome (Edwards) of the southern Rocky Mountains and S. m. artonis
(Edwards) of the Great Basin. Likewise, the Utah populations of S.
atlantis chitone (Edwards) ( -wasatchia dos Passos & Grey) are highly
variable and intermediate between S. a. nikias (Ehrmann) of the south-
ern Rocky Mountains and S. a. tetonia dos Passos & Grey of the northern
Rocky Mountains.

It is useful to examine the intergradation between subspecies in
greater detail. In most cases, hybrid suture zones are located along some
type of geographic isolating barrier that either exists today or once
existed in the past. Indeed, the distributions of many subspecies of North
American butterflies strongly correlate with the major biogeographic
regions and suture zones identified by Remington (1968).

As an example, Speyeria aphrodite (Fabr.) occupies a continuous
distribution over much of eastern North America, and there are no
geographic barriers within this region at present. Nevertheless, three
distinctive subspecies join together near the Great Lakes as shown in
Figure 1. Moreover, the hybrid suture zones among these races are fairly
abrupt, suggesting some degree of incompatibility. The typical S. a.
aphrodite is an East Coast subspecies with a westward extension into the
Great Lakes region, S. a . alcestis (Edwards) occupies the native tail-grass
prairies of the southern Great Plains, and S. a. manitoha (Chermock)
occupies the northern Great Plains and Rocky Mountains.

Although these subspecies are not isolated at present, it is known that
the Great Lakes region was buried under deep glacial ice fields during the
Pleistocene some 15,000 years ago (Wells & Stewart, 1987). At that time,
the three S. aphrodite subspecies were probably isolated in widely
disjunct refugia along the East Coast, on the southern Great Plains, and
in the northern Rocky Mountain region respectively. Following the
retreat of the glaciers, the subspecies must have expanded their ranges
to join together in the modern Great Lakes suture zone.

In western North America, high mountain ranges and lowland deserts
have combined with the climatic fluctuations of the Pleistocene to form

60

J. Res. Lepid.

Fig. 2. Distribution of Speyeria callippe subspecies in western North America.

29(l-2):54-76, 1990(91)

61

strong isolating barriers at various times during the past two million
years. Specifically, the mountains would serve as barriers during cold
periods of glacial maxima, while lowland deserts would serve as barriers
during warm interglacial periods. As a consequence, many butterfly
genera exhibit extremely complex patterns of speciation and subspeciation
within this region.

A good example is the geographic variation of Speyeria callippe (Bdv.)
shown in Figure 2. The callippe group of subspecies are isolated along the
West Coast from S. c. semivirida (McD.) and S. c. nevadensis (Edwards)
by the high Cascade and Sierra Nevada ranges. Likewise, the high
mountains along the Continental Divide have served to isolate two
subspecies east of the Divide, including S. c. meadi (Edwards) along the
Colorado Front Range and S . c. calgariana (McD.) on the northern Great
Plains. In sharp contrast, lowland deserts have served to isolate S. c.
semivirida , S. c. nevadensis , and S. c. harmonia dos Passes & Grey in the
Great Basin and Intermountain regions. In fact, one of the major
biogeographic suture zones discussed by Remington (1968) extends
between these regions as shown by the lorquini-weidemeyeri suture zone
in Limenitis and the euryalus- glover i suture zone in Hyalophora. In
addition to S. callippe , other species of Speyeria that exhibit an identical
biogeographic distribution of subspecies across this region include S.
mormonia , S. zerene, and S. egleis (Behr)

Polytopic Variation and Vicariance

One of the major criticisms directed towards the subspecies concept is
the frequent spatial disjunction and discordant distribution of indepen-
dent characters within a biological species (Gillham, 1956; Gould &
Johnston, 1972; Arnold, 1985). This has been called “polytopic variation”
(Mayr, 1963), and it is frequent in Speyeria callippe populations along the
West Coast (Arnold, 1985). Such patterns may be explained as local and
independent fluctuations in gene frequencies within a cohesive and
panmictic gene pool. However, this can only be true for species with a
cohesive population structure of the Mayrian model. For species with a
fragmented population structure of the Ehrlich-Raven model, alterna-
tive explanations for polytopic variation must be considered. In addition,
discordance among diagnostic taxonomic characters is frequently ob-
served at the full species level (Hammond, 1986). Different species share
different combinations of the same characters as subspecies. Of the 13
species of Speyeria, only four have consistent diagnostic characters that
are not present in other species, and which do not vary geographically.

One explanation for discordance among different characters is conver-
gent adaptation to similar environmental conditions by populations only
distantly related to each other. A prime example is seen in pallid
subspecies of Speyeria atlantis (Edwards). Most populations of this
species live under cool, moist conditions, and display heavy, melanic
basal suffusion on the dorsal wing surfaces combined with very dark disc

62

J. Res. Lepid.

colors on the ventral hindwing. However, a number of »$. atlantis
subspecies are found in areas with warm, dry conditions, and these
exhibit very pale wing colors. Such taxa include S. a. Helena dos Passos
& Grey on the Canadian prairies, S. a. ratonensis Scott in northeastern
New Mexico, S. a. greyi Moeck in the Ruby Mountains of Nevada, S. a.
elko Austin in the Independence and North Humboldt Mountains of
Nevada, and S. a. Irene (Bdv.) in the Sierra Nevada of California. Not only
are these five subspecies widely disjunct in distribution, but their pale
wing colors are quite discordant with other characters of the wing
pattern. One may reasonably conclude that these populations are not
closely related, and have acquired similar pale colors through indepen-
dent convergence.

A second explanation for polytopic variation is vicariance (Rosen, 1978;
Erwin, 1981). In species with a fragmented population structure, iso-
lated populations tend to function as independent evolutionary units and
phylogenetic lineages. Ancient subspecies that once enjoyed wide distri-
butions may now survive only as relicts in widely disjunct refugia, while
newer subspecies may now surround the refugia of the older subspecies.
In these situations, there are often sharp ecological differences between
the subspecies, which allows the older populations to survive within their
restricted refugia. In addition, hybrid suture zones between the subspe-
cies are often abrupt step-clines resulting from some degree of incompat-
ibility. Otherwise, the older populations would tend to experience genetic
swamping from the newer populations.

The recognition of vicariance is dependent upon two important factors.
First, one must consider a time dimension for populations or taxa, since
vicariance suggests a distribution through time. This will be discussed
later under processes of macroevolution. Second, convergent similarities
must be distinguished from phylogenetic similarities, not only to recog-
nize examples of convergence or vicariance, but to establish a natural
taxonomic classification. This is not always easy. Problems with charac-
ter interpretation are ultimately reflected in taxonomic difficulties with
the group in question. In Speyeria evolution, wing markings appear to be
highly conservative and reliable as diagnostic characters, while wing
colors are less stable. However, the darkness of color ( i.e . melanic basal
suffusion) is extremely plastic, and subject to repeated convergence and
reversal (homoplasy). This is seen with the phylogenetic interpretations
of S. callippe (Fig. 3, Table 1).

There is much evidence that vicariance is a relatively common phenom-
enon among polytypic butterflies in such diverse genera as Papilio,
Colias , Euphydryas , Speyeria , Coenonympha, Lycaena, Callophrys, and
Icaricia. For example, Speyeria atlantis atlantis is widely distributed in
the Appalachians and across Canada to Alaska, but it also occurs in
widely disjunct refugia through the Rocky Mountains from southern
British Columbia to northern New Mexico. These refugia are surrounded
by more divergent and probably newer subspecies of S. atlantis (Ferris,
1983).

29(l-2):54-76, 1990(91)

63

Because most subspecies represent discrete evolutionary units, a
phylogenetic analysis can be applied to them in a study of vicariance as
discussed by Cracraft (1982). Thorpe (1984) used this approach with
European snakes in order to distinguish between primary and secondary
dines. As an example, a cladistic analysis of Speyeria callippe subspecies
is shown in Figure 3, with individual character changes listed in Table
1. The species is probably derived from a West Coast isolate of the
Appalachian-type S. a. atlantis. Indeed, the Oregon S. c. elaine is still
remarkably similar to S. a. atlantis , while the California S. c. callippe and
S. c.juba (Bdv.) are slightly more divergent from the putative ancestral
type. One daughter species is apparently derived from S. callippe.
Speyeria edwardsi probably evolved from a population of *S. c. semivirida
that became isolated on the northern Great Plains east of the Continen-
tal Divide in the same manner as S. c. calgariana, but at a much earlier
time. In other words, S. edwardsi is likely a Pliocene or early Pleistocene
isolate, while S. c. calgariana is probably an isolate of the late Pleisto-
cene.

However, the oldest subspecies and the most complex geographic
variation are found with the callippe group distributed along the West
Coast in Oregon and California (Fig. 4). In phylogeny, this group divides

64

J. Res. Lepid .

callippe - comstocki
juba - macaria
elaine - liliana
rupestris - inornata
semivirida - nevadensis

Fig. 4. Distribution of Speyeria callippe subspecies along the West Coast.

into four distinctive pairs of subspecies, including (1.) S. c. callippe and
S. c. comstocki (Gunder), (2.) S. c.juba and S. c. macaria (Edwards), (3.)
S. c. elaine and S. c. liliana (Hy. Edwards), and (4.) S. c. rupestris (Behr)
and S. c. inornata (Edwards). But when the phylogenetic pairs are
compared with the distributions shown in Figure 4, wide disjunctions are
apparent.

The callippe-comstocki pair has a continuous distribution in the south-
ern California Coast Range. However, the juba-macaria pair is frag-
mented into three distinct isolates in the northern California Coast
Range, in the northern Sierra Nevada, and in the Tehachapi Mountains.
Likewise, the liliana- elaine pair is widely disjunct between Napa and

29(l-2):54-76, 1990(91)

65

Lake Counties on the central California coast and western Oregon.
Finally, the rupestris-inornata pair is apparently derived from S. c.
elaine, and occupies the Salmon-Trinity Mountains of northwestern
California, extending southward in the western foothills of the Sierra
Nevada to Tulare County.

Thus, the coastal S. c.juba population in Glenn, Tehama, and Mendocino
Counties is inserted between the unrelated populations of S. c. Uliana to
the south and S. c. rupestris to the north. Likewise, the Sierran S. c.juba
population is inserted between S. c. inornata to the west and S. c.
nevadensis to the east, while S. c. macaria is connected to S. c. comstocki
in the Coast Range and to S. c. inornata in the Sierras through the
intermediate hybrid population called S. c. laurina (Wright) in the
Greenhorn Mountains. Although hybridization is observed among all of
these subspecies pairs, the suture zones are usually sharp step-clines
that suggest some degree of incompatibility. Extensive reticulate fusion
is only evident in the laurina population. In addition, sharp ecological
differences are also present among the inornata, juba, and nevadensis
populations of the northern Sierra Nevada.

The present distribution pattern is consistent with alternating con-
tractions and expansions in diverse S. callippe populations during the
climatic fluctuations of the Pleistocene. During cool glacial periods, the
species may have disappeared from the mountains of northwestern
California, only to re-expand into this region during warm interglacial
periods. Originally only three subspecies were likely present along the
West Coast if this interpretation is correct. These include the common
ancestor of callippe-comstocki in the southern Coast Range, ancestral
Uliana- elaine in the northern Coast Range and western Oregon, and
ancestral juba-macaria in the Sierra Nevada. An early glacial period
could be responsible for the initial disjunction between S. c. Uliana and
S. c. elaine , followed by an expansion of the Sierran S. c. juba into the
Coast Range during a subsequent interglacial period. Likewise, a later
glacial period may have resulted in the disjunction of the coastal and
Sierran juba populations. During a still later interglacial period, S. c.
rupestris expanded and evolved from the Oregon S. c. elaine in the
Salmon-Trinity Mountains, and eventually spread southward in the
western foothills of the Sierra Nevada as S. c. inornata. Although this
evolutionary hypothesis is complex, it would explain the complicated
geographic variation present today along the West Coast.

Incipient Speciation

Patterns of geographic variation are also very complex in various
degrees of incipient speciation, a process that is usually an extension of
isolation and vicariant disjunction. Because polytypic species appear to
lack genetic and evolutionary cohesion, the significance of speciation is
largely of an ecological nature. Through the acquisition of reproductive
isolation, closely related populations are able to co-exist in sympatry,

66

J. Res. Lepid.

dividing available resources into ecological niches. Such resource parti-
tioning results in greater diversity and stability for the total ecosystem.
Thus, Speyeria edwardsi and S. callippe are widely sympatric on the
northern Great Plains.

The term “semispecies” has been applied to geographic segregates that
exhibit a trend towards speciation (reproductive isolation), and polytypic
species that consist of semispecies have been called “superspecies”
(Mayr, 1963). In nature, there is often no clear-cut distinction between a
geographic subspecies and a fully distinct biological species as Ehrlich
(1961), Lorkovic (1962), Ehrlich & Murphy (1983), and Clarke & Larsen
(1986) have discussed with butterflies. This fact supports Mayr’s belief
that most speciation is an allopatric process.

Because speciation is an important result of the geographic differentia-
tion of populations, it is useful to examine this process in greater detail,
and butterflies provide many examples at four different stages of diver-
gence. These include (1.) divergent subspecies with no reproductive
isolation, (2.) divergent subspecies with reproductive isolation in re-
stricted local areas, (3.) allopatric populations that are intermediate in
morphology between fully distinct sympatric species, and (4.) fully
distinct sympatric species that exhibit reticulate hybrid fusion in re-
stricted local areas.

There are many examples of highly, divergent subspecies that lack
reproductive isolating mechanisms, as with the Speyeria callippe sub-
species along the West Coast. Most of these populations have been
regarded as distinct taxonomic species in the past (dos Passes & Grey,
1947). Other examples in North America include the Limenitis arthemis
complex (Platt, 1983), the Papilio glaucus-rutulus complex (Brower,
1959), and the Papilio machaon complex. In the latter group, the taxa
asterius Cramer, zelicaon Lucas, and hudsonianus Clark have always
been regarded as distinct species, and they are quite divergent in both
wing color pattern and allozyme patterns (Sperling, 1987). Nevertheless,
these taxa lack reproductive isolation, and they form extensive hybrid
swarms within their suture zones. Moreover, there is evidence that this
secondary intergradation is not a recent or temporary phenomenon, but
has persisted for hundreds or even thousands of years since the last
glaciation (Sperling, 1987). Similar stability of hybrid swarms has also
been detected in Hyalophora moths (Collins, 1984).

Other highly divergent subspecies intergrade in most areas of their
ranges, but exhibit reproductive isolation in a few local areas of overlap-
ping sympatry. For example, Speyeria atlantis subspecies exhibit exten-
sive intergradation throughout most of the species’ range. Indeed,
populations in the Rocky Mountains of Montana are joined together in a
massive, three-way hybrid swarm between the northern S. a. heani
(Barnes & Benj.), the central S. a. tetonia, and S. a. Helena of the
Canadian prairies.

However, in some local areas, S. atlantis subspecies co-exist together

29(l-2):54-76, 1990(91)

67

in sympatry with strong reproductive isolation, for example in the Riding
Mountains of Manitoba (Moeck, 1957) and the Black Hills of South
Dakota (Grey, Moeck & Evans, 1963). Likewise, Ferris (1983) and Scott
(1988) have recently looked at the partial segregation between S. a.
atlantis {-electa Edwards) and S. a. hesperis (Edwards) along the
Colorado Front Range. My own field observations in western Colorado
indicate that S. a. atlantis and S. a . nikias are widely sympatric with
strong reproductive isolation south of the Gunnison River. However, the
S. atlantis populations north of this river are an intermediate mixture
between typical atlantis and nikias , apparently resulting from a hybrid
fusion between these races.

A slightly different version of incipient speciation involves sympatric
populations with allochronic flight periods. Mattoni (1989) has provided
several examples of sympatric allochrony in the Euphilotes battoides
complex. David V. McCorkle and I have observed a good example in
Euphydryas editha (Bdv.). Two subspecies of similar phenotype are
distributed along the western slopes of the Cascade and Sierra Nevada
ranges. The Cascadian race is named E. e. colonia (Wright), the larvae
feed on Castilleja spp. , and the adults fly in June and July. By contrast,
the Sierran race is named E. e. rubicunda (Hy. Edwards), the larvae feed
on C ollinsia spp. , and the adults fly in April and May (Ehrlich & Murphy,
1981). Very similar populations are also found in the Siskiyou Mountains
of southwestern Oregon and adjacent California. These have been called
“baroni” by Dornfeld (1980), although the true E. e. baroni (Edwards) is
apparently restricted to the California coast in Mendocino County
(Murphy, 1983).

In fact, the Siskiyou populations consist of two sympatric and allochronic
units, although the adults are nearly identical in phenotype. A rubicunda-
like population lives in dry, rocky habitats, and the adults fly from late
March through May with oviposition on Collinsia. A sympatric colonia-
like population lives in Darlingtonia bogs and riparian areas, and the
adults fly in June and July with oviposition on Castilleja. Although the
adults occur in the same areas with a mixture of bogs and rocky outcrops,
the different flight periods must preclude most gene exchange between
these populations. In terms of reproductive isolation, they are function-
ing as distinct species.

The third situation arises from the geographic fragmentation of popu-
lations that precedes speciation. These are allopatric populations that
are intermediate in morphology between distinct sympatric species, and
are difficult to classify at the species level. For example, Speyeria zerene
and S. coronis are sympatric throughout most of the western United
States. Although closely related, they clearly function as distinct species.
However, two allopatric populations are intermediate between these
species, including carolae (dos Passos & Grey) in the Spring Mountains
of Clark County, Nevada and semiramis (Edwards) in southern Califor-
nia. The carolae population was originally described as a S. coronis

68

J. Res. Lepid.

subspecies, but was later transferred to S. zerene by Grey & Moeck
(1962). Likewise, the very similar semiramis exhibits reproductive link-
age with other S. coronis populations, but it still retains many of the
characteristics of S. zerene.

There are similar examples in other genera such as Colias and Papilio.
Colias pelidne Bdv. & LeC. and C. gigantea Strecker are widely sympa-
tric in the northern Rocky Mountains, and are fully distinct with
different larval foodplants (Ferris, 1987). However, an allopatric popula-
tion in the Rocky Mountains of Colorado is intermediate and apparently
feeds on both foodplants. Ferris (1987) has chosen to deal with this
difficult taxonomic problem by treating the Colorado population as a
third distinct species, C. scudderi Reakirt. Likewise, Papilio astyalus
Godart andP. androgens Cramer are sympatric through much of tropical
America. However, an allopatric population named P. thersites Fabr. is
isolated on the island of Jamaica, and is intermediate in wing pattern
between the continental species. Although treating allopatric popula-
tions as distinct species provides an easy solution to embarrassing
taxonomic problems, it also obscures the intermediate transitions be-
tween sympatric species. Such transitions are quite common in nature,
but their existence is not recognized by the Linnaean system of taxo-
nomic nomenclature.

Finally, the fourth situation involves widely sympatric species that
remain distinct in most regions, but exhibit hybridization and reticulate
fusion in certain local areas. Euphydryas chalcedona (Dbl.) andP. anicia
(Dbl.) are usually distinct species in most parts of their sympatric ranges
(Ferris, 1988a), but reticulate fusion between them is evident in parts of
Nevada and northeastern California (Scott, 1980). A similar situation
exists between Colias pelidne and C. interior Scudder. Although these
species appear to remain distinct in most parts of their ranges, apparent
hybrid populations are found in Idaho (Ferris, 1988b). There is also
evidence that local areas of hybridization and fusion occur in several
other butterflies of the Pacific Northwest, including hybrid swarms
between Colias occidentalis Scudder and C. alexandra Edwards, and
between Lycaeides melissa (Edwards) and L. idas (L.) (unpublished
data).

Processes of Macroevolution

The theory of macroevolution has changed considerably since Darwin,
and even since formulation of the “modern synthesis”. Stebbins & Ayala
(1981) have suggested that macroevolution is an autonomous field of
evolutionary study, because macroevolutionary patterns cannot be de-
duced from the microevolutionary principles of mutation, gene flow,
random drift, and natural selection. Thus, the processes of macroevolu-
tion can only be examined by studying actual patterns of past divergence
and adaptive radiation.

Various theories of macroevolution have been reviewed by Mayr ( 1963,

29(l-2):54-76, 1990(91)

69

1976), Stanley (1979), Eldredge & Cracraft (1980), and Eldredge (1989).
At first, it was thought that most evolution took place by phyletic
gradualism within cohesive gene pools. This model suggested that
established populations slowly change over long periods of time in
response to gradually changing environmental conditions. However, the
fossil record provided support for an entirely different pattern of macro-
evolution; one characterized by short, explosive bursts of evolutionary
change followed by long periods of relative stasis. Such evidence resulted
in theories of saltation, quantum evolution, and punctuated equilibria
(Gould & Eldredge, 1977).

Under the model of punctuated equilibria, macroevolution does not
take place by phyletic change within established populations, but through
the creation of entirely new populations (cladogenesis). Moreover, the
process rarely involves a dichotomous splitting of a pre-existing ances-
tral population into two daughter populations. Instead, macroevolution
is usually associated with peripheral budding (peripatric cladogenesis)
at the perimeter of the ancestral population’s range where environmen-
tal conditions are different (Mayr, 1976 p. 455; 1982). The ancestral
population remains intact and unchanged within the original environ-
ment, while the new daughter population actively invades and adapts to
an entirely new environment (Eldredge & Cracraft, 1980 p. 125). Indeed,
only a few generations of intense directional selection within a small
founder population maybe required to achieve significant adaptive shifts
(Ford, 1945 pp. 268-270; Carson & Templeton, 1974; Dimock & Mattoni,
1986). Thus, peripatric cladogenesis combined with adaptive shifts to
new environmental conditions are now thought to be the most important
factors behind macroevolution.

However, Mayr and Eldredge still believe that cladogenesis is essen-
tially synonymous with speciation due to the purported cohesion of the
biological species (Eldredge, 1989). This view is rejected in the present
paper for the reasons previously discussed. Instead, I would argue that
the basic process of macroevolution is the geographic differentiation of
populations, and not necessarily speciation. Likewise, if one wants to
directly observe the process of macroevolution, they should examine
patterns of geographic variation, and not patterns of speciation.

There are other far-reaching implications. It is frequently argued that
populations or taxa do not have the dimension of time, because a
population of today is never precisely identical to a population of yester-
day. Thus, no living population can be directly ancestral to any other
living population. While this argument is partly a matter of semantic
definition, it is also a corollary of phyletic gradualism. However, if
punctuated equilibria and peripatric cladogenesis are real phenomena,
then populations do have dimensions of both time and space. Presently
extant populations may be regarded as “ancient” or “recent” on the time
scale, and may be regarded as “ancestral” or “descendant” in phylogeny.

In addition, ancestral forms of insects face a low probability of extinc-

70

J. Res . Lepid,

tion as descendants evolve, in contrast to the extinction patterns of large
vertebrates seen in the fossil record. Insects are able to survive as relicts
in small, isolated refugia that could never support a viable population of
large vertebrates. At the full species level, reproductive isolation and
resource partitioning allow large numbers of related insect populations
to co-exist in sympatry, including both ancestral and descendant popu-
lations. For example, eight species of Speyeria are widely sympatric in
western North America. But the four species of the primate genus Homo
follow each other sequentially in the fossil record, and never co-existed
together for long periods of time (Rightmire, 1985). Although these
primates present the pattern of punctuated equilibria, paleontologists
cannot decide if Homo evolution has been phyletic or cladogenic (Eldredge,
1989 p. 75). There is no such doubt with extant butterfly taxa. It is only
through cladogenesis that related populations can exist simultaneously
in time.

The previous examples of Speyeria atlantis and S. callippe serve to
illustrate punctuated equilibria, peripatric cladogenesis and ancestor-
descendant relationships. In particular, the Appalachian-type S. a.
atlantis appears to represent a basal-stem ancestor within the genus
Speyeria, and is a prime example of punctuated equilibria. This subspe-
cies displays a classical vicariant pattern, with a continuous Appala-
chian population and three widely disjunct Rocky Mountain populations.
One extends from northern Idaho to southeastern British Columbia, a
second occurs in the Big Horn Mountains of Wyoming, and a third occurs
in Colorado. A fourth population in the Black Hills of South Dakota is
more derived, and appears to be closely related to S. a. hollandi (Chermock)
in the Riding Mountains of Manitoba. All of these populations are
surrounded by highly derived subspecies of S. atlantis.

There has certainly been no direct genetic contact among the Rocky
Mountain populations of S. a, atlantis for the past 12,000 years and
generations since the last glacial period, and possibly for much longer
considering the distributions of the surrounding subspecies. Genetic
contact between the Appalachian and Rocky Mountain populations has
clearly been broken for a very long time. Yet no major differentiation has
taken place among these populations. This long-term stasis is closely
correlated with the ecology of S. a. atlantis , since all disjunct isolates of
this subspecies still occupy a similar habitat consisting of cold, wet
spruce or birch forests. In sharp contrast, the surrounding subspecies
occupy warm, dry forests of pine, fir, and aspen, or even open grasslands
on the Canadian prairies.

The cladogram in Figure 3 depicts the distribution of derived charac-
ters among the taxa of the Speyeria callippe group. However, cladograms
do not give a time dimension to taxa, and simply arrange taxa in a
dichotomous branching pattern. If cladogenesis takes place by periph-
eral budding, with ancestral populations remaining intact and mostly
unchanged, then the actual phylogeny should be represented by ances-

29(l-2):54-76, 1990(91)

71

caigariana meadi

\

harmonia

nevadensis inornata

atlantis

Fig. 5. Phylogenetic interpretation of Speyeria caiiippe subspecies.

tor-descendant relationships among extant taxa in the absence of extinc-
tions. Such an interpretation for the S. caiiippe group is presented in
Figure 5, and is based upon the cladistic data combined with the
distributions shown in Figures 2 and 4.

This family tree suggests that S. caiiippe elaine originated as a West
Coast isolate from S. a. atlantis , probably in western Oregon. Aside from
characters 1 and 2 ,S.c. elaine only differs from S. a, atlantis by a single
derived character (13), and this is a rather weak and inconsistent
character. Moreover, S. c. elaine still occupies a forest habitat like S. a.
atlantis , although it is a warm, dry forest of oak, pine, and fir. The
vicariance of S. c. elaine and the similar S. c. Uliana is also suggestive of
considerable antiquity.

Other subspecies of S. caiiippe in California are more divergent, and
some races have shifted from the ancestral forest habitat to xeric
grasslands (i.e. comstocki , macaria , inornata). However, the most ex-
treme divergence shown in Figure 3 separates the semivirida group from
the caiiippe group. This change is closely correlated with an extreme

72

J. Res. Lepid.

ecological shift, from the ancestral forest habitat to the semi-desert
sagebrush steppes of the Great Basin and intermountain regions. Two of
these taxa, S. c. calgariana and S. edwardsi, have moved east of the
Rocky Mountains to occupy the xeric short-grass prairies of the northern
Great Plains. In other words, the degree of morphological divergence is
directly related to ecological divergence, from the cold, wet spruce forests
of S. a. atlantis to the hot, dry grasslands of S. c. calgariana.

I believe that these examples provide at least indirect, circumstantial
evidence in support of the theories of peripatric cladogenesis and punc-
tuated equilibria. From the probable point of origin on the West Coast,
Speyeria callippe populations have evolved and spread by the peripheral
budding process southward and eastward across much of western North
America. Lowland deserts and high mountain ranges combined with
Pleistocene climatic fluctuations have likely served as isolating barriers
during this process. Major morphological changes appear to be associ-
ated with the creation of new, descendant populations combined with
major ecological shifts into new environments. At the same time, ances-
tral populations remain intact and locked into long-term stasis within
their ancestral environments. Over long periods of time, ancestral types
such as S. c. elaine may give rise to multiple descendants that differ
greatly in age of divergence. For example, S. c.juha appears to be much
older than S. c. rupestris by reason of its vicariance.

The above discussion does not mean that phyletic evolution never takes
place within established gene pools. However, such change may be
limited to simple genetic traits, rather than a major restructuring of the
genome. Sperling (1987) has hypothesized that the black wing morph
gene of Papilio polyxenes asterius has been introduced into sympatric
populations of P. bairdi through hybridization, probably in New Mexico
or eastern Arizona. The black gene has since spread throughout south-
western populations of P. bairdi , extending to southern California and
southern Idaho. Such gene flow and phyletic evolution is most likely to
take place in strongly vagile organisms with a cohesive population
structure, and is much less likely in sedentary organisms such as
Speyeria or Euphydryas species.

Acknowledgments. I would like to thank David McCorkle, Henri Descimon, Rudi
Mattoni, and an anonymous reviewer for their helpful comments and ideas.

Literature Cited

Arnold, R.A. 1985. Geographic variation in natural populations of Speyeria
callippe (Boisduval) (Lepidoptera: Nymphalidae). Pan-Pacific Entomologist
61:1-23.

Barton, N. & G. M. Hewitt. 1983. Hybrid zones as barriers to gene flow. In: Protein
polymorphism: adaptive and taxonomic significance. G.S. Oxford & D. Rollinson,
eds. Academic Press, New York. pp. 341-359.

Bowden, S.R. 1979. Subspecific variation in butterflies: adaptation and dissected
polymorphism in Pieris (. Artogeia ) (Pieridae). J. Lepid. Soc. 33:77-111.

29(l-2):54-76, 1990(91)

73

Brower, H.L. 1959. Speciation of butterflies of the Papilio glaucus group. I.

Morphological relationships and hybridization. Evolution 13:40-63.

Carson, H.L. 1982. Speciation as a major reorganization of polygenic balances. In:
Mechanisms of speciation, C. Barigozzi, ed. AlanR. Liss, Inc. New York. pp. 411-
433.

Carson, H.L. & A.R. Templeton. 1984. Genetic revolutions in relation to speciation

phenomena: The founding of new populations. Ann. Rev. Ecol. Syst. 15:97-131.
Clarke, C.A. & T.B. Larsen. 1986. Speciation problems in the Papilio machaon
group of butterflies (Lepidoptera: Papilionidae). Syst. Ent. 11:175-181.
Collins, M.M. 1984. Genetics and ecology of a hybrid zone in Hyalophora
(Lepidoptera: Saturniidae). University of Calif. Pub. in Entomology 104:1-93.
Cracraft, J. 1982. Geographic differentiation, cladistics, and vicariance
biogeography: Reconstructing the tempo and mode of evolution. Amer. Zool.
22:411-424.

Dimock, T.E. & R.H.T. Mattoni. 1986. Hidden genetic variation in Agraulis vanillae
incarnata (Nymphalidae). J. Res. Lepid. 25:1-14.

Dornfeld, E.J. 1980. The butterflies of Oregon. Timber Press, Forest Grove,

Oregon. 276 pp.

dos Passos, C.F. & L.P. Grey. 1947. Systematic catalogue of Speyeria (Lepidoptera,
Nymphalidae) with designations of types and fixations of type localities.
American Museum Novitates, No. 1370.

Ehrlich, P.R. 1961. Has the biological species concept outlived its usefulness. Syst.
Zool. 10:167-176.

Ehrlich, P.R. &D.D. Murphy. 1981. The population biology of checkerspot butterflies
(. Euphydryas ). Biol. Zbl. 100: 613-629.

. 1983. Butterflies and biospecies. J. Res. Lepid. 21:219-225.

Ehrlich, P.R. & P.H. Raven. 1969. Differentiation of populations. Science 165:1228-
1232.

Ehrlich, P.R. & R.R. White. 1980. Colorado checkerspot butterflies: isolation,
neutrality, and the biospecies. Am. Nat. 115:328-341.

Eldredge, N. & J. Cracraft. 1980. Phylogenetic patterns and the evolutionary
process. Columbia University Press, New York. 349 pp.

Eldredge, N. 1989. Macroevolutionary dynamics. McGraw-Hill Publishing Co.
New York. 226 pp.

Endler, J.A. 1977. Geographic variation, speciation and dines. Princeton Univ.
Press, Princeton, N.J. 246 pp.

Epling, C. & W. Gatlin. 1950. The relation of taxonomic method to an explanation
of organic evolution. Heredity 4:313-325.

Erwin, T.L. 1981. Taxon pulses, vicariance, and dispersal: an evolutionary synthesis
illustrated by carabid beetles. In: Vicariance biogeography: a critique. G.
Nelson & D.E. Rosen, eds. Columbia University Press, New York. pp. 159-183.
Ferris, C.D. 1983. Speyeria atlantis phenotypes in the southern Rocky Mountains
(Lepidoptera: Nymphalidae: Argynninae). J. Res. Lepid. 22:101-114.

Ferris, C.D. 1987. A revision of the North American SaZZx-feeding Colias species
(Pieridae: Coliadinae). Bull. Allyn Mus. No. 112.

. 1988a. Euphydryas anicia and E. chalcedona in Idaho (Lepidoptera:

Nymphalidae). J. Res. Lepid. 26:109-115.

. 1988b. Revision of the North American Ericaciae-feeding Colias species

(Pieridae: Coliadinae). Bull. Allyn Mus. No. 122.

Ford, E.B. 1945. Butterflies. Collins St. James Place, London. 368 pp.

74

J. Res. Lepid.

Futuyma, D.J. 1979. Evolutionary biology. Sinauer Associates, Inc. Sunderland,
Mass. 565 pp.

Gillham, N. W. 1956. Geographic variation and the subspecies concept in butterflies.
Syst. Zool. 5:110-120.

Gould, S.J. & N. Eldredge. 1977. Punctuated equilibria: The tempo and mode of
evolution reconsidered. Paleobiology 3:115-151.

Gould, S.J. & R.F. Johnston. 1972. Geographic variation. Ann. Rev. Ecol. Syst.
3:457-498.

Grey, L.P. & A.H. Moeck. 1962. Notes on overlapping subspecies I. An example in
Speyeria zerene. J. Lepid. Soc. 16:81-97.

Grey, L.P., A.H. Moeck, & W.H. Evans. 1963. Notes on overlapping subspecies. II.
Segregation in the Speyeria atlantis of the Black Hills. J. Lepid. Soc. 17:129-
147.

Hammond, P.C. 1986. A rebuttal to the Arnold classification of Speyeria callippe
(Nymphalidae) and defense of the subspecies concept. J. Res. Lepid. 24:197-
208.

Johnson, G.B. 1976. Enzyme polymorphism and adaptation in alpine butterflies.
Ann. Missouri Bot. Gard. 63:248-261.

Lorkovic, Z. 1962. The genetics and reproductive isolating mechanisms of the
Pieris napi-hryoniae group. J. Lepid. Soc. 16:5-19, 105-127.

Mattoni, R.H.T. 1989. The Euphilotes battoides complex: recognition of a species
and description of a new subspecies (Lycaenidae). J. Res. Lepid. 27:173-185.
Mayr, E. 1963. Animal species and evolution. Harvard University Press, Cambridge,
Mass. 797 pp.

. 1976. Evolution and the diversity of life. Harvard University Press, Cambridge,

Mass. 721 pp.

. 1982. Processes of speciation in animals. In: Mechanisms of speciation, C.

Barigozzi, ed. Alan R. Liss, Inc. New York. pp. 1-19.

McCorkle, D.V. & P.C. Hammond. 1988. Biology of Speyeria zerene hippolyta
(Nymphalidae) in a marine-modified environment. J. Lepid. Soc. 42:184-195.
Miller, L.D.&F.M. Brown. 198 1 . A catalogue checklist of the butterflies of America
north of Mexico. The Lepidopterists’ Society Memoir No. 2. 280 pp.

Moeck, A.H. 1957. Geographic variability in Speyeria. Milwaukee Entomol. Soc.
Spec. Pap. 48 pp.

Murphy, D.D. 1983. On the status of Euphydryas editha baroni with a range
extension of E. editha luestherae. J. Res. Lepid. 21:194-198.

Pimentel. R.A. 1959. Mendelian infraspecific divergence levels and their analysis.
Syst. Zool. 8:139-159.

Platt, A.P. 1983. Evolution of North American admiral butterflies. Bull. Entomol.
Soc. Am. 29:10-22.

Remington, C.L. 1968. Suture zones of hybrid interaction between recently joined
biotas. Evolutionary Biology 2:321-428.

Rightmire, G.P. 1985. The tempo of change in the evolution of mid-Pleistocene
Homo. In: Ancestors: the hard evidence, E. Delson, ed. Alan R. Liss, Inc. New
York. pp. 255-264.

Rindge, F.H. 1987. Speyeria collection of Paul Grey to the American Museum of
Natural History. J. Lepid. Soc. 41:123.

Robinson, R. 1971. Lepidoptera genetics. Pergamon Press, Oxford. 687 pp.
Rockey, S.J., J.H. Hainze, & J.M. Scriber. 1987. A latitudinal and obligatory
diapause response in three subspecies of the eastern tiger swallowtail Papilio
glaucus (Lepidoptera: Papilionidae). Am. Midi. Nat. 118:162-168.

29(l-2):54-76, 1990(91)

75

Rosen, D.E. 1978. Vicariant patterns and historical explanations in biogeography.
Syst. Zool. 27:159-188.

Rothschild, W. & K. Jordan. 1906. A revision of the American Papilios. Novit. Zool.
13:411-752.

Scott, J.A. 1980. A survey of valvae of Euphydryas chalcedona, E. c. colon, and E.
c. anicia. J. Res. Lepid. 17:245-252.

— — . 1988. Speyeria atlantis in Colorado: rearing studies concerning the relation
between silvered and unsilvered forms. J. Lepid. Soc. 42:1-13.

Scriber, J.M. 1986. Origins of the regional feeding abilities in the tiger swallowtail
butterfly: ecological monophagy and the Papilio glaucus australis subspecies in
Florida. Oecologia 71:94-103.

Sperling, F.A.H. 1987. Evolution of the Papilio machaon species group in western
Canada (Lepidoptera: Papilionidae). Quaest. Ent. 23:198-315.

Stanley, S.M. 1979. Macroevolution: pattern and process. W.H. Freeman & Co. ,
San Francisco. 332 pp.

Stebbins, G.L. & F.J. Ayala. 1981. Is a new evolutionary synthesis necessary?
Science 213:967-971.

Thorpe, J.P. 1984. Primary and secondary transition zones in speciation and
population differentiation: a phylogenetic analysis of range expansion. Evolution
38:233-243.

Wells, P.V. & J.D. Stewart. 1987. Cordilleran-boreal taiga and fauna on the
central Great Plains of North America, 14,000- 18,000 years ago. Am. Midi. Nat.
118:94-106.

White, M.J.D. 1978. Modes of speciation. W.H. Freeman & Co., San Francisco. 455

pp.

Wilson, E.O. & W.L. Brown, Jr. 1953. The subspecies concept and its taxonomic
application. Syst. Zool. 2:97-111.

76

J. Res. Lepid.

Table 1 . Derived character state changes in the phylogeny of Speyeria callippe

subspecies

1 . black median band at end of discal cell of dorsal hindwing distally elongate
(versus not elongate)

2. silver spot in discal cell of ventral hindwing (vhw) large (vs. small discal spot)

3. yellow-orange ground color on dorsal wing (dw) (vs. medium orange)

4. very heavy melanic basal suffusion on dw (vs. heavy suffusion)

5. disc on vhw yellow-brown or pure yellow, red ground color on ventral
forewing (vfw) reduced (vs. brown disc & heavy red color)

6. median spots on vhw large and elongate (vs. short, pointed spots)

7. small median spots on vhw (vs. larger spots)

8. wide yellow submarginal band on vhw (vs. narrow band)

9. disc on vhw yellow-brown or pure yellow (vs. brown disc)

10. melanic basal suffusion on dw greatly reduced (vs. heavy suffusion)

1 1 . forewings rounded (vs. pointed'forewings)

12. reduced melanic basal suffusion on dw (vs. heavy suffusion)

13. large and rounded median spots on vhw (vs. pointed spots)

14. spots on vhw unsilvered (vs. silver spots)

15. reduced melanic basal suffusion on dw (vs. heavy suffusion)

16. ground color of vfw yellow in males (vs. orange color)

1 7. veins in dorsal forewing of males thin with reduced dark scaling (vs. thick
dark veins)

1 8. yellow-orange ground colot on dw (vs. medium orange)

19. disc greenish brown on vhw (vs. brown disc)

20. median spots on vhw large and elongate (vs. rounded spots)

21 . melanic basal suffusion on dw greatly reduced (vs. heavy suffusion)

22. disc pure green without brown on vhw (vs. greenish brown disc)

23. green suffusion over yellow submarginal band on vhw (vs. clear yellow
band)

24. disc gray or gray-green (vs. pure green disc)

25. disc olive-green (vs. pure green disc)

26. heavy melanic basal suffusion on dw (vs. little suffusion)

27. large wind size (forewing length over 32 mm.) (vs. small wing size less than
32 mm.)

28. long valve process on male genitalia (vs. short, club-shaped process)

Journal of Research on the Lepidoptera

29(l-2):77-84, 1990(91)

Cryptic larval polychroinatism in Rekoa marius
Lucas and R. pa /eg on Cramer (Lycaenidae:
Theclinae).

Ricardo Ferreira Monteiro

Universidade Federal do Rio de Janeiro, Institute de Biologia, Deptartamento de Ecologia,
Lab. de Ecologia de Insetos, Caixa Postal 68020, CEP 21941 - Rio de Janeiro, BRASIL

Abstract, Rekoa marius and R. palegon , are widely distributed
neotropical hairs treaks that show cryptic larval polychromatism which
is food dependent. Larvae of R. marius showed a short-term color shift
when reared from different colored flower buds. A great number of host
plant species belonging to several families were recorded for both R.
marius and R. palegon larvae. Biological aspects of the larvae are
presented and the relationships between cryptic polychromatism and
polyphagy are discussed.

Introduction

The neotropical hairstreak species Rekoa marius Lucas and!?, palegon
Cramer (Lycaenidae: Theclinae) are widely ranged from southern Texas
(USA) to Argentina, and are known in butterfly literature as Theda
ericusa and T. palegon (Robbins, 1991).

The genus Rekoa has recently been revised by Robbins (1991) who
refers to Rekoa marius and R. palegon larvae as polyphagous species,
citing records mainly on Leguminosae and Malpighiaceae for the former
and Asteraceae for the latter.

Apparently crypsis is very common among lycaenid larvae (e.g.,
Maschwitz et al., 1984, 1985; Ballmer and Pratt, 1989), but records of
larval cryptic polychromatism are poorly documented. For example,
food-determined color pattern was reported in Callophrys mossii hayensis
larvae (Brown, 1969), but was disputed by Emmel and Ferris (1972), and
Orsak and Whitman (1988).

Malicky (1970) described Newcomer organs and perforated cupolas in
R. palegon larvae, and most Theclinae are known to be ant associated
(Hinton, 1951;Atsatt, 1981; Henning, 1983; Cottrell, 1984). However, no
ants have ever been recorded for the two species studied here.

The purpose of this paper is to show the cryptic coloration in two Rekoa
larvae species and discuss their food-dependent color mechanism in
relation to polyphagy.

Materials and Methods

From June 1985 to June 1989 1 examined flower buds of about a hundred plant
species in Barra de Marica (22° 57'S, 42° 52'W) State of Rio de Janeiro, Brasil,
consisting of beach sand dune vegetation (“restinga”), and covering an area
approximately 200 ha.

78

J. Res. Lepid.

Immatures of Rekoa species were collected, and larvae were reared in transpar-
ent plastic boxes, with moist paper on the bottom. Adult voucher specimens were
deposited in Museu Nacional do Rio de Janeiro (Rio de Janeiro, Brasil) and in the
National Museum of Natural History (Washington, DC).

To verify the effect of host plant on larval color, I collected 30 larvae of R. marius
in the first and second instar, and five larvae in the third and fourth instar and
reared them on different host species with four different flower bud colors. Five
second instar larvae of R. palegon were also submitted to the same experiment.
All larvae were observed daily.

The presence of attendant ants and the occurrence of parasitoids were also
recorded.

Results

LIFE HISTORY AND CRYPTIC COLORATION

Immatures of R. marius are most common in summer and fall, but eggs
and larvae may be found during most of the year. Larval development
takes 22 days with four instars and pupal duration of 10 days (26.5°C ±
2; 65% ± 10 RH; N=10).

R. marius was found feeding on 30 host species of 26 genera belonging
to 10 dicot families (Table 1). The main host plant species had Extra
Floral Nectaries (EFN) on flower buds (. Arrabidaea conjugata andLundia
cordata ), leaves 0 Senna bicapsularis and S. australis ) or bracts ( Ouratea
cuspidata ). In these plant species ants were observed frequently visiting
their EFN.

After eclosion larvae are yellowish and feed on small flower buds; as
they grow they feed on larger flower buds and become cryptic. Although
several host species were used, R. marius larvae always resemble the
color of the flower bud on which they are fed (Fig. 1, A— F). Hence, in some
host plants, such as O. cuspidata, S. bicapsularis, Coccoloba arborescens
and Lundia cordata, young larvae of R. marius are green or whitish-
yellow, corresponding to the color of the calyx of these plant species. The
third and fourth instar larvae are predominantly yellow on the first two
plant species, white on C. arborescens and red on L. cordata. Both young
and mature larvae feeding on Arrabidaea conjugata were always lilac
colored. This plant species, in contrast to the others, has a lilac calyx and
corolla. The adults show no color variation correlating to larval host
plant species.

R. palegon occurred in the study area mainly in fall and winter, when
most of its host plant species, almost all Asteraceae, (Table 1) bloom.
Larval development takes 23 days and pupa duration is 10 days (N=10).
The larvae are yellowish when they hatch. During the three following
instars they become predominantly green, white, yellow or wine (Fig. 1,
G-L). Larvae from inflorescences of Mikania stipulacea present two
cryptic coloration patterns, wine-green when feeding on younger flower
heads or green- white on mature flower heads.

29(l-2):Y7-84? 1990(91)

79

Table 1 . Host records of Rekoa marius and R. paiegon larvae collected since
1385, from “restinga” of Barra de Marica, Marica, Brazil.

Rekoa marius

Rekoa paiegon

Apocynaceae

Aspidosperma pyricollum

Bignoniaceae

Arrabidaea agnus-castus

A. conjugata

Lundia cordata

Adenocaiymma marginatum
Anemopaegma venusta

Jacaranda jasminoides

Tabebuia chrystoricha

Boraginaceae

Cordia verbenaceae

Leguminosae

Senna bicapsuiaris

S. australis

Caesalpinia bonduc

CratyHa hypargyrea

Andira iegaiis

Diocleavioiacea

Swartzia apeiata

Malpighiaceae

Byrsonima sericea

Peixotoa hispidula

Stygmaphylion paralias

Heteropteris chrysophylia
Mefastomataceae

Marcetia taxifolia

Tibouchina aff. holosericea

Myrtaceae

Eugenia uniflora

E. ovalifolia

Neomitranthes obscura

Ochnaceae

Ouratea cuspidate

Polygonaceae

Coccoloba arborescens

C. alnifoiia

Sapindaceae

Paullinia weinmanniaefolia

Asteraceae

Mikania hoehnei

M. stipulacea

M. micrantha

Eupatorium laxum

E. odoratum

Vernonia scorpioides

V. geminata

Wedelia paludosa

Wulfia baccata

Baccharis punctulata

Trixis antimenorrhoea

Polygonaceae

Coccoloba arborescens

FEEDING BEHAVIOR

Larvae of both species feed all day. They bore into the flower bud (or
inflorescence) making a circular hole in which they insert their retractile
heads and eat the reproductive tissue and part of the calyx and corolla.

80

J. Res. Lepid.

Each R. marius larva ate about 40 to 50 flower buds of O. cuspidata or
S. bicapsularis during its development. Food deprivation led to cannibal-
ism; larger larvae generally ate smaller ones or pupae.

In the laboratory, R. palegon larvae reared on S. bicapsularis and O.
cuspidata (. R . marius host species) developed into adults in 90% of the
cases (N=20). On the other hand 95% of R. marius larvae (N=20) failed
to develop to pupae when reared in the three main host species used by
R. palegon : Mikania hoehnei , M. stipulacea and Vernonia scorpioides.

NATURE OF COLORATION

Food plant shift experiments showed that first and second instar larvae
change their coloration and become similar to the new host flower bud
color. Twenty green or yellow first instar larvae of R. marius collected
from S. bicapsularis , S. australis and O. cuspidata became lilac when
reared on A. conjugata flower buds. Ten first and second instar lilac
larvae from A. conjugata became yellowish after being reared with the
three above host species.

Five green second instar larvae of R. palegon became yellowish when
changed from Eupatorium laxum (green flower head) to Wedelia paludosa
inflorescences (yellow flower head).

Coloration change occurs within three to five days after the food plant
shift, but coloration becomes more pronounced after the first moult.
Third and fourth instar larvae transferred to different colored host
plants partially change their color and eventually showed intermediate
colors.

PARASITISM AND MYRMECOPHILY

Rekoa marius and R. palegon larvae are hosts for the same four
parasitoids: Telenomus sp. (Hym.: Proctotrupoidea: Scelionidae — egg),
Rogas sp. (Hym.: Braconidae larva), Conura n. sp. (Hym. : Chalcididae
— larva/pupa) and a tachinid species (Diptera — larva/pupa). Total

Figure 1 . Rekoa marius (A-F) and R. palegon (G-L) color patterns on some food
plants.

(A) Ouratea cuspidata (young flower bud)

(B) O. cuspidata (mature flower bud)

(C) Coccoloba arborescens

(D) Caesalpinia bonduc

(E) Arrabidaea conjugata

(F) Lundia cordata

(G) Eupatorium laxum

(H) Mikania hoehnei

(I) M. stipulacea

(J) Vernonia scorpioides

(K) Wulfia baccata

(L) Wedelia paludosa

82

J. Res. Lepid.

parasitism of each butterfly species in both years (1987-1988) was no
more than 15% (N=300) for eggs and 5% (N=500) for larvae.

Several ant species were observed attending third and fourth instar
larvae of R. marius (Fig.l: B,C,E) ( Camponotus crassus , C. rufipes, C.
cingulatus, Solenopsis sp., and Conomyrma sp.) and R. palegon
0 Camponotus crassus, Crematogaster sp. and a dolichoderine species).
The ants did antennation on the Newcomer gland and pore cupola but fed
only from the former.

Discussion

Color and feeding behavior of R. marius and R. palegon larvae provide
them with camouflage (Fig. 1). Such cryptic host plant dependent
polychromatism is caused by flavonoid and carotenoid pigments accumu-
lated by larvae (Kaplan & Monteiro, unpublished data). Wilson (1987)
showed that flavonoid content of several lycaenid species is dependent on
the flavonoid content of larval food, and although she suggested that
flavonoid pigments act as chemical defenses, I suspect that they confer
crypsis rather than aposematism in Rekoa larvae.

Brower (1958) speculated on the selective advantage of polyphagy in
cryptic insects, suggesting that populations using several host plant
species would be more difficult for insectivorous birds to find. Color
polymorphism would break the insect population into several “visual
species” each of which would have to be learned independently by avian
predators. In both Rekoa species, food plant induced coloration produced
cryptic poly chromatism, dividing the population into several “visual
species.” Such a mechanism may be of great importance because there
would be no non-cryptic individuals. Hence, polyphagy would not be
dependent on the number of cryptic color morphs, as in polymorphism,
but quite the contrary, the number of morphs would be determined by the
host plant coloration patterns. Polyphagy in these hairstreaks may also
allow them to reproduce for longer periods in the year, overcoming the
problems of ephemeral food resources in nature.

The wide range of host plants used by many thecline species, consid-
ered uncommon among other butterflies, may be due in part to their
flower-feeding habit (Robbins & Aiello, 1982). However, I believe that
the color determination mechanism found in Rekoa may also play an
important role on host range width providing larval cryptic protection for
whichever host plant used. Such a mechanism indicates that a perfect
camouflage does not necessarily require a monophagous habit of larva,
as has been suggested by Maschwitz et al. (1984, 1985).

Although for flower bud feeding and ant associated lycaenids the
presence of Extra Floral Nectaries (EFN) in their host plants may be less
important than in foliage feeding species, it is remarkable that the five
main food plants of R. marius (S. hicapsularis, S. australis , L. cordata,
A. conjugata and O. cuspidata ) bear EFN. As these larvae do not feed on
EFN, in contrast to what was observed in a riodinid species by De Vries

29(l-2):77-84, 1990(91)

83

and Baker (1989), it is possible that this host species preference is based
on a direct correlation between a higher frequency of ants in the host
plants and the rate of larvae survival as verified in other lycaeriid species
(Atsatt, 1981; Henning, 1983; Pierce & Elgar, 1985).

Ant protection against parasitoids recorded for some lycaenids (Pierce
& Mead, 1981; Atsatt, 1981; Pierce & Easteal, 1986) may also occur in
Rekoa as indicated by the low parasitism rate found in both myrmecophi-
lous species.

Acknowledgements. Margarete V. Macedo helped me in all phases of this work.
Rogerio P. Martins, Curtis Callaghan, Donald Harvey and Robert K. Robbins and
two anonymous reviewers gave valuable critical comments on the manuscript.
Robert Robbins, Marta Loyacono, Paul Marsh, Carlos Roberto Brandao and G.
Delvare provided the insect identifications. Janie Silva identified the plants. The
field work was financially supported by FINEP. Doctoral scholarship came from
CAPES.

Literature Cited

Atsatt, P. R. 1981. Ant-dependent food plant selection by the mistletoe butterfly
Ogyris amaryllis (Lycaenidae). Oecologia 48: 60-63.

Ballmer, G. R. and G. F. Pratt. 1989. A survey of the last instar larvae of the
Lycaenidae (Lepidoptera) of California. J. Res. Lepid. 27: 1-81.

Brower, L. P. 1958. Bird predation and food plant specificity in closely related
procryptic insect. Amer. Natur. 92 (864): 183-187.

Brown, R. M. 1969. Larva and habitat of Callophrys fotis bayensis (Lycaenidae).
J. Res. Lepid. 8: 49-50.

Callaghan, C. J. 1981. Notes on the immature biology of two myrmecophilous
Lycaenidae: Juditha molpe (Riodininae) and Panthiades bitias (Lyceninae). J.
Res. Lepid. 20 (1): 36-42.

Cottrell, C. B. 1984. Aphytophagy in butterflies: its relationship to myrmecophily.
Zool. J. Lin. Soc. 79: 1-57.

DeVries, P. J. and I. Baker. 1989. Butterfly exploitation of ant-plant mutualism:

adding insult to herbivory. J. New York Ent. Soc. 97 (3): 332-340.

Emmel, J. F. and C. D. Ferris. 1972. The biology of Callophrys ( Incisalia ) fotis
bayensis (Lycaenidae). J. Lepid. Soc. 26: 237-244.

Henning, S. F. 1983. Biological groups within the Lycaenidae (Lepidoptera). J.
Ent. Soc. So. Africa 46 (1): 65-85.

Hinton, H. E. 1951. Myrmecophilous Lycaenidae and other Lepidoptera - a
summary. Proc. So. London Ent. Nat. Hist. Soc. 1949-1959, 111-175.
Malicky, H. 1970. New aspects on the association between lycaenid larvae
(Lycaenidae) and ants (Formicidae, Hymenoptera). J. Lepid. Soc. 24: 190-202.
Maschwitz, U., M. Schroth, H. Hanel and T. Y. Pong. 1984. Lycaenids parasitizing
symbiotic plant-ant partnerships. Oecologia 64: 78-80.

Maschwitz, U., M. Schroth, H. Hanel and Y. P. Tho. 1985. Aspects of the larvae
biology of myrmecophylous lycaenids from West Malaysia (Lepidoptera). Nach.
Ent. Ver. Apollo, Frankfurt 6: 181-200.

Orsak, L. and D. W. Whitman. 1986. Chromatic polymorphism in Callophrys mossii
bayensis larvae (Lycaenidae): spectral characterization, short-term color shift,
and natural morph frequencies. J. Res. Lepid. 25: 188-201.

84

J. Res. Lepid.

Pierce, N. E. and S. Easteal. 1986. The selective advantage of attendant ants for
the larvae of lycaenid butterfly, Glaucopsyche lygdamus. J. An. Ecol. 55: 451-
462.

Pierce, N.E. and M. A. Elgar. 1985. The influence of ants on host plant selection
by Jalmenus evagoras, a mymecophilous lycaenid butterfly. Behav. Ecol. and
Sociob. 16: 209-222.

Pierce, N. E. and P. S. Mead. 1981. Parasitoids as selective agents in the symbiosis
between lycaenid butterfly larvae and ants. Science 211 (132): 1185-1187.

Robbins, R. K.‘ 1991. Evolution, comparative morphology, and identification of the
Eumaeine butterfly genus Rekoa Kaye (Lycaenidae: Theclinae). Smithsonian
Contributions to Zoology 498: 64.

Robbins, R. K. and A. Aiello. 1982. Foodplant and oviposition for Panamanian
Lycaenidae and Riodinidae. J. Lepid. Soc. 36 (2): 65-75.

Wilson, A. 1987. Flavonoid pigments in chalkhill blue (Ly sandra coridon Poda)
and other lycaenid butterflies. J. Chem. Ecol. 13 (3): 473-493.

Journal of Research on the Lepidoptera

29(l-2):85-91 , 1990(91)

On Pieris ( Artogeia ) marginalis macdunnoughii
Remington (Pieridae). Part II.

S. R. Bowden

Lydeard, Merryfield Way, Storrington, W. Sussex, RH20 4NS, U.K.

Abstract. The crossing of Pieris ssp. macdunnoughii (almost without
marking on the upperside) with a European P. napi produced F x male
hybrids even more heavily marked than normal napi. It was not
thought that the funebris gene which was present in the European
stock was responsible. The female hybrids, carrying an approximation
to the napi female pattern, showed slight asymmetry of marking, and
“smudging”. Comparable results from other wide crosses within the
napi group suggest that these, unlike closer hybrids, often suffer
disturbance of the canalizing mechanisms which control patterns. In
the present case abnormally short diapause failed to produce the full
“spring” characters.

Introduction

A recent paper (Bowden 1988) described the hybridization of Pieris
{napi) macdunnoughii Remington with European P. napi L., and
concluded that that the Colorado taxon (like ssp. marginalis Scudder of
Oregon) carried two genetically distinct systems of yellow coloration.
However, no adequate description of the melanic markings of the
hybrids was given.

The present publication is intended to supply this, and to discuss the
realized expression of the upperside pattern of this species-group,
though further experimental work with unrelated macdunnoughii
material, including European crosses in both directions, is desirable to
confirm conclusions, and particularly to quantify environmental
effects.

It should be explained that the work now reported was initiated as an
attempt to transfer the funebris (Lorkovic 1971) gene, as homozygote,
to substantially pure ssp. macdunnoughii. There it might produce an
informatively different phenotype. But the series of pairings (starting
with an inbreeding of brood 1986-y — see below) necessary for this was
not obtained, and in the present connection the presence of the funebris
gene was probably irrelevant. Any future worker on macdunnoughii x
napi hybridization will avoid this complication, as the funebris gene
appears, most unfortunately, to have been lost everywhere.

Several American subspecies related to Pieris napi carry no more
than faint indications of the spot markings which normally charac-
terize the Palaearctic members of the group (fig. 1). Such indications as
there are (e.g. in fig. 6) seem to suggest that the pattern is present, but

86

J. Res. Lepid .

1. Corsican J'64dvm 6.

2. Corsican cT'64</v,59 (underside).

3. Calabrian X Irish cT'67 / 6a.

4. Calabrian X Irish }'67/9a.

5. F2 neobryoniae (ex Karnten) X adalwinda (ex Kiruna) cf'54p33.

6. virginiensis X oieracea 2'65qr" 19.

1. Nff.Q cf'85/fl .

8. Q.NffQ 9 '86/ 2.

9. Q.Nff $'80n38.

10. Nff.NffQ 9'86c15.

29(l-2):85-91, 1990(91)

87

has been suppressed. Breeding experiments do reveal the presence in
some Nearctic populations of a recessive epistatic gene, restricta, which
allows apparent dominance of the European marking over its American
absence: thus the F1 hybrids in general follow the European model.
This happens, indeed, in the case of macdunnoughii.

Summary of Brood Relationships

See also Bowden, 1988.

In the following table, N stands for European napi (with or without
the funebris gene) and Q stands for macdunnoughii. Hybrids are
indicated by juxtaposition of these symbols, the female parent being
placed first. Lower-case italics identify separate broods.

Q

2'84R10 x c?'84R7

— »

1985-r,

14 cf + 17 $

$'85r21 x cf'85r22

->

1985-tf,

12 C? + 11 2

NQ

2'84p27ffx c?'85r6

-—>

1985-/?,

95 Cf + 0 2

2'86A‘’10ff x cf'85QlO

1986-^,

94 cf + 0 2

2'86/i"'llff x c?Cf '85(714-16

->

1986-g“,

10 Cf +?2 2

N.NQ

?2'85/i23,24 x cfcf ’85667,68

1986-c,

41 Cf + 36 2

Q.NQ

2'85r31 x cfcf'85&78,85

->

1986-y,

5 cf + 3 2

QN

$'85^18 x cfcf'85/i88-91

— >

1986-ra,

38 cf + 26 2

Here the symbol ff following an individual identification number
indicates a funebris homozygote.

Marking Emphasis in Hybrid Males

In the non-diapause emergence of our Fx hybrid brood 1985-/? (fig. 7),
which gave no adult females, the black upperside markings of the 39
males were generally intense, with the fore wing apical patch extend-
ing rather smoothly to vein Ml (= v3 of the Herrich- Schaffer system as
used by Higgins & Riley 1970 and Higgins 1975.) So great an extension
is usual in Pieris brassicae L., but seldom attained in P. rapae L. and P.
napi. Further, the veins of the hindwings were marked distally with
spots, which spread laterally from the veins to produce a marginal
band only slightly interrupted (fig. 7). However, less extensively
marked males were eight in number.

At first sight the almost continuous band recalled the fact that the
European mother of 1985-/? was of funebris form. However, although
all the individuals of k must have been heterozygous for the (normally
recessive) funebris gene, a small proportion (eight as mentioned)
showed little or no vein-end expansion. Also, in the similarly con-
stituted later broods 1986-^, gl\ the hindwing distal vein-marking
was less. Such variation, attributable to environmental fluctuation,
interferes with quantitative assessment.

Besides, examination of many earlier napi f. funebris specimens,
especially those of less extreme form, had shown that the marginal
band of funebris appears to be built up primarily from interneural

88

J. Res . Lepid

pigmentation, though the veins themselves may also become black. It
is known that, in Pieris ( Artogeia ), patterns are assembled from sys-
tems based on either the veins or interneural (transverse) bars, and
that these systems are separately controlled, whether genetically or by
environment. The marking in 1985-/2 is always vein-based, and is in fact
close to that of P. canidia Sparrman and even of P. deota deNiceville. A
connection with funebris can probably therefore be excluded.

Markings of Female Hybrids

Females as well as males were obtained when the Fj males were
back-crossed in either direction — as also when, later, the reverse Fi
(J macdunnoughii x cf napi funebris heterozygote) succeeded (1986-ft).

In the back-cross (1986-j) to macdunnoughii only about half the off-
spring should have lacked homozygous restricta; accordingly four out
of eight expressed European marking. But then as well as the hind-
wing vein-marking there was heavy black napi patterning on the fore-
wings of two females, locally smudged on to small areas usually white
(e.g. fig. 8). Such smudging, when it occurred in Lorkovic’s earliest
funebris broods, was attributed by him to partial expression of heter-
ozygous funebris (which is normally quite recessive), but I have
obtained it (for example) in Calabrian x Irish napi (e.g. fig. 4). It may
be that it can occur as some consequence of very wide crosses. Another
disturbance in these back-cross females was a certain imbalance: left-
and right-side markings were not perfect mirror-images (fig. 8). This
feature too may be evidence of an instability (basically genetic) when
widely distinct genomes are combined.

In the 1986-c back-cross to napi (funebris heterozygote) about one-
quarter of the offspring (17 or 17 out of 77) were of regular f. funebris,
as expected, the phenotypes being within the range generally found
(not ilustrated here — cf. Bowden 1983). The female Jcl5 (fig. 10)
looked rather questionable as a possible partial funebris: though its
underside showed no funebris characters whatever, this is not con-
clusive negative evidence. There was little sign of bilateral imbalance
in 1986-c; most of the non-funebris males looked not unlike the Fx
hybrid males.

In the reciprocal Fx cross, 1986-n (which had relatively few males
marked more extensively than napi at apices and hindwing vein-ends),
the female veins were marked distally, but not very heavily. Some (e.g.
Jn38, fig. 9) showed once more a left-right asymmetry of marking. The
writer was rather surprised not to find in the females even larger
melanic areas, presaged by those unusually extensive male markings
in 1985-/2. But what conclusions are to be drawn from a comparison of
figs. 4 and 10?

29(l-2):85-91, 1990(91)

89

Pattern Interpretation

There was considerable individual variation within each brood of our
hybrids, and a significant diminution of average male patterning be-
tween brood 1985-& and its successors 1986-g*, gl\ though parentage of
these was similar (only for 1986-72 and c was the “European” parent
merely heterozygous for funebris). As is usual in such cases, it is un-
certain exactly how much of this variation was due to cryptic genetic
differences, and how much more to environmental fluctuation. But
except in 1986-c, where funebris became homozygous in some in-
dividuals, it is doubtful whether this gene had any visible effect in
these macdunnoughii hybrids.

Nevertheless, the use of funebris-form napi introduced enough
possible genetic complications to increase the uncertainty of inter-
preting the pattern of the hybrids. Funebris itself is far from being
understood.

However, we still have much to learn even about the typical Pieris
(. Artogeia ) design, which is in fact a very peculiar one. Schwanwitsch
(1956) claimed that in Pieridae in general the principal wing-pattern
components that he recognized as typical for other butterflies were
present, but he acknowledged, “Most of the components disappear from
the [genus] Pieris wing-pattern”. There are no eye-spots, no ocelli, no
ripple-patterns, no central symmetry system. Even the true discoidal
spots (which Nijhout 1978 remarks are “present in virtually every
species of Lepidoptera”) do not appear, whereas they are prominent in
Euchloe , Pontia and Tatochila.

Close hybridization within the sub-genus Artogeia usually produces
no significant pattern disturbance: the forms that result are simply
intermediate, as far as dominant and epistatic relationships allow.
More distant hybrids seem to have their specific canalizing mech-
anisms disoriented, and this can be exaggerated in secondary and
back-cross combinations. The pattern of the Arctic/Alpine adalwinda
neohryoniae F2 hybrid (fig. 5) is startling in a male, but of course on the
underside strong radial markings persist normally in this sex too. It
appears that extreme imbalance of modifiers may be responsible for
many apparent reversions.

I have previously (1983) quoted Riedl (1978) to support the view that
hybridization can allow the re-activation of suppressed gene-systems.
It is not intended even to consider here the two questions as to how the
basic “ Artogeia ” pattern of these closely related butterflies arose and
how its variants are now (for the most part) maintained. The historical
question (at present almost beyond conjecture) is of the greater interest.
Are ecological/selection proposals convincing?

90

J. Res. Lepid.

Pattern and Diapause

Other disturbances are to be expected in wide hybrids, and were in
fact encountered in the case of napi x macdunnoughii. The particular-
ly common failure of the female sex to reach maturity, at least in one
direction of the cross (cf. Lorkovic 1978) has been detailed for these
hybrids in our earlier report (Bowden 1988). One amendment to that is
due: a second female from 1986-g“ eclosed (after diapause) subsequent-
ly to the preparation of the report and was classified as a true hybrid,
not a waif. So such failure may not be absolute, and I believe that other
workers have used injections of ecdysone to induce hybrid females
among Papilionidae to complete their imaginal development.

The determination and onset of diapause may also become irregular,
or its connection with seasonal forms be distorted.

In brood 1985-&, 39 normal non-diapause emergences occurred 25 ix-85
to Tx-85, then ceased, as if at initiation of diapause, though remaining
pupae were not cold-stored until 4 xii*85. However, pupae subsequently
restored to room temperature on 12 i-86 and on 27-iv-86 produced
12 + 15 male adults after about 15 days, but still of the non-diapause
upperside marking! These “summer-form” insects of February and
May 1986 did nevertheless approach the post-diapause underside pat-
tern, with rather complete hindwing veining. On the other hand,
pupae kept in the refrigerator until 25-vi-86 and 26-viii-86 produced
12 + 14 “spring-form” males in about eight and five days respectively.
These “spring” insects showed only a very slight blackening of upper-
side hindwing vein-ends.

I believe that the production of “non-diapause” forms after a short but
apparently true diapause has perhaps not previously been reported; it
deserves study, with stricter environmental control. But all the irre-
gularities of development described above do accord with the author’s
1988 conclusion, that macdunnoughii Remington is appropriately
separated from the species napi L.

Acknowledgements. I am very grateful to C. L. Remington, W. B. Watt, Frances
Chew and A. M. Shapiro for sending me material of Colorado macdunnoughii
over a period of 30 years, and also for the background information which they
supplied.

Nearly all the specimens obtained in the course of this work are, like others
bred by the writer, to be deposited in the British Museum (Natural History),
South Kensington, London.

Literature Cited

BOWDEN, S. R., 1983. A palaeomorph of ArtogeiaP. — f. funebris Lorkovic. Proc.

Brit. ent. nat. Hist. Soc. 16:76-80.

1988. On Pieris ( Artogeia ) marginalis macdunnoughii Remington. J. Res.

Lepid. 26:82-88.

HIGGINS, L. G., 1975. The Classification of European Butterflies. London.

29(l-2):85-91, 1990(91)

91

HIGGINS, L. G. & N. D. RILEY, 1970. A Field Guide to the Butterflies of Britain and
Europe. London.

LORKOVIC,Z., 1971. Pieris napi (L.) morfa funebris, osebujna nova rekombinaeija
krizanja. Acta entomologica Jugoslavia 7:1-9.

1978, Types of hybrid sterility in diurnal Lepidoptera: speciation and

taxonomy. Acta entomologica Jugoslavia 14:13-26.

RIEDL, R., 1978. Order in Living Organisms: a Systems Analyais of Evolution
Chichester.

SCHWANWITSCH, B. N., 1956. Wing-pattern of Pierid butterflies. Revue d’Entorn
ologie de 1TJRSS 35(2):285-301.

Journal of Research on the Lepidoptera

29(l-2):92-104, 1990(91)

An Annotated List of Lepidopterological Journals

Gerardo Lamas

Museo de Historia Natural, Universidad Nacional Mayor de San Marcos, Apartado 14-0434,
Lima- 14, Peru.

Abstract. One hundred and seventy-five periodical publications,
dealing exclusively or almost exclusively with lepidopterological mat-
ters, are listed herein. Each entry includes data on the title(s) of the
serial, place(s) and date(s) of publication, and issues examined by the
author.

During preparation of a Bibliography on the Neotropical Butterflies
(Lamas, Robbins & Field, in press), a large number of periodicals were
searched. Naturally, special efforts were made to consult runs as
complete as possible of those journals dealing mostly with Lepidoptera.
As a consequence of that bibliographic research, I became interested in
learning how many lepidopterological journals were in existence in the
world. The results were quite unexpected, as 175 different periodicals
have been located, over half of which (94) still are being published. Only
periodical publications treating Lepidoptera exclusively or almost exclu-
sively have been considered.

Each entry in this list is arranged alphabetically by journal title, and
includes data on any title changes, place(s) and date(s) of publication,
and a list of the issues actually examined by the author (indicated by
“Rev:”). In several cases, it has proven impossible to locate complete runs
of a periodical, in spite of having had access to the superb holdings of the
British Museum (Natural History), Smithsonian Institution, National
Agricultural Library (Maryland) and Library of Congress (Washington,
DC), besides many other smaller libraries. Particularly frustrating have
been some of the ephemeral serials of restricted circulation, including
many Japanese journals and, especially, periodicals treating matters
related to silk moths and sericulture. Unfortunately, natural history
libraries commonly do not have large holdings of the latter, and for that
reason I have been unable to consult most of such titles listed herein
which have been extracted from secondary sources.

As with other bibliographic publications, the present one certainly will
prove incomplete; I trust readers will be kind enough to point out to me
any inaccuracies and omissions that I have incurred. Information for the
present list was gathered prior to May 31, 1991.

Acknowledgments. My heartiest thanks go to all librarians who have helped me
throughout the years in gathering bibliographic information and, especially, to my
friends Mayumi Takahashi and Scott E. Miller, who commented on an earlier
manuscript list and provided several additions to it. Ulf Eitschberger sent
important data on several European journals, and John T. Polhemus a most useful
list of Japanese periodicals.

29(1 -2):92- 104, 1990(91)

93

1. ACTA RHOPALOCEROLOGICA. Gengen-Sha Company. Fukuoka.

[1977=1982] Rev: 1977(1,2), 1979( , ,3).

2. ACTA SERICOLOGXCA. Sericultural Experiment Station. Norinsho.

[1952— » ] Rev: None.

3. ALEXANOR. Revue des Lepidopteristes Frangais. Paris.

[1959— > ] Rev: 1-16.

4. L AMATEUR DE PAPILLONS. Journal de Lepidopterologie. Paris.

[1922-1938] Rev: 1-8, 9(1-4) (all published).

Continued as: Revue francaise de Lepidopterologie (1938-1958).
ANNALES DU LABORATOIRE D’ETUDES DE LA SOIE. Lyon.

See Rapport du Laboratoire d’Etudes de la Soie. 5.

5. ANTHOCARIS. Matsumoto Konchu Danwakai. Nagano.

[1954-1959] Rev: None.

6. APATURA. Re vista de Lepidopterologla de la Sociedad Asturica de

Lepidopterologia. Oviedo.

[1979— > ] Rev: 1.

7. APOLLO. Bulletin of the Lepidoptera Specialist Group. International

Union for the Conservation of Nature and Natural Resources.
Gray’s River, Washington.

[1979] Rev: 1 (all published).

8. ARQUIVOS DE ENTOMOLOGIA. SERIE A. Escola de Agronomia “Eliseu

Maciel”. Pelotas.

[1958-1963] Rev: 1A, IB, 3-4, 8A, 13-14 (all published).

9. ARQUIVOS DE ENTOMOLOGIA. SERIE B. Escola de Agronomia “Eliseu

Maciel”. Pelotas.

[1959-1962] Rev: 1A, IB, 3-4, 8A, 13-14 (all published).

10. ATALA. The Journal of Invertebrate Conservation. The Xerces Society.

New Haven, Connecticut.

[1973— > ] Rev: 1-16.

Note: Subtitle varies \ Atala. Communication of the Xerces
Society (vols. 1-2); Atala. Journal of the Xerces Society (vols. 3-
12).

11. ATALANTA. Norsk Lepidopterologisk Selskaps Tidsskrift. Oslo.

[1967-1968] Rev: 1(1-2) (all published).

Continued as: Atalanta Norvegica (1969— > ).

12. ATALANTA. Zeitschrift der Deutschen Forschungszentrale fur

Schmetterlingswanderungen. Wurzburg.

[1964^ ] Rev: 1-21.

13. ATALANTA NORVEGICA. Norsk Lepidopterologisk Selskaps Tidsskrift.

Oslo.

[1969^ ] Rev: 1(3-5), 2-3, 4(1-2).

Formerly: Atalanta (Oslo) (1967-1968).

14. THE AURELIAN. A Journal of Lepidoptera. Beckley, Sussex.

[1978-1980] Rev: 1(1-4) (all published).

15. BAPTRIA. Suomen Perhostutkij ain Seura r. y. Lepidopterologiska

Sallskapet i Finland r. f. Helsinki.

[1976— > ] Rev: 3(1-2, ,4), , 5( , ,3-4), 6-11, 12(1).

16. BERICHT DES SCHLESISCHEN TAUSCH-VEREINS FUR

SCHMETTERLINGE. Brieg.

[1840-1846] Rev: 1-6.

94

J. Res. Lepid.

17. BIBLIOGRAPHIA EUROPAEA LEPIDOPTEROLOGICA. Societas

Europaea Lepidopterologica. Karlsruhe.

[1982-1988] Rev: 1979/80, 1981-1986.

Continued as: Bibliographica Palaearctica Lepidopterologica
(1989— > ).

18. BIBLIOGRAPHICA PALAEARCTICA LEPIDOPTEROLOGICA. Societas

Europaea Lepidopterologica. Karlsruhe.

[1989— > ] Rev: 10.

Formerly: Bibliographia Europaea Lepidopterologica (1982-
1988).

19. BOLETIM TECNICO DE SERICICULTURA. Campinas.

[1955-> ] Rev: None.

Formerly: Sericicultura. Servigo de Sericicultura (Campinas)
(1936-1952).

20. BOLETfN INFORMATIVO. SOCIEDAD MEXICANA DE

LEPIDOPTEROLOGfA, A.C. Mexico.

[1975-1982] Rev: 1-8, 8(1) (all published).

21. BULLETIN OF THE ALLYN MUSEUM. Sarasota, Florida.

[1971 — > ] Rev: 1-131.

22. THE BULLETIN OF THE ASSOCIATION OF MINNESOTA

ENTOMOLOGISTS. Saint Paul, Minnesota.

[1967-1972] Rev: 2-4, 5(1-2) (all published).

Formerly: Newsletter of the Association of Minnesota
Entomologists (1966-1967).

23. BULLETIN. CERCLE DES LEPIDOPTERISTES DE BELGIQUE.

BELGISCHE LEPIDOPTEROLOGISCHE KRING. Bruxelles.
[1972^ ] Rev: 3-15, 16(1-2).

24. BULLETIN OF THE CHEYENNE MOUNTAIN MUSEUM. Colorado

Springs, Colorado.

[1938-1939] Rev: 1(1-2) (all published).

25. BULLETIN. THE ENTOMOLOGISTS’ EXCHANGE ASSOCIATION.

Denver, Colorado.

[1936] Rev: 1(1-3) (all published).

26. THE BULLETIN OF THE HILL MUSEUM. A Magazine of

Lepidopterology. Witley, Surrey.

[1921-1932] Rev: 1-4 (all published).

27. THE BULLETIN OF THE LEPIDOPTEROLOGICAL SOCIETY OF

JAPAN. Kyoto.

[1946] Rev: 1(1-4) (all published.

28. BULLETIN OF THE SERICULTURAL EXPERIMENT STATION.

Tokyo.

[1918 — > ] Rev: None.

29. BULLETIN OF SERICULTURE AND SILK INDUSTRY. Uyeda.

[19 ?-1941?] Rev: None.

30. BULLETIN DE LA SOCI)T) DES L)PIDOPT)RISTES FRAN’AIS. Paris.

[1976-1978]

Rev: 1(1-3), 2(1-4) (all published).

31. BULLETIN DE LA SOCIETE LEPIDOPTEROLOGIQUE DE GENEVE.

Geneve.

[1905-1945] Rev: 1-7, 8(1-5).

29(l-2):92-104, 1990(91)

95

32. BUTLLETf DE LA SOCIETAT CATALANA DE LEPIDOPTEROLOGf A.

Mataro.

[1979-> ] Rev: 21-46, , , , ,51-52.

Formerly: Comunicacions de la Comissio de Lepidopterologia
(1977-1978).

33. BUTTERFLIES AND MOTHS. The Transactions of the

Lepidopterological Society of Japan. Kyoto.

[1949-1955] Rev: 1(2), 2-5 (all published).

Formerly: The Transactions of the Nippon Lepidopterological
Society (1945).

Continued as: Tyo to Ga. Transactions of the Lepidopterological
Society of Japan (1955^ ).

34. THE BUTTERFLY CLUB. Howell Mountain Butterfly Club. Sanitarium,

California.

[1946-1947] Rev: 1( , ,3-11, ), 2(1).

35. BUTTERFLKY & DISTRIBUTION. Society for the Study of Butterflies.

Aomori.

[1980-> ] Rev: None.

36. THE BUTTERFLY FARMER. A monthly magazine for amateur

entomologists. Truckee, California.

[1913-1914] Rev: 1(1-12) (all published).

37. BUTTERFLY NEWS. The popular butterfly & conservation newspaper.

Global Butterfly Services. Dorset.

[1985— > ] Rev: 1-13.

38. BUTTERFLY PARK NATURE CLUB NEWS. Roscoe, California.

[1929-1932] Rev: 1(1-12), 2(1-5), 3(1-8), 4(1).

39. CELASTRINA. Tsugaru Entomological Society. Aomori.

[1978-^ ] Rev: None.

40. CHO CHO. The Rhopalocerists’ Magazine. Fukuoka.

a [1978— > ] Rev: 1-4, ,6, 7(1-10).

41. CHOHO. Butterflies Talk. Club Musashino. Tokyo.

[1966-1969] Rev: None.

42. CLUB NOTES, MOTH AND BUTTERFLY CLUB. Gravel Switch,

Kentucky.

[71947-1953] Rev: [7](l-5).

Continued as: Notes on Moths and Butterflies (1953-1955).

43. COENONYMPHA. Hokkaido Lepidopterological Society. Hokkaido.

[1955— > ] Rev: None.

COMPTE RENDU DES TRAVAUX DU LABORATOIRE D’ETUDES DE
LA SOIE. Lyon.

See Rapport du Lahoratoire d’Etudes de la Soie.

44. COMUNICACIONS DE LA COMISSIO DE LEPIDOPTEROLOGfA.

Institucio Catalana d’Historia Natural. Mataro.

[1977-1978] Rev: 11-20.

Continued as: Butlleti de la Societal Catalana de
Lepidopterologia (1979— > ).

45. CONTRIBUTIONS TO THE NATURAL HISTORY OF THE

LEPIDOPTERA OF NORTH AMERICA. Decatur, Illinois.
[1911-1924] Rev: 1-4, 5(1-3) (all published).

96

J. Res. Lepid.

46. CORIDON. A Magazine for the Lepidopterist. Bourne End, Bucks.

[1960] Rev: 1(1-2) (all published).

Continued as: Coridon. A Review for the Lepidopterist. Series B
(1961-1963).

47. CORIDON. A Review for the Lepidopterist. Series A. Bourne End, Bucks.

[1961-1964] Rev: 1-6 (all published)

48. CORIDON. A Review for the Lepidopterist. Series B. Bourne End, Bucks.

[1961-1963] Rev: 1-5 (all published).

Formerly: Coridon. A Magazine for the Lepidopterist (1960).

49. CORRESPONDENZ-BLATT DES ENTOMOLOGISCHEN VEREINS

“IRIS” ZU DRESDEN. Dresden.

[1884-1888] Rev: 1(1-5) (all published).

Continued as: Deutsche Entomologische Zeitschrift
herausgegeben von der Entomologischen Verein “Iris” zu
Dresden. Lepidopterologische Hefte. { 1889-1902).

50. COSMIA. Gunma Heterocerists’ Society. Gunma.

[1965-1971] Rev: None.

51. DANAUS. The Newsletter about the Migratory Monarch Butterfly. Santa

Monica, California.

[1988^ ] Rev: 1-5.

52. DATA ON THE SURVEY OF SILKWORM RAISING. Ministry of

Agriculture and Forestry. Tokyo.

[1952— > ] Rev: None.

53. DEUTSCHE ENTOMOLOGISCHE ZEITSCHRIFT HERAUSGEGEBEN

VON DER GESELLSCHAFT IRIS ZU DRESDEN IN
VERBINDUNG MIT DER DEUTSCHEN
ENTOMOLOGISCHEN GESELLSCHAFT ZU BERLIN.
LEPIDOPTEROLOGISCHE HEFTE. Dresden.

[1889-1902] Rev: 2-14 (all published).

Formerly: Correspondenz-Blatt des Entomologischen Vereins
“Iris” zu Dresden (1884-1888).

Continued as: Deutsche Entomologische Zeitschrift Iris (1902-
1944).

54. DEUTSCHE ENTOMOLOGISCHE ZEITSCHRIFT IRIS [or “IRIS”].

Deutsche Entomologische Verein Iris. Dresden.

[1902-1944] Rev: 15-57 (all published).

Formerly: Deutsche Entomologische Zeitschrift herausgegeben
von der Gesellschaft Iris zu Dresden in Verbindung mit der
Deutschen Entomologischen Gesellschaft zu Berlin.
Lepidopterologische Hefte (1889-1902).

55. DISCUSSION ON MULBERRY AND SILKWORM TECHNOLOGY.

Japan Sericulture Information Association. Tokyo.

[1947— > ] Rev: None.

56. ENCYCLOPEDIE ENTOMOLOGIQUE. SERIE B. III. LEPIDOPTERA.

Recueil d’Etudes Biologiques et Systematiques sur les
Lepidopteres du Globe. Paris.

[1925-1930] Rev: 1-3 (all published).

29(l-2):92-104, 1990(91)

97

57. THE ENTOMOLOGIST’S EXCHANGE NEWS. Colorado Springs,

Colorado.

[1937-1942] Rev: 2( , ,3-12), 3(1-7), 4(1-10), 5(1-11), 6(1-10),
7(1-2).

Formerly: Bulletin. The Entomologists' Exchange Association
(1936).

58. FRASS. An Ocassional Journal of Faralepidopterology .

[1973-1976] Rev: 1-4 (all published).

59. FUTAO. Futao-Kai. Osaka.

[1989-> ] Rev: 1-3.

60. GALATHEA. Berichte des Kreises Niirnberger Entomologen e.V.

Niirnberg.

[1985— » ] Rev: 1(1-4), 2(1-3).

61. GARUI TSUSHIN. The Japan Heterocerists’ Journal. Japan

Heterocerists’ Society. Tokyo.

[1954 > ] Rev: 1-142, suppls. 1-2.

62. HAMPSHIRE BUTTERFLY REPORT. British Butterfly Conservation

Society. Hampshire Branch.

[1985— > ] Rev: 1986.

63. HERBIPOLIANA. Buchreihe zur Lepidopterologie. Marktleuthen.

[1983-> ] Rev: 1(1-2).

64. INDIAN JOURNAL OF SERICULTURE. Central Silk Board. Bombay.

[1962-> ] Rev: None.

65. INDIAN SILK JOURNAL. A quarterly devoted to Indian

Sericulture and Silk. Central Silk Board. Bombay.

[ ? - ? ] Rev: None.

66. INSECT MIGRATION STUDIES. Annual Newsletter to Research

Associates. West Hill, Ontario.

[1973— > ] Rev: 10-12.

Formerly: Insect Migration Studies. Newsletter to Research
Associates (19 7-1972).

67. INSECT MIGRATION STUDIES. Newsletter to Research Associates.

Cobourg, Ontario.

[19 7-1972] Rev: 8-9.

Continued as: Insect Migration Studies. Annual Newsletter to
Research Associates { 1973— » ).

68. IWASE. Osaka.

[1983— > ] Rev: 1-4.

THE JAPAN HETEROCERISTS’ JOURNAL.

See Garui Tsushin.

JOURNAL OF THE LEPIDOPTEROLOGICAL SOCIETY OF JAPAN.

See Yadoriga.

69. JOURNAL OF THE LEPIDOPTERISTS’ SOCIETY.

[1959-> ]

Rev: 13-43, 44(1-3).

Formerly: The Lepidopterists’ News (1947-1959).

70. THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA. Lepidoptera

Research Foundation, Inc. Beverly Hills, California.

[1962— > ] Rev: 1-28.

98

J. Res. Lepid.

71. THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA

NEWSLETTER. Lepidoptera Research Foundation, Inc.
Beverly Hills, California.

[1980— > ] Rev: 1-9.

72. THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA.

SUPPLEMENT. Lepidoptera Research Foundation, Inc.
Beverly Hills, California.

[1985^ ] Rev: 1.

73. JOURNAL OF SERICULTURAL SCIENCE. Japanese Society of

Sericultural Science. Tokyo.

[1930— > ] Rev: None.

74. KAKOCHO. Nagoya Entomological Society. Aichi.

[1949-> ] Rev: None.

75. KENTUCKY LEPIDOPTERIST. Nesletter of the Society of Kentucky

Lepidopterists. Louisville, Kentucky.

[1975— > ] Rev: 1-16, 17(1).

76. KONAMUSHI. Numata Rhopalocerists’ Society. Gunma.

[1978^ ] Rev: None.

77. KOREAN JOURNAL OF SERICULTURAL SCIENCE.

[19 ?— » ] Rev: None.

78. KORRESPONDENZBLATT. BEILAGE ZUR DEUTSCHEN

ENTOMOLOGISCHEN ZEITSCHRIFT “IRIS”. Dresden.
[1910] Rev: 1-12 (all published).

79. KUROSESERI. Kyushu Rhopalocerists’ Society. Kumamoto.

[1966-1974] Rev: None.

80. KYUSHU NO CHO. Kyushu Rhopalocerists’ Society. Kumamoto.

[1956-1957] Rev: None.

81. LEPIDOPTERA. Medlemsblad for Lepidopterologisk Forening.

Kpbenhavn.

[1946-1951] Rev: 1946-1948/9, 1950(1).

Continued as: Lepidoptera. Ny Serie (1965-^ ).

82. LEPIDOPTERA. NY SERIE. Lepidopterologisk Forening. Kpbenhavn.

[1965^ ] Rev: 1-4, 5(1-2), suppls. 1-5.

Formerly: Lepidoptera (Kpbenhavn) (1946-1951).

83. LEPIDOPTERA. Official Bulletin of the Boston Entomological Club.

Forest Hills, Massachusetts.

[1918-1921] Rev: 2-4, 5(1-2) (all published).

Formerly: The Lepidopterist (Roslindale) (1916-1917).

84. THE LEPIDOPTERA GROUP OF 1968 NEWSLETTER. Vejle.

[197? — > ] Rev: 1(6), 2(5-7, 11-15).

85. THE LEPIDOPTERA GROUP OF 1968 SUPPLEMENT. Vejle

[1977— > ] Rev: 1-2, , , , , ,8.

86. THE LEPIDOPTERIST. Official Bulletin of the Boston Entomological

Club. Roslindale, Massachusetts.

[1916-1917] Rev: 1(1-13) (all published).

Continued as: Lepidoptera (Forest Hills) (1918-1921); and The
Lepidopterist (Salem) (1918-1931).

87. THE LEPIDOPTERIST. Salem, Massachusetts.

[1918-1931] Rev: 2-4, 5(1-3) (all published).

Formerly: The Lepidopterist (Roslindale) (1916-1917).

29(l-2):92-104, 1990(91)

99

88. LEPIDOPTERISTS’ NEWS. Florida Society of Lepidopterists of the

Florida Society of Natural History. Miami, Florida.

[1932-1933] Rev: 1(1-2) (all published).

89. THE LEPIDOPTERISTS’ NEWS. The Monthly Newsletter of the

Lepidopterists’ Society. Cambridge, Massachusetts.

[1947-1959] Rev: 1-12 (all published).

Continued as: Journal of the Lepidopterists’ Society (1959^ ).

90. LEPIDOPTEROLOGISCHE RUNDSCHAU. Wien.

[1927-1928] Rev: 1-2 (all published).

91. LILAC. Minoo Rhopalocerists’ Society. Osaka

[1952-1962] Rev: None.

92. LINNEANA BELGICA. Revue Beige d’Entomologie. Bruxelles.

[1958^ ] Rev: 1-11, 12 (1-4).

93. LUEHDORFIA NEWS. Luehdorfia Lovers Association. Nagano.

[1965-1970] Rev: None.

94. MELANARGIA. Nachrichten der Arbeitsgemeinschaft rheinisch-

westfalischer Lepidopterologen. Leverkusen.

[1989-^ ] Rev: 1(1-3).

Formerly: Mitteilungen der Arbeitsgemeinschaft rheinisch-
westfalischer Lepidopterologen (1977-1988).

95. MEMOIR. THE LEPIDOPTERISTS’ SOCIETY. New Haven,

Connecticut.

[1964— > ] Rev: 1-3.

96. MEMOIRS OF THE TSUKADA COLLECTION. Plapac Co., Ltd. Tokyo.

[19 ?-> ] Rev: 3.

97. MEMORIES DE LA SOCIETAT CATALANA DE LEPIDOPTEROLOGf A.

Mataro.

[1986^ ] Rev: None.

98. MICROLEPIDOPTERA OF THAILAND. Scientific Results of the

Lepidopterological Expeditions of the University of Osaka
Prefecture to Thailand. Entomological Laboratory, University of
Osaka Prefecture. Sakai, Osaka.

[1987-* ] Rev: 1-2.

99. THE MID-CONTINENT LEPIDOPTERA SERIES. Saint Paul,

Minnesota.

[1967-1972] Rev: 1-3, 4(50-61), 5( ,63).

100. MIKABO. Tano Rhopalocerists’ Society. Gunma.

[1950-1955] Rev: None.

101. MITTEILUNGEN DER ARBEITSGEMEINSCHAFT

OSTWESTFALISCH-LIPPISCHER ENTOMOLOGEN.

Bielefeld.

[1965^ ] Rev: None.

102. MITTEILUNGEN DER ARBEITSGEMEINSCHAFT RHEINISCH-

WESTFALISCHER LEPIDOPTEROLOGEN. Diisseldorf.
[1977-1989] Rev: 1-5, Beiheft 1 (all published).

Continued as: Melanargia. Mitteilungen der Arbeitsgemeinschaft
rheinisch-westfalischer Lepidopterologen (1989— > ).

103. MITTEILUNGEN DES ENTOMOLOGISCHEN VEREINES

POLYXENA. Wien.

[1906-1913] Rev: 1-6 (all published).

100

J. Res. Lepid.

104. MIYAZAKI NO CHO. Miyazaki Rhopalocerists’ Society. Miyazaki.

[1968- > ] Rev: None.

105. MONTHLY BULLETIN OF THE REPORT ON MOTHS AND

BUTTERFLIES. Middle Thames Natural History Society.

[ ? -1965] Rev: 12a, 13-17, , ,20-35, , ,38-43.

106. NACHRICHTEN DES ENTOMOLOGISCHEN VEREINS APOLLO.

Frankfurt-am-Main.

[1976-1979] Rev: 1-4 (all published).

Continued as: Nachrichten des Entomologischen Vereins Apollo.
Neue Folge (1980— » ).

107. NACHRICHTEN DES ENTOMOLOGISCHEN VEREINS APOLLO.

NEUE FOLGE. Frankfurt-am-Main.

[1980-^ ] Rev: 1-10, suppls. 1-7.

108. NAGANO GAHO. Nagano Heterocerists’ Society. Nagano.

[1976-> ] Rev: None.

109. NEOTROPICAL LEPIDOPTERA NEWS. Gainesville, Florida.

[1987] Rev: 1 (all published).

Continued as: Tropical Lepidoptera News (1989— > ).

1 10. NEUE ENTOMOLOGISCHE NACHRICHTEN AUS DEM

ENTOMOLOGISCHEN MUSEUM DR. ULF EITSCHBERGER.
Beitrage zur Okologie, Faunistik und Systematik von
Lepidopteren. Marktleuthen.

[1988^ ] Rev: 24-27.

111. NEWS. THE BRITISH BUTTERFLY CONSERVATION SOCIETY.

Quorn, Leicestershire.

[1973— > ] Rev: 12-37.

Formerly: Quarterly News. The British Butterfly Conservation
Society (1972-1973).

1 12. NEWS OF THE LEPIDOPTERISTS’ SOCIETY.

[1959— > ] Rev: 1959-1990, 1991(1-3).

113. NEWSLETTER OF THE ASSOCIATION OF MINNESOTA

ENTOMOLOGISTS. Saint Paul, Minnesota.

[1966-1967] Rev: 1(1-4) (all published).

Continued as: The Bulletin of the Association of Minnesota
Entomologists (1967-1972).

114. NEWS. NOUVELLES. NACHRICHTEN. SOCIETAS EUROPAEA

LEPIDOPTEROLOGICA. Karlsruhe.

[1976— > ] Rev: 1-17.

115. NEWS SHEET. THE BRITISH BUTTERFLY CONSERVATION

SOCIETY. Over Compton, Sherborn, Dorset.

[1968-1971] Rev: 1, ,3, ,5-8.

Continued as: Quarterly News. The British Butterfly
Conservation Society (1972-1973).

116. NOTA LEPIDOPTEROLOGICA. Societas Europaea Lepidopterologica.

[1977-* ] Rev: 1-11, 12(1-3).

117. NOTA LEPIDOPTEROLOGICA. SUPPLEMENT. Societas Europaea

Lepidopterologica.

[1989— > ] Rev: 1-2.

118. NOTES ON MOTHS AND BUTTERFLIES. Gravel Switch, Kentucky.

[1953-1955] Rev: 7(9).

Formerly: Club Notes, Moth and Butterfly Club (71947-1953).

29(l-2):92-104, 1990(91)

101

119. NOVITATES MACROLEPIDOPTEROLOGICAE. Dresden.

[1926-1930] Rev: 1-5 (all published).

120. THE OHIO LEPIDOPTERIST. Newsletter of the Ohio Lepidopterists.

Columbus, Ohio.

[1979-> ] Rev: 1-12.

121. THE OHIO LEPIDOPTERISTS RESEARCH REPORTS. Columbus,

Ohio.

[1980-^ ] Rev: None.

122. ORIENTERING FRA LEPIDOPTEROLOGISKE MEDDELELSER.

Aarhus.

[1975— > ] Rev: 2-12.

123. PAPILIO. Gunma Rhopalocerists’ Society. Gunma.

[1950-1953]. Rev: None.

124. PAPILIO. Organ of the New York Entomological Club. New York, New

York.

[1881-1885] Rev: 1-3, 4(1-10) (all published).

125. PAPILIO. NEW SERIES. Lakewood, Colorado.

[1981— > ] Rev: 1-5.

126. PAPILIO INTERNATIONAL. The Lepidoptera Group of 1968. Vejle.

[1983— > ] Rev: 1-6.

127. PARNASSIANA. Neubrandenburg.

[1930-1939] Rev: 1-5 (all published).

128. PARNASSIUS. Awaji Entomological Society. Hyogo.

[1967^ ] Rev: None.

129. PHEGEA. Vereniging voor Entomologie van de Koninklijke

Maatschappij voor Dierkunde van Antwerpen. Antwerpen.
[1973— > ] Rev: 1-14, 15(1-2).

130. PRIAMUS. Results of the Scientific Research on Entomology. Ankara.

[1981— > ] Rev: 1-4, 5(1/2).

131. PROCES VERBAUX/VERSLAGEN. CERCLE DES

LEPIDOPTERISTES DE BELGIQUE. BELGISCHE
LEPIDOPTEROLOGISCHE KRING. Bruxelles.

[1973] Rev: 1968-1971.

132. PUBLICACIONES ESPECIALES. SOCIEDAD MEXICANA DE

LEPIDOPTEROLOGfA, A. C. Mexico.

[1976 — > ] Rev: 1-4.

133. THE PUPAPER. Butterfly World. A classified Newsletter. Pendleton,

Indiana.

[1978? — > ] Rev: 5(1-3).

134. QUARTERLY NEWS. THE BRITISH BUTTERFLY CONSERVATION

SOCIETY. Over Compton, Sherborn, Dorset.

[1972-1973] Rev: 9-11 (all published).

Formerly: News Sheet. The British Butterfly Conservation
Society (1968-1971).

Continued as: News. The British Butterfly Conservation Society
(1973^ ).

102

J. Res. Lepid.

135. RAPPORT DU LABORATOIRE D’ETUDES DE LA SOIE. Lyon.

[1887-1931] Rev: 1-17 (all published).

Note: Volumes 1-7 (1887-1895) and 16-17 (1923-1931) as
“Rapport...”; vols. 8-9 (1897-1899) and 13-15 (1909-1919) as
“Compte Rendu des Travaux du Laboratoire d’Etudes de la
Soie”; and vols. 10-12 (1901-1906) as “Annales du Laboratoire
d’Etudes de la Soie”.

136. REVISTA DE LA SOCIEDAD MEXICANA DE

LEPIDOPTEROLOGIA, A. C. Mexico.

[1975-* ] Rev: 1-12.

137. REVUE FRANgAISE DE LEPIDOPTEROLOGIE . Paris.

[1938-1958] Rev: 9(5/6), 10-15, 16(1-4) (all published).
Formerly: L Amateur de Papillons (1922-1938).

138. REVUE DU VER A SOIE. JOURNAL OF SILKWORM. Station de

Recherches Sericicoles d’Ales. Ales,.Gard.

[1949-1977] Rev: 1-18 (all published).

Continued as: Sericologia (1979^ ).

139. THE RHOPALOCEOLOGICAL [sic] MAGAZINE. The

Rhopaloceological [sic] Society. Tokyo.

[1936] Rev: 1( ,2-4).

THE RHOPALOCERISTS’ MAGAZINE.

See Cho Chd.

140. SAPHIRINUS. Amagasaki Rhopalocerists’ Society. Hyogo.

[1953-1956] Rev: None.

141. SCOPARIA. Otaru Rhopalocerists’ Society. Hokkaido.

[1960-1961] Rev: None.

142. SERICICULTURA. Consiliul Superior al Agriculturii. Rucuresti.

[1965^ ] Rev: None.

143. SERICICULTURA. Jornal. Industria de Seda Nacional. Campinas.

[1927-1938] Rev: None.

144. SERICICULTURA. Revista Mensal da S. A. Industria de Seda

Nacional. Campinas.

[1934-1935] Rev: None.

145. SERICICULTURA. Servigo de Sericicultura. Campinas.

[1936-1952] Rev: None.

Continued as: Boletim Tecnico de Sericicultura (1955— > ).

146. SERICOLOGIA. Revue de Vers a Soie. Journal of the Silkworms.

Commission Sericicole Internationale. Lyon.

[1979— > ]

Formerly: Revue du Ver a Soie. Journal of Silkworm (1949-
1977).

147. SERICULTURE. Journal of the Faculty of Textiles and Sericulture.

Shinshu University. Series E. Ueda.

[1955-> ] Rev: None.

148. SHILAP. REVISTA DE LEPIDOPTEROLOGIA. Sociedad Hispano-

Luso-Americana de Lepidopterologia. Madrid.

[1973— » ] Rev: 1-18.

149. SOUTHERN LEPIDOPTERISTS’ BULLETIN. Jacksonville, Florida.

[19 ?— > ] Rev: None.

150. SOUTHERN LEPIDOPTERISTS’ NEWS. Jacksonville, Florida.

[1979-> ] Rev: 1-3, 4(1-3).

29(l-2):92-104, 1990(91)

103

151. SPECIAL BULLETIN OF THE LEPIDOPTEROLOGICAL SOCIETY

OF JAPAN. Osaka.

[1965-1971] Rev: 1-5 (all published).

152. SUGI TOGE. Kyoto Rhopalocerists’ Society. Kyoto.

[1976^ ] Rev: None.

153. THE SWALLOWTAIL. The Journal of the Butterfly and Moth Stamp

Society. Wiltshire.

[1986— > ] Rev: 1(1-4).

154. TERRITORY. Setagaya Rhopalocerists’ Association. Tokyo.

[1977 — > ] Rev: None.

155. THECLA. Yamashiro Rhopalocerists’ Society. Kyoto.

[1942-1944] Rev: None.

156. TINEA. The Japan Heterocerists’ Society. Tokyo.

[1953 — > ] Rev: 1-11, 12(1-25).

157. TOKURANA. Acta Rhopalocera. Mishima City.

[1981— > ] Rev: 1-5, , ,8(1), 9-15, 16(1-2).

Note: Vols. 1-4 as “Acta Rhopalocera”; 5-9 as “Acta
Rhopalocerologica”; and 10-16 as “Reports on Rhopalocera”.

158. TOKURANA SPECIAL. The Butterflies’ Magazine. Mishima City.

[1985^ ] Rev: l(pp. 1-36).

THE TRANSACTIONS OF THE LEPIDOPTEROLOGICAL SOCIETY
OF JAPAN.

See Tyo to Ga.

159. THE TRANSACTIONS OF THE NIPPON LEPIDOPTEROLOGICAL

SOCIETY. Kyoto.

[1945] Rev: 1(1) (all published).

Continued as: Butterflies and Moths. The Transactions of the
Lepidopterological Society of Japan (1949-1955).

160. TREBALLS DE LA SOCIETAT CATALANA DE

LEPIDOPTEROLOGIA. Mataro.

[1978— > ] Rev: 1-7.

161. TROPICAL LEPIDOPTERA. Association for Tropical Lepidoptera.

Gainesville, Florida.

[1990— > ] Rev: 1(1-2).

162. TROPICAL LEPIDOPTERA NEWS. Association for Tropical

Lepidoptera. Gainesville, Florida.

[1989— > ] Rev: 2-4, 1991(1-2).

Formerly: Neotropical Lepidoptera News (1987).

163. TYO TO GA. Transactions of the Lepidopterological Society of Japan.

Osaka.

[1955— > ] Rev: 6-40.

Formerly: Butterflies and Moths. The Transactions of the
Lepidopterological Society of Japan (1949-1955).

164. UTAHENSIS. Journal of the Utah Lepidopterists’ Society. Salt Lake

City, Utah.

[1981— > ] Rev: 7(1).

165. VLINDERS.

[1986— > ] Rev: None.

166. WEEKLY BUTTERFLY’S TSU. Mokuyosha Company. Tokyo.

[1975^ ] Rev: None.

104

J. Res. Lepid.

167. WINGS. Newsletter of the Xerces Society. New Haven, Connecticut.

[1974— > ] Rev: 1-15.

168. XERCES SOCIETY BUTTERFLY COUNTS. The Xerces Society. New

Haven, Connecticut.

[1981 — > ] Rev: 1979-1989.

Note: Some issues are supplements to Atala, q.v.

169. XERCES SOCIETY EDUCATIONAL LEAFLET. The Xerces Society.

New Haven, Connecticut.

[1978— > ] Rev: 1-9.

170. XERCES SOCIETY INTERIM NEWSLETTER. The Xerces Society.

New Haven, Connecticut.

[1978— > ] Rev: 1-2.

171. THE XERCES SOCIETY SELF-HELP SHEET. The Xerces Society.

New Haven, Connecticut.

[71973— > ] Rev: 2-4, ,6-7.

172. YADORIGA. Lepidopterological Society of Japan. Osaka.

[1955 — > ] Rev: 22-23, ,25-131.

173. YUGATO. Yugakai (Heterocerists’ Association). Niigata.

[1959-^ ] Rev: None.

174. ZEITSCHRIFT FUR LEPIDOPTEROLOGIE. Fachorgan fur die

Gesamte Schmetterlingskunde. Krefeld.

[1950-1955] Rev: 1-3 (all published).

175. ZEPHYRUS. Chorui Dokokai. Fukuoka.

[1929-1947] Rev: 1-9 (all published).

Journal of Research on the Lepidoptera

29(1-2):105-133, 1990(91)

The Butterflies (Lepidoptera) of the Tuxtlas Mts.,
Veracruz, Mexico, Revisited: Species-Richness and
Habitat Disturbance.

Robert A. Raguso

Dept, of Biology, Yale University, New Haven, CT 06511 USA!

Jorge Llorente-Rousquets

Museo de Zoologia, Facultad de Ciencias, Universidad Nacional Autonoma de Mexico,
Apartado Postal 70-399 Mexico D.F., CP 04510

Abstract. Checklists of the butterflies (Lepidoptera) collected in two
rainforest study sites in the Tuxtlas Mts., Veracruz, Mexico are pre-
sented. A total of 182 species of butterflies were recorded at Laguna
Encantada, near San Andres Tuxtla, and 212 species were recorded
from the nearby Estacion de Biologia Tropical “Los Tuxtlas”
(EBITROLOTU). We collected 33 species not included in G. Ross’
(1975-77) faunistic treatment of the region, 12 of which are new species
records for the Tuxtlas. We present a list of the skipper butterflies
(Hesperioidea) of the Tuxtlas, including a state record for the giant
skipper, Agathymus rethon.

At both study sites, we observed seasonal patterns in species abun-
dance during periods of reduced precipitation. Our data indicate an
apparent increase in butterfly species-richness in the Tuxtlas over the
last 25 years. This increase reflects more efficient sampling due to
advances in lepidopteran ecology and improved collecting methods, as
well as the effects of habitat disturbance. A comparison between the
butterfly faunas of the two rainforest sites revealed that a higher
percentage of weedy, cosmopolitan species were present at Laguna
Encantada, the smaller, more disturbed site. We anticipate further
changes in butterfly species-richness and faunal composition as the
mosaic of habitats in the Tuxtlas continue to be modified.

Introduction

Historically, the rainforests of the Sierra de Los Tuxtlas, Veracruz,
Mexico have been the focus of varied and extensive ecological research
(Ross 1966, 1975-77, Soto 1976, Horvitz and Beattie 1980, Horvitz and
Schemske 1988, de la Cruz and Dirzo 1987, Popma, et al. 1988, etc.).
Following Ross’( 1975-77) comprehensive three year survey ofthe Tuxtlas
butterfly fauna, studies have generated additional records and range
extensions for many butterfly taxa (Welling 1982, 1983, G. Busby,
unpub. data, R. Robbins and C. Beutelspacher, pers. comms.). These
records provide a comparative base from which to examine changes in the
profile of a Neotropical lepidopteran fauna in a region of high human
impact.

* Present address: Department of Biology, University of Michigan, Ann Arbor, MI 48109-
1048 USA.

106

J. Res. Lepid.

Recent expansion of cattle ranches, agricultural development and
logging have reduced much of the formerly extensive rainforest of the
Tuxtlas to scattered enclaves on steep slopes, in secluded volcanic craters
or in local parks, resulting in a mosaic of heterogeneous rainforest
patches. A lowland Neotropical wet forest may suffer local extinction of
highly specialized plant species as a result of clear cutting and land
conversion (Gomez-Pompa 1973). How has the Tuxtlas butterfly fauna
responded to such human disturbance?

Checklists of the butterfly species collected in two separate rainforest
study sites in the Tuxtlas are given here. Laguna Encantada and the “Los
Tuxtlas” biological field station of the Universidad Nacional Autonoma
de Mexico (EBITROLOTU, by Mexican convention) are similar in eleva-
tion (ca. 350 m) and vegetation, but differ greatly in both forest area
(Laguna Encantada = 56 ha, EBITROLOTU » 700 ha) and the extent
of disturbance. We report new species records for the region and compare
our findings to those of earlier studies in the Tuxtlas. Preliminary
evidence is presented indicating seasonal peaks in adult flight activity
among the butterfly faunas of Laguna Encantada and EBITROLOTU,
corresponding to periods of reduced precipitation. Finally, we calculate
the percentages of weedy, cosmopolitan species associated with dis-
turbed habitats that presently comprise the butterfly faunas of each site
as an index of disturbance.

Materials and Methods

STUDY SITES

The Tuxtlas Mountains are an isolated volcanic range 88.5 km long and 53.1
km wide stretching northwest to southeast along the Isthmus of Tehuantepec
(Ross 1975-77, see Figure 1). The region is characterized by a warm-humid
climate and represents the northernmost extension of evergreen tropical rainforest
in the Americas (Gomez-Pompa 1973). Most annual precipitation occurs from
June to October and varies greatly across the Sierra, ranging from 1996 mm at
San Andres Tuxtla to 4700 mm at EBITROLOTU (Soto 1976, Alvarez 1982).

Our objectives were to collect and identify all butterfly species occuring at two
study sites within the Tuxtlas. Ross’ study generated lists of butterflies charac-
teristic of vegetational formations such as montane rainforest, swamp forest,
savanna, etc. Both Laguna Encantada and EBITROLOTU are examples of lower
montane rainforests ( selva alta perenifolia, Gomez-Pompa 1973, Ross 1975-77).
Laguna Encantada (el. 350 m.) is located 2 km northeast of San Andres Tuxtla
and features a small freshwater caldera flanked by steep volcanic slopes and
dense rainforest dominated by Ceiba pentandra and Ficus sp. trees. Cattle
grazing and logging have created numerous disturbed areas on the slopes, where
successional trees such as Cecropia mexicana, Cassia sp., Piper sp., and Annona
sp. grade into open grassy pastures and clusters of Agave sp. dot the high lip of
Laguna Encantada’s crater. Plants such as Crotalaria vitellina, Bidens pilosa
var. bimucronata, Cordia alliodora, C. spinescens and Lantana sp. provide
nectar resources not usually found within primary rainforests. Due to selective
logging, human disturbance occurs within the forest as well as on its periphery.
Numerous paths traverse treefall gaps, deep forest, stream beds and grassy
meadows over an area of 54 ha (C. Horvitz, pers. comm.).

29( 1-2): 105-133, 1990(91)

107

Raguso and Llorente

Figure 1 . Maps of the Sierra de Tuxtlas, Veracruz, Mexico, after Ross (1975 above)
and Lot-Helgueras (1979 below).

92°15' 92°10' 92°05’ 92°00'

108

J. Res. Lepid.

The second study was conducted ca. 15 km northeast of the Laguna Encantada
site at EBITROLOTU. This 700 ha reserve is located ca. 34 km north of Lago
Catemaco and towards the Gulf of Mexico and ranges from 150 to 530 m in
elevation (Lot-Helgueras 1976, Alvarez 1982). Its steep slopes are cloaked in
lower montane rainforest, with an understory of Astrocarium mexicanum palms
(Ross 1975-77). Although primary rainforest extends well beyond EBITROLOTU’s
western and southern boundaries, butterflies were only collected within the
confines of the reserve. Disturbed areas are found to the north of the station and
secondary forest flanks the road to Montepio at the reserve’s eastern end (S.
Guevara-Sada and A. Gomez-Pompa 1976). The EBITROLOTU site features
greater topodiversity and more primary forest than the smaller Laguna Encantada
site (see Lot-Helgueras 1976, de la Cruz and Dirzo 1987).

Cursory surveys were performed from 0900 to 1500 hrs on 25. vii and 4.viii. 1985
at nearby Playa Azul, situated on the northeast shore of Lago Catemaco (see
Figure 1). Volcanic craters, swamps, coffee and banana plantations and rainforest
remnants can all be found within a three km radius of this site (Ross 1975-77).
The Playa Azul surveys were performed with the sole intention of identifying
regional species records. We will include these surveys in our discussion of
butterfly species-richness in the Tuxtlas but exclude them from our comparisons
of the Laguna Encantada and EBITROLOTU study sites.

COLLECTION OF SPECIMENS

Butterflies were collected from 20.vii to 31.viii. 1985 at Laguna Encantada, and
during most of 1985, 1986 and parts of 1987 at EBITROLOTU. Skippers
(Hesperioidea) were not collected at EBITROLOTU. We employed similar
collecting methods at both study sites. Prominent nectar sources, oozing sap,
puddles, treefall gaps and creek beds were inspected during three half-hour
intervals: 0730-0800, 1200-1230, and 1500-1530 hrs. Van Someren-Rydon
traps of vertical cylindrical netting were hung above rotting bananas, mangos,
animal waste and carrion, and placed near lightgaps, forest margins and moist
trails (cf. Beutelspacher 1982, J. de la Maza and R. de la Maza 1985a, DeVries
1987). Traps were inspected three times daily for captured specimens at both
Laguna Encantada and EBITROLOTU.

IDENTIFICATION OF SPECIMENS

Butterflies from Laguna Encantada were curated and deposited at the Yale
Peabody Museum, New Haven, CT with representative specimens retained by
the Museo de Zoologia, Facultad de Ciencias, U.N.A.M., Mexico City. Specimens
collected at EBITROLOTU are currently housed in the systematic insect collec-
tion at that field station.

Specimens were identified using a variety of sources including Evans (1951-
1955), Singer, DeVries and Ehrlich (1983), Lamas (1987), DeVries (1987),
Godman and Salvin (1879...), Hoffmann (1940), Jenkins (1983, 1984, 1985a) and
Scott ( 1986). We used reference collections housed at the Yale Peabody Museum
of Natural History, American Museum of Natural History, Los Angeles County
Museum of Natural History, Museo de Zoologia U.N.A.M., California Academy
of Sciences, and the Essig Museum of Entomology, University of California,
Berkeley. Individual experts were consulted in order to identify difficult taxa (see
Acknowledgements).

29(1-2):105-133, 1990(91)

109

FAUNAL SURVEYS AND COMPARISONS
Butterfly species data from Laguna Encantada and EBITROLOTU were
examined in three different contexts. First, our findings were compared to those
of previous studies performed within the Tuxtlas (Ross 1975-77, etc.) and
throughout tropical Mexico in an effort to identify new regional records.

Second, we suspected that the brevity of our study at Laguna Encantada may
have produced unrepresentative samples due to seasonality effects. Beutelspacher
(pers. comm.) warns that a six week survey may not account for highly seasonal
or reclusive butterfly species. Clench (1979) described a method by which
sampling effort is used to calculate the species total for a given locality:

S = Se (N)/(N + K)

where S = cumulative total of species observed
Se = total species theoretically present
N = cumulative total of collector/observer hours
K = constant of “collectability”

This equation describes saturation curves such as a substrate-limited enzy-
matic reaction (e.g. Michaelis-Menten equation) or, in this case, the observation
and collection of butterfly species in a finite area over time. Given local
species introductions, extinctions and chance events, a plot of species collected (S)
versus time spent collecting (N) shows a sharply rising curve which tapers
asymptotically at the limit of total species (Se) theoretically present at that site.
This limit is calculated by solving simultaneous equations for (Se) and (K) with
data from two well-spaced points on the fitted curve (Clench 1979). Deviations
may accrue if the collecting protocol is changed or if butterfly abundance is
seasonal. A double reciprocal transformation linearizes these data and facili-
tates the calculation of Se (cf. Lineweaver-Burke [case II] equation in the
enzymatic analogy; W. Watt, pers. comm.).

Finally, we sought to assess the effects of habitat disturbance upon the
butterfly faunas of the two rainforest sites. Recent studies have focused upon the
presence or absence of bird species restricted to primary forests as indicators of
human disturbance in temperate rainforests (e. g. Strix occidentalis ; Franklin,
1988). While certain butterfly species such as Nessaea aglaura and Heliconius
sapho leuce are thought to be quite habitat-restricted and intolerant of distur-
bance, other forest species (e. g. Parides iphidamas, Cissia libye, Battus sp.,
ithomiines, sphingid moths) are known to occur in numerous habitat types or to
migrate between wet and dry forests or along altitudinal gradients (DeVries
1987, Janzen 1984). Ross' listings of the butterfly species indicative of human
disturbance were more useful for our purposes. We compiled a list of species
characteristic of human disturbance by pooling Ross’ (1975-77) species listings
from two such conditions present near our study sites: abandoned maize fields
( milpas ) and pastures (see Appendix 1). From this list, we calculated the
percentage of such species among the respective faunas of each site and compared
them using a 2 x 2 G-test of independence (Sokal and Rohlf, 1981: pp. 737-738).
We used the ratio of collector hours/hectare calculated for each site to standardize
these data for sampling effort and then repeated the 2x2 G-test comparison.

110

J. Res. Lepid.

Results

CHECKLISTS, SPECIES-RICHNESS AND NEW RECORDS:

We collected a total of 1293 specimens representing 254 species of
butterflies and skippers. Of this number, 33 species were not listed in
Ross’ survey of the Tuxtlas butterfly fauna, and 13 are new regional
records (see Appendix 2).

During a six week study period we collected 146 species of butterflies
and 36 species of skippers at Laguna Encantada. Although skipper
species records for Veracruz (especially the Catemaco area) can be found
throughout the literature (Evans 1951-1955, Hoffmann 1940, Freeman
1966, 1969b), our list of Hesperioidea is the first published for the
Tuxtlas, and includes the first recorded observation of Agathymus rethon
(Megathymidae) in this area. The surveys at Playa Azul produced one
regional butterfly species record, Calydna sturnula hegias (Riodinidae),
and four additional skipper species.

A total of 684 specimens representing 212 species of butterflies
(Papilionoidea) were collected from February 1985 to June 1987 at
EBITROLOTU (see checklist, Appendix 2). This total includes 11 new
species records for the Tuxtlas region: Parides lycimenes septentrionalis ,
Eury tides marchandi (Papilionidae), Sarota chrysus, Cremna thasus
subrutila,Napaea eucharila picina, Emesis vulpina (Riodinidae), Zizula
tulliola (Lycaenidae), Memphis neidhoeferi , Memphis xenocles, Cissia
renata disafecta , and Megisto ruhricata anahelae (Nymphalidae).

Compared with Ross’ data for the entire region, the number of butterfly
species at Laguna Encantada, Playa Azul and EBITROLOTU combined
(254) was low. This was expected, as Ross’ study had encompassed a
diversity of habitats from sea level to the highest elevations of the
Tuxtlas, over a much larger area. It now appears that Ross’ findings,
particularly among the Lycaenidae, were conservative. After a five year
survey of butterflies and skippers conducted in the vicinity of Lago
Catemaco, G. Busby and his colleagues found that Ross’ estimates of
butterfly species-richness in the Tuxtlas may represent less than 77% of
the actual butterfly fauna (Welling 1982, 1983, G. Busby, unpub. data).
The 13 regional records included among our 254 butterfly species further
indicate that Ross’ totals underestimate the present species-richness of
the Tuxtlas butterfly fauna.

SEASONALITY

A cumulative plot of the butterfly species collected at Laguna Encantada
is given in Figure 2a. Upon closer inspection, the species tally levels off
slightly after the first ten days, then rises sharply again during the first
week in August, 1985 (see arrow). A double reciprocal plot of S and N
(Figure 2b) reveals what appear to be two separate functions intrinsic to
this plot, visualized more clearly when plotted independently (Figures 2c
and 2d). These functions may represent the temporal overlap of two
distinct seasonal faunas. A similar pattern appears in a cumulative plot

29(1-2):105-133, 1990(91)

111

S = # species

Figure 2a. Cumulative species collected, Laguna Encantada, Veracruz, 1985.

1/S = 1/# species

1/N = 1 /Collector hours

Figure 2b. Reciprocal Plots of species vs. collector-hours, Laguna Encantada, 1 985.

112

J. Res. Lepid.

0.03

0.02

1/S = 1/# species

0.01

0.00

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70

1/N = 1/Collector hours

Figure 2c. Reciprocal Plot 21 - 29 July 1985, Laguna Encantada.

1/N = 1 /Collector hours

Figure 2d. Reciprocal Plot 30 July - 31 Aug. 1985, Lag. Encantada.

29(1-2):105-133, 1990(91)

113

Figure 3. Cumulative species collected, EBITROLOTU, Veracruz, 1985-1987.

of butterflies collected at EBITROLOTU (see Figure 3), where the
sharpest increase in number of species collected occurred between 15 and
22. iv. 1985 (see arrow).

Comparison of Study Sites

According to the theoretical species total (Se - 234) calculated for
Laguna Encantada, we sampled 78% of its butterfly fauna during a six
week period. Similarly, our species total at EBITROLOTU (212) repre-
sents about 72% of the theoretical species-richness calculated for that
site (Se = 295). The ratios of collector hours to hectares (Laguna
Encantada = 1.125, EBITROLOTU = 0.484), which we use as a rough
index of sampling efficiency, emphasize the great size disparity between
our two study sites.

The butterfly species that presently constitute the faunas of our
respective study sites are markedly dissimilar. Only about 70% of the
butterfly species collected at Laguna Encantada were also a subset of the
212 species sample collected at EBITROLOTU (see Appendix 3). The
number of species associated with disturbed habitats at both Laguna
Encantada and EBITROLOTU is given as a percentage of the total
number of species collected at each respective study site. The outcome

114

J. Res. Lepid.

Table 1 . Comparison of butterfly species composition between two rainforest
study sites, Veracruz, Mexico.

Study site

Laguna Encantada

EBITROLOTU

Collector-hours (N)

63 hrs

337.5 hrs

Area

56 ha

700 ha

Sampling effort (hrs/ha)

1.125

0.484

Total species Papilionoidea

collected and observed (S)

146

212

Number of species characteristic of
disturbed habitats* (pasture and milpa ) 31

27

Percentage of total for each study site 21 .23%

12.74%

I. 2 x 2 G-test of Independence G = 4.52 P < 0.05

II. Data transformed to standardize sampling effort

2x2 G-test of Independence G = 2.76 P < 0.1 0

* sensu Ross 1 975-77

of a 2 x 2 G-test of independence illustrates that the butterfly fauna of
Laguna Encantada includes a significantly higher percentage of weedy,
cosmopolitan species (21.23%) than that of the more extensive forest at
EBITROLOTU (12.74%, G = 4.52, see Table 1). When we repeat this
comparison after standardizing the weedy species data by multiplication
with the collector hours/hectare ratios, the differences between the
butterfly faunas of these two rainforest sites are still significant to the >
90% level (G - 2.76, see Table 1).

Discussion

SPECIES RICHNESS AND NEW RECORDS
Some patterns in butterfly species-richness emerge from a comparison
of recent faunal surveys conducted in tropical Mexico (see Table 2). An
increase in species-richness accompanies the transition from a Nearctic
to a Neotropical butterfly fauna as elevation decreases from Jalapa,
Veracruz southeast to the Gulf Coast lowlands (Beutelspacher 1975a,
Llorente, et al. 1986). All authors attribute high species-richness in the
Tuxtlas to the combination of habitats, vegetational types, geographic
isolation and climatic stability that characterize the region (Beutelspacher
1975a, Ross 1975-77, J. de la Maza and R. de la Maza 1985a, 1985b,
Llorente, et al. 1986). Our findings, when combined with those of Ross
(1975-77) and Busby (unpublished data) indicate a total of roughly 719
butterfly and skipper species for the Tuxtlas Region. In southern Mexico,
this total dwarfs that of neighboring Tabasco (C. Routledge 1977) and is

29(1-2):105-133, 1990(91)

115

Table 2. Comparison of Lepidoptera surveys, Mexico and Costa Rica.

Author Year,

Locality,

elev., duration, area Total

Ross 1975-77, 359

Tuxtlas, Veracruz,

0-1700 m, 15 mon., NA

C. Routledge 1977, 141

Tabasco,

0-900 m, 14 mon., 532,656 ha

Raguso and Llorente 1990, 182

a. Laguna Encantada, Veracruz,

350 m, 1 .5 mon., 56 ha

b. EBITROLOTU, Veracruz, 212

170-350 m, 27 mon., 700 ha

Llorente, et at. 1986, 3331 2 3

Teocelo/Jalapa, Veracruz,

600-1350 m,< 6 yrs.,NA

Beutelspacher 1975, 127

Las Minas, Veracruz,

1500 m, 16 mon., NA

Beutelspacher 1981, 150

Chamela, Jalisco,

0-500 m, 12 mon., 4.44 ha

Beutelspacher 1982, 174

El Chorreadero, Chiapas,

650 m, > 24 mon., NA

J. &R. delaMaza 1985, 544

Boca del Chajul, Chiapas,

150 m, 4 yrs., 800 ha

DeVries 1983, 1467

Costa Rica,

0-3500m, NA, NA.

Number of species by family1

Hes.

Pap.

Pie.

Lyc.

Rio.

Nym.

—

21

36

88

48

166

—

15

23

7

8

88

36

9

16

20

17

84

—

14

19

39

27

113

—

20

36

66

49

162

33

9

24

12

4

45

45

14

22

8

10

52

36

11

25

13

14

78

148

26

31

64

76

199

353

40

71

275s

3004

428

1 Key to abbreviations: Hes. = Hesperioidea, Pap. = Papilionidae, Pie. = Pieridae,
Lyc. = Lycaenidae, Rio. = Riodinidae, Nym. = Nymphalidae., NA= not available

2 plus 20 unidentified species

3 R. Robbins pers. comm.

4 P. DeVries pers. comm.

116

J. Res. Lepid.

exceeded only by the butterfly fauna of Chiapas (see J. de la Maza and R.
de la Maza 1985a, 1985b, Beutelspacher 1982).

In the 25 years since Ross’ study took place, four new butterfly species,
Adelpha leucerioides , A. diazi (Beutelspacher 1975b), A. milleri
(Beutelspacher 1976) and Mesosemia gemina (J. de la Maza and R. de la
Maza 1980), and two new subspecies, Prepona brooksiana escalantiana
(Descimon, Mast de Maeght and Stoffel 1974) and Dismorphia eunoe
popoluca (Llorente and Luis 1988) have been described from the Tuxtlas.
Numerous species records have appeared in the works of Freeman ( 1966,
1969b), Welling (1982, 1983), Busby (unpub. data) and this paper.
Recent censuses of hairstreak butterflies (Lycaenidae) in the Tuxtlas
have accounted for nearly twice as many species as Ross had collected (R.
Robbins, G. Busby, pers. comm.). If butterfly species-richness in the
Tuxtlas has increased, what are the causes?

Let us examine the traits of the 13 species newly observed in the
Tuxtlas. Five of these species are members of the Riodinidae; a family of
butterflies whose life histories are poorly understood. Many adult
riodinids perch on the undersides of leaves and are crepuscular in nature
(Ross 1975-77). These habits may have allowed riodinids to remain
undetected during previous butterfly censuses.

Butterfly species that fly high above the forest floor, avoid flowers, or
are difficult to identify may also escape notice. Memphis neidhoeferi and
M. xenocles are elusive charaxines that fly in primary and secondary
rainforest canopies, descending to feed on rotting fruit, feces and carrion
(DeVries 1987). The use of Van Someren-Rydon traps, the most reliable
method for capturing Memphis butterflies, was not widespread during
the years of Ross’ study (1962—65). Many of the metallic blue-colored
Memphis species are also variable and difficult to identify (Comstock
1961, DeVries 1987). Similarly, females of the swallowtail butterfly
Parides lycimenes are easily confused with those of Parides iphidamas,
P. erithalion and P. sesostris. The satyrine Cissia renata flies in bright
sunlight in all forest habitats and also visits animal dung and rotting
fruit (DeVries 1987). Llorente and Luis (1989) have collected increased
numbers of riodinid, charaxine and satyrine species by using Van
Someren-Rydon traps and by collecting in early morning and late
evening. The butterfly species discussed above were probably omitted
from previous surveys due to incomplete sampling or misidentification.

The observed increase in lycaenid species-richness may reflect a
combination of habitat disturbance and sampling efficiency. Ross col-
lected most of his 83 species of hairstreak butterflies near Catemaco at
flowers of Cordia spinescens , C. alliodora, Crotalaria uitellina and
Bidens pilosa ; plant species indicative of forest margins and pastures.
Most of Busby’s 150 hairstreak species were collected at blossoms of
Cordia (unpublished data). The availability of Bidens , Cordia and
Crotalaria nectar resources, which appear to increase in direct propor-
tion to the development of forest land, may lure lycaenids from the forest

29(1-2):105-133, 1990(91)

117

canopy down to ground level, and thus facilitate their capture. This
phenomenon may account for our record of Zizula tulliola at
EBITROLOTU.

A similar combination of disturbance and increased sampling effi-
ciency could explain the current abundance of Opsiphanes cassina
fabricii and O. tamarindi sikyon; two brassolines not observed during
Ross’ study. These species utilize plants of the Musaceae, Arecaceae and
Heliconiaceae as hosts and exploit the coconut palms and banana
plantations which accompany urban development and agriculture. Coco-
nut and banana are not new to the Tuxtlas, and Opsiphanes cassina and
O. tamarindi have probably inhabited local secondary forests with native
Heliconia and palms for centuries. However, these butterflies are now
regularly seen flying at dusk through the streets of Catemaco and San
Andres Tuxtla, and it is likely that agriculture and urbanization have
contributed to their abundance and detection.

Our three remaining species records for the Tuxtlas are butterflies that
were probably overlooked because they are rare. Eurytides marchandi
is a distinctively golden-colored swallowtail that flies near rivers and
forest edges. It may be a recent introduction from the southeast, where
it is known from lowland Chiapas ( J. de la Maza and R. de la Maza 1985a),
or it may simply be scarce in the Tuxtlas.

Megisto rubricata is a grass-feeding satyrine which is generally found
in shady oak-pine forests and arid canyons from Texas to Guatemala
(Scott 1986). We collected M. rubricata at EBITROLOTU from February
to October in 1985 and 1986. Since the only substantial oak-pine forests
in the Tuxtlas lie roughly 40 km southeast of EBITROLOTU on the
southern slopes of the Santa Marta volcano (Ross 1975-77), the habitat
requirements for M. rubricata may not be as strict as was previously
thought.

The presence of Agathymus rethon at Laguna Encantada is an enigma.
A. rethon has been collected at 1200 m in Puebla, Morelos, Guerrero and
the Oaxaca/Chiapas border. Like other megathymids, it is closely wed to
its foodplant, Agave sisalana (Stallings and Turner 1957, Freeman
1969a). In early August 1985, we observed two adults perched on a small,
dark green-leaved Agave on the northern lip of Laguna Encantada’s
crater. Gomez-Pompa (1973) identified Agave species from three locali-
ties in Veracruz - the high pine forests of Jalapa, the arid hills of Perote,
near Puebla, and the desert bordering Hidalgo - but made no mention of
Agave in the Tuxtlas. Likewise, Agave seeds were not discussed in
studies of seed dispersal by birds (Trejo-Perez 1976, Van Dorp 1985) or
bats (Orozco-Segovia, et al. 1985) conducted at EBITROLOTU. Never-
theless, Ross (1975-77) found Agave sp. to be locally abundant in elfin
woodland on the exposed upper ridges of the San Martin and Santa
Marta volcanoes in the Tuxtlas. We think that the Agave /Agathymus
rethon association at Laguna Encantada is probably one of a few scat-
tered relicts of a drier period in the Tuxtlas’ history.

118

J. Res. Lepid.

In summary, we attribute the apparent increase in butterfly species-
richness in the Tuxtlas to the following factors:

1. improved collecting methods

2. increased knowledge of species’ life histories

3. the effects of human disturbance on local habitats and vegetation

4. the gradual detection of rare species as a function of cumulative
sampling effort.

SEASONALITY

Mid to long-term faunal surveys such as Ross’work in the Tuxtlas (15
months) and our study at EBITROLOTU (27 months) bear the important
feature of having sampled the butterflies of those sites during at least one
entire year. Shorter-term censuses, on the scale of our study at Laguna
Encantada (6 weeks), may highlight seasonal fluctuations in butterfly
species abundance. Shapiro (1975) and Hill (1988) have discussed the
importance of temporal distribution and seasonality to the measurement
of butterfly species-richness. Studies of the butterfly faunas of tropical
rainforests in Panama (Emmel and Leek 1970) and Queensland, Austra-
lia (Hill 1988) indicate that butterfly seasonal abundance is a complex
phenomenon linked to environmental factors such as precipitation. Fox,
et al. ( 1965) observed peak butterfly abundance in Liberia during the dry
season, while Owen (1971) found butterflies in adjacent rainforests in
Sierra Leone to be most abundant during the rainy season.

In their analysis of the butterfly fauna of southern Chiapas, J. de la
Maza and R. de la Maza (1985b) described two peaks of adult butterfly
flight activity during the Mexican summer; the first occurring toward the
end of the dry season in April and the second during the sunniest segment
of the wet season in August and September. These authors predicted that
such patterns would be consistent throughout the rainforests of the
Mexican Gulf Coast.

We observed an increase in butterfly species at Laguna Encantada in
early August 1985 that corresponded to reduced precipitation during this
period (personal observation). While it is important not to extrapolate
meteorological data overzealously across the Tuxtlas, a record of precipi-
tation collected at EBITROLOTU (C. Field, unpublished data, see Figure
4) illustrates reduced rainfall during August 1985. By the end of August,
the number of new butterfly species encountered at Laguna Encantada
had diminished to nearly zero.

Seasonal patterns were also evident at EBITROLOTU. The sharp
increase in butterfly species encountered in mid-April 1985 also corre-
sponded to reduced precipitation (see Figure 4). We did not, however,
observe an additional species increase at EBITROLOTU during the first
week of August 1985. The species that first appeared during this time
at Laguna Encantada (eg. Parides erithalion polyzelus, Colobura dirce,
Epiphile adrasta ) are multiple-brooded and had already been collected at
EBITROLOTU during the spring months. Using Clench’s method,

29(1-2):105-133, 1990(91)

119

Month

Figure 4. Rainfall Data, EBITROLOTU, Veracruz, 1 985 (C. Field, unpublished data).

seasonal effects are visible only during the early stages of a cumulative
species census. The use of mark-release-recapture studies (Ehrlich and
Davidson 1960), malaise traps (Covell and Freytag 1979), and the
observational methods of Pollard (1977) and Hill (1988) are better suited
to the specific study of butterfly seasonality.

COMPARISON OF STUDY SITES

At Laguna Encantada and EBITROLOTU, we have sampled what
appear to be distinctly different butterfly faunas. At Laguna Encantada
we encountered a significantly higher percentage of generalist butterfly
species than that found at the larger, less disturbed EBITROLOTU site.
In addition, at least 13 species putatively associated with lower montane
rainforest were present at EBITROLOTU but absent at Laguna Encantada
(see Appendix 3). Do our results highlight the effects of habitat distur-
bance, the artifacts of imperfect experimental design or simply an
unforseen dissimilarity between rainforest microhabitats in the Tuxtlas?

Many workers have addressed the effects of habitat disturbance on
local resource availability and biological diversity (May 1973, 1981;
Connell 1978; see Denslow 1985 for review). Although our discussion has
stressed disturbance brought about by human activities, natural events
such as hurricanes, fires and treefalls may cause ecological disturbance
of comparable magnitude. Depending on its historical scale and fre-
quency, habitat disturbance may elicit a variety of responses from the
members of a given Neotropical butterfly community.

According to Blau (1980), who studied populations of Papilio polyxenes
in Costa Rica, there is a large assemblage of Neotropical insect species
adapted for exploiting habitats produced by localized disturbances. In
his ecological analysis of the moth fauna of Costa Rica’s Santa Rosa
region, Janzen (1988a) extends this discussion to the spatial and tempo-

120

J. Res. Lepid.

ral variation in land use and ecological disturbance near Guanacaste
National Park. Janzen (1988b) notes that the resulting mosaic of habi-
tats and successional stages of vegetation presently accomodates more
species of Lepidoptera than a pristine dry tropical forest, past or future,
could realistically support.

There are important differences between cosmopolitan species that can
exploit disturbed forest habitats and organisms that are adapted to
undisturbed primary forests. We have discussed the prevalence of
weedy, cosmopolitan species at the disturbed Laguna Encantada site and
note that a similar pattern was observed by Welling (1966) for Papilio ,
Euptoieta, Zerene and Phoebis species in patches cut into dense thorn
forests of the Yucatan Peninsula. Many of these butterflies are migra-
tory habitat generalists with catholic hostplant requirements and broad
distributions throughout the tropical Americas, and could rapidly invade
disturbed rainforest patches in the Tuxtlas. For example, at Laguna
Encantada we frequently observed Phoebis philea and Anteos clorinde
nivifera flying at the forest’s borders and over its canopy and ovipositing
on Cassia trees in light gaps within the forest. If the disturbances that
promote the recruitment of Cassia trees in rainforest gaps are frequent,
we believe that these pierid species, which range from Texas to Argen-
tina, will persist at Laguna Encantada.

Have we censused the butterfly faunas of Laguna Encantada and
EBITROLOTU thoroughly enough to legitimize the comparisons pre-
sented in this paper? As discussed by R. Routledge (1980) and Pielou
(1960, see Peet 1974 for review), it is difficult to remove sampling bias
from experimental measurements of diversity in large communities.
Despite the limitations discussed by Clench (1979, especially regarding
K, the collectability constant.), we chose the enzymatic model to analyze
the results of our surveys because of its applicability to our collecting
methods and its emphasis on sampling effort.

Although we sampled roughly 78% of the butterfly fauna of Laguna
Encantada in six weeks, we calculate from our model that an additional
4000 collector-hours would have been required to account for its complete
fauna. By this reasoning, many more collector hours would have been
necessary to fully sample the butterfly fauna of the larger EBITROLOTU
site. Sampling effort on this scale is only practical for long-term studies
conducted by numerous researchers. Although we spent five times as
many collecting hours at EBITROLOTU than at Laguna Encantada, our
sampling at the former site was slightly less thorough. This result and
the ratios of collector hours/hectare illustrate the great size disparity
between the two study sites and the difficulties inherent to biotic surveys
of different-sized habitats.

It is not clear whether the distinctive differences between the butterfly
faunas of EBITROLOTU and Laguna Encantada are a consequence of
human disturbance or simply a reflection of intrinsic differences between
these two sites. Were primary forest specialists present at EBITROLOTU

29(1-2):105-133, 1990(91)

121

(e.g. Nessaea aglaura) driven to local extinction by disturbance at
Laguna Encantada, or were they never there in the first place? It will be
important to return to Laguna Encantada and look for these potential
indicator species. The presence of older Cecropia , Ceiba and Ficus trees
at Laguna Encantada suggests that disturbance has played a historical
role in the ecology of that site. On the other hand, differences in altitude,
precipitation and topography may render the rainforests at Laguna
Encantada and EBITROLOTU more dissimilar than they appear. An
experiment such as Brazil’s “Minimum Critical Area” project, in which
forest patches of different sizes were cut from continuous lowland
rainforest, would be better suited to address these questions.

Rainforests throughout the Tuxtlas Mountains of Veracruz, Mexico are
becoming highly fragmented, and our study sites, particularly Laguna
Encantada, are likely to become even more disturbed and isolated.
Which members of a fragmented forest’s butterfly fauna will be sus-
tained if external sources of immigrants have been eliminated? In Costa
Rica, the protected dry forests of Guanacaste National Park will become
more homogeneous and pristine with succession. Janzen (1988b, also see
Gilbert 1980) predicts the disappearance of many species of Lepidoptera
in these forests, because potential sources of species immigration exter-
nal to the park were destroyed long ago and thus will be unable to counter
local extinction. In time, the extensive faunistic data bases from
Guanacaste should facilitate the testing of that hypothesis. Unlike
Guanacaste, the Tuxtlas region of Veracruz, Mexico will continue to
suffer human disturbance and cannot presently be said to have reached
its zenith in butterfly species-richness. We simply don’t understand the
history of human occupation and disturbance in the Tuxtlas well enough
to predict the fates of its forests and their faunas. It will be important to
update surveys such as ours periodically, paying particular attention to
species introductions and disappearances and the state of proximal
forests as potential sources of immigrants.

Acknowledgements. Field work at Laguna Encantada was conducted with the
permission of the landowner, Sr. Luis Vila of San Andres Tuxtla. Dr. R. Dirzo,
director of EBITROLOTU made available his insect collection and observations.
Additional specimens were provided by A. Ibarra. Access to museum collections
was granted by Drs. C. Remington, F. Rindge, J. Donahue, N. Penny and S.
Stockwell. Determination of some taxa collected at Laguna Encantada was
provided by the following: Riodinidae — D. Harvey; Taygetis — L. Miller; Theclinae
— R. Robbins; Satyrinae — M. Singer and Ithomiinae, Adelpha — the late G. Small.

Andre Chapp, Jorge Herrera, Gonzalo Quino and especially C. Horvitz and D.
Schemske assisted with field work. C. Boggs, D. Clark, R. Dial, M. Ellis, K. Freas,
D. Furth, M. Gallardo, D. Martinez, S. Naeem, B. Rathcke, W. Wagner, W. Watt,
S. Weiss, and two anonymous reviewers provided helpful advice and comments. G.
Busby and C. Field generously furnished unpublished data.

Support for travel and field expenses was provided by NSF grant BSR-841566 to
C. Horvitz and D. Schemske and a CONACYT D- 112-903646 (Mexico) grant and
DGAPA-UNAM IN201789 to J. Llorente.

122

J. Res. Lepid.

Literature Cited

Alvarez, J. A. S. 1982. Caida de hojarasca en una selva humeda tropical de Los
Tuxtlas, Veracruz, pp. 7-13. Thesis. U.N.A.M. Mexico.

Beutelspacher, C. 1975a. Notas sobre el suborden Rhopalocera de Las Minas,
Veracruz. Rev. Soc. Mex. Lep. 1: pp. 11-20.

. 1975b. Dos nuevas especies del genero Adelpha Hubner (Nymphalidae) de

Mexico. Rev. Soc. Mex. Lep. 1: 31-37.

. 1976. Estudios sobre el genero Adelpha Hubner (Nymphalidae) en Mexico.

Rev. Soc. Mex. Lep. 2: 8-14.

. 1981. Lepidopteros de Chamela, Jalisco, Mexico I. Rhopalocera. Anales Inst.

Biol. U.N.A.M. Ser. Zool. 52: 371-388.

. 1982. Mariposas diurnas de “El Chorreadero”, Chiapas. Anales Inst. Biol.

U.N.A.M. Ser. Zool. 53: 341-366.

Blau, W. S. 1981. The effects of environmental disturbance on a tropical butterfly
population. Ecology 61: 1005-1012.

Clench, H. K. 1979. How to make regional lists of butterflies: some thoughts. Jour.
Lep. Soc. 33: 215-231.

Comstock, W. P. 1961. Butterflies of the American tropics: the genus Anaea.

American Mus. of Natural History, New York, NY.

Connell, J. H. 1978. Diversity in tropical rainforests and coral reefs. Science 199:
1302-1309.

Covell, Charles V. Jr. and Paul H. Freytag. 1979. The malaise trap as a means
of sampling butterfly populations in Kentucky. Jour. Lep. Soc. 33: 153-161.
de la Cruz, M. and R. Dirzo. 1988. A survey of the standing levels of herbivory in
seedlings from a Mexican rainforest. Biotropica 19: 98-106.
de la Maza, J. and R. de la Maza, 1980. Rev. Soc. Mex. Lep. 6: 12-14.

. 1985a. La fauna de mariposas de “Boca de Chajul”, Chiapas, Mexico. Parte

I. Rev. Soc. Mex. Lep. 9: 23-44.

. 1985b. La fauna de mariposas de “Boca de Chajul”, Chiapas, Mexico. Parte

II. Rev. Soc. Mex. Lep. 10: 1-24.

Denslow, J. S. 1985. Disturbance-mediated coexistence of species, in S. T. A. Picket
and P. S. White eds. “The ecology of natural disturbance and patch dynamics.
The Academic Press, Inc. Orlando, FL.

Descimon, H. J. Maest de Maeght, J. R. Stoffel. 1974. Contribution a 1’etude des
nymphalides neotropicales; description de trois nouveau xPrepona mexicain.
Alexanor 8: 235-240.

DeVries, P. J. 1983. Checklist of Costa Rican butterflies, pp. 659-678 in D. H.
Janzen, ed± “Costa Rican natural history.” University of Chicago Press,
Chicago, IL.

. 1987. “The butterflies of Costa Rica.” Princeton University Press, Princeton,

NJ.

Ehrlich, P. R. and S. E. Davidson. 1960. Techniques for capture- recapture studies
of Lepidoptera populations Jour. Lep. Soc. 14: 227-229.

Emmel, T. C. and C. F. Leck. 1970. Seasonal changes in organization of tropical
rainforest butterfly populations in Panama. J. Res. Lep. 8: 133-152.

Evans, W. H. 1951. A catalogue of the American Hesperiidae in the British
Museum. Part I. Pyrrhopyginae. British Museum, London.

. 1952. A catalogue of the American Hesperiidae in the British Museum. Part

II. Pyrginae section I. British Museum, London.

29(1~2):105-133, 1990(91)

123

1953. A catalogue of the American Hesperiidae in the British Museum. Part

III. Pyrginae section II. British Museum, London.

. 1955. A catalogue of the American Hesperiidae in the British Museum. Part

IV. Hesperiinae and Megathyminae. British Museum, London.

Fox, R. M., A. W. Lindsey, H. K. Clench and L. D. Miller. 1965. The butterflies of
Liberia. Mem. Amer. Ent. Soc. Number 19.

Franklin, J. F. 1988. Structural and functional diversity in temperate forests, pp.
166-175. in E. O. Wilson, ed. “National forum on biodiversity”. National
Academy Press. U.S.A.

Freeman, H. A. 1966. New Hesperiidae records for Mexico. J. Lep. Soc. 20: 226-228.
— — . 1969a. A systematic review of the Megathymidae. Jour. Lep Soc. 23,
Supplement 1.

. 1969b. Records, new species and a new genus of Hesperiidae from Mexico.

Jour. Lep. Soc. 23, Supplement 2.

Gilbert L. E. 1980 Food web organization and conservation of neotropical
diversity, pp. 11-34. in M. Soule and B. Wilcox, eds. “Conservation biology: an
evolutionary approach.” Sinauer Associates, Inc. Sunderland, MA.

Godman, F. C. and O. Salvin. 1879-1900. Biologia centrali-americana. Insecta.

Lepidoptera: Rhopalocera. Three Vols. London.

Gomez-Pompa, A. 1973. Ecology of the vegetation of Veracruz, pp. 73-148 in A.
Graham, ed. “Vegetation and vegetational history in Northern Latin America.”
Elsevier, Amsterdam, Netherlands.

Guevara-Sada, S. and A. Gomez-Pompa. 1976. Determinacion del contenido de
semillas en muestras de suelo superficial de una selva tropical de Veracruz,
Mexico, pp. 203-232 in A. Gomez- Pompa et al., eds. “Investigaciones sobre la
regeneracion de selvas altas en Veracruz, Mexico. I.” Cortipania Editorial
Continental S.A., Mexico.

Hill, C. S. 1988. The temporal distribution of butterfly species richness in a
subtropical region of Australia. Proc. Ecol. Soc. Australia. 15: 115-126.
Hoffmann, C. C. 1940. Catalogo sistematico y zoogeografico de los lepidopteros
mexicanos. Pt.I. Papilionoidea. Anales Inst. Biol., U.N.A.M. 11(2): 639-739.
Horvitz, C. C. and A. J. Beattie. 1980. Ant dispersal of Calathea (Marantaceae)
seeds by carnivorous ponerines (Formicidae) in a tropical rainforest. Am. Jour.
Bot. 67: 321-326.

Horvitz, C. C. and D. W. Schemske. 1988. A test of the pollinator-limitation
hypothesis for a neotropical herb. Ecology 69: 200-206.

Janzen, D. H. 1984. Two ways to be a tropical big moth. Santa Rosa saturniids and
sphingids. Oxford Surveys in Evolutionary Biology 1: 85-140.

. 1988a. Tropical dry forests, the most endangered major tropical ecosystem.

pp. 130-137. in E. O. Wilson, ed. “National forum on biodiversity.” National
Academy Press. Washington, DC.

. 1988b. Ecological characterization of a Costa Rican dry forest caterpillar

fauna. Biotropica 20: 120-138.

Jenkins, D. W. 1983. Neotropical Nymphalidae I. Revision of Hamadryas. Bulletin
of the Allyn Museum 81: 1-146.

. 1984. Neotropical Nymphalidae II. Revision of Myscelia. Bull. Allyn Mus. 87:

1-64.

. 1985. Neotropical Nymphalidae III. Revision of Catonephele. Bull. Allyn

Mus. 92: 1-65.

124

J. Res. Lepid.

Lamas, Gerardo. 1987. The identity of Mechanitis doryssus Bates and M. saturata
Godman (Lepidoptera: Nymphalidae; Ithomiinae). Rev. Per. Ent. 30: 48-50.

Llorente, J. and A. Luis. 1988. Nuevos Dismorphiini de Mexico y Guatemala
(Lepidoptera: Pieridae). Folia Entomologica Mexicana 74: 159-178.

. 1989. Mariposas en el Valle de Mexico: Introduccion e Historia. 1. Distribucion

local y estacional de los Papilionoidea de la Canada de los Dinamas, Magdalena
Contreral, D. F., Mexico. Folia Entomologica Mexicana 80: 188pp + 15 figs. + 5
tables, ms in press.

Llorente, J., A. Garces and A. Luis. 1986. El paisaje Teoceleno IV. Las Mariposas
de Jalapa-Teocelo, Veracruz. Teocelo 3: 14-37.

Lot-Helgueras, A. 1976. La estacion de biologia tropical, “Los Tuxtlas”: pasado,
presente y futuro. in A. Gomez-Pompa, et al., eds. “Investigaciones sobre la
regeneracion de selvas altas en Veracruz, Mexico. I.” Compania Editorial
Continental. S.A. Mexico.

May, R. 1973. “Stability and complexity in model ecosystems.” Princeton Univ.
Press. Princeton, NJ.

. 1981. Modelling recolonization by neotropical migrants in habitats with

changing patch structure, with notes on the age structure of populations, in R.
L. Burgess and D. M. Sharpe, eds. “Forest island dynamics in man-dominated
landscapes.” pp. 207- 213. Springer- Verlag, Inc. New York, NY.

Orozco-Segovia, A., M. Angel Arnella, N. A. Correa, and C. Vazquez Yanes. 1985.
Interacciones entre una poblacion de murcielagos de la especie Arribeus
jamaicensis y la vegetacion del area circundante la region de Los Tuxtlas,
Veracruz, pp. 365-378. in A. Gomez-Pompa and S. del Amo Rodriguez, eds.
“Investigaciones sobre la regeneracion de selvas altas en Veracruz. II.” Inst, de
Investigaciones sobre Recursos Bioticos. Editorial Alhambra Mexicana. S.A.
Mexico.

Owen, D. F. 1971. Tropical Butterflies: the ecology and behavior of butterflies in
the tropics with special reference to African species. Clarendon Press, Oxford.

Peet, R. K. 1974. The measurement of species diversity. Ann. Rev. Ecol. and Syst.
5: 285-307.

Pielou, E. C. 1960. The measurement of diversity in different types of biological
collections. J. Theoret. Bio. 13: 131-144.

Pollard, E. 1977. A method for assessing changes in abundance of butterflies. Biol.
Conserv. 12: 115-134.

Popma, J., Franz Bongers, M. Martinez-Ramos and E. Vaneklaas. 1988. Pioneer
species distribution in treefall gaps in neotropical rainforests. A gap definition
and its consequences. Jour. Trop. Ecol. 4: 77-88.

Ross, G. N. 1966. Life history studies on Mexican butterflies IV. Ann. Entom. Soc.
Amer. 59: 985-1003.

. 1975-1977. An ecological study of the butterflies of the Sierra de Tuxtla, in

Veracruz, Mexico. Jour. Res. Lep.14: 103-124, 169-188, 233-255; 15: 41-60, 109-
128, 185-200, 225-240; 16: 87-130.

Routledge, C. E. 1977. El suborden Rhopalocera (Lepidoptera) del estado de
Tabasco. Su lista, frecuencia, diversidad y distribucion. Rev. Soc. Mex. Lep. 3:
57-75.

Routledge, R. 1980. Bias in estimating the diversity of large uncensused
communities. Ecology 61: 276-281.

Scott, J. 1986. “The butterflies of North America.” Stanford Univ. Press. Stanford,
CA.

29(1-2):105-133, 1990(91)

125

Shapiro, A. M. 1975. The temporal component of butterfly species diversity, in M.
L. Cody and J. M. Diamond, eds. “Ecology and evolution of communities.” pp.
181-195. Harvard University Press, Cambridge, MA.

Singer, M. C., P. J. DeVries, and P. E. Ehrlich. 1983. The Cissia confusa species
group in Costa Rica and Trinidad. Zool. Jour. Linn. Soc. 79: 101-119.

Sokal, R. R. and F. J. Rohlf. 1981. Biometry. Second Edition. W. H. Freeman and
Co., San Francisco, CA.

Soto, E. M. 1976. Algunos aspectos climaticos de la region de Los Tuxtlas. in A.
Gomez-Pompa, et al. , eds. “Investigaciones sobre la regeneracion de selvas altas
en Veracruz, Mexico. I.” Compania Editorial Continental. S. A. Mexico.

Stallings, D. M. and J. R. Turner. 1957. A review of the Megathymidae of Mexico
with a synopsis of the classification of the family. Lepid. News 11: 113-125.

Trejo-Perez, L. 1976. Diseminacion de semillas por aves en “Los Tuxtlas”,
Veracruz, pp. 467-490. in A. Gomez-Pompa, et al., eds. “Investigaciones sobre
la regeneracion de selvas altas en Veracruz, Mexico. I.”. Compania Editorial
Continental. S. A. Mexico.

Van Dorp, D. 1985. Frugivoria y dispersion de semillas por aves. pp. 533-564. in A.
Gomez-Pompa and S. del Amo Rodriguez, eds. “Investigaciones sobre la
regeneracion de selvas altas en Veracruz, Mexico. II.” Inst, de Investigaciones
sobre Recursos Bioticos. Editorial Alhambra Mexicana. S. A. Mexico.

Welling, E. 1966. Devastation of Yucatecan forests with notes on insect
abundance and formation of local climates. J. Lep. Soc. 20: 201-205.

. 1982. “Season Summary”, Neotropical Region. News of the Lepidopterists’

Society. 1982 (4): 33.

-. 1983. “Season Summary”, Neotropical Region. News of the Lep. Soc. 1983 (4):

35.

126

J. Res. Lepid.

Appendix 1 .

Papilionidae

Pieridae

Nymphalidae

Lycaenidae

Riodinidae

Butterfly species characteristic of disturbed habitats, after Ross
(1977) and Robbins (pers. comm.).

* Papilio thoas

* Ascia monuste
Eurema daira
Eurema mexicana

* Eurema nicippe
Gluthophrissa drusilla

* Phoebis sennae

* Phoebis philea

* Phoebis argante
Pyrisitia boisduvaliana
Pyrisitia proterpia

* Pyrisitia lisa
Pyrisitia dina
Pyrisitia nise

* Zerene cesonia
Anaea aidea
Anartia fatima
Anartia jatrophae
Anthanassa ardys
Castilia myia
Castilia griseobasalis
Ghiosyne janais
Chiosyne lacinia

* Danaus gilippus
Dry as iulia

* Euptoieta hegesia
Hermeuptychia hermes

* Marpesia chiron
Phyciodes vesta
Thessalia theona
Calycopis isobeon
Eve res corny ntas
Hemiargus ceraunus
Hemiargus huntingtoni
Leptotes cassius striata
Rekoa palegon
Rekoa marius

t Strymon columella

f Tmolusazia

f Tmolus echion
Juditha molpe

* cited by Ross (1975-1977) or DeVries (1987) as being migratory.

f Robbins pers. comm.

29(1-2):105-133, 1990(91)

127

Appendix 2. A Checklist of the Butterflies of Laguna Encantada and La Estacion de
Biologia Tropical “Los Tuxtlas” (EBITROLOTU), LJ.N.A.M., Veracruz, Mexico.

Totals: 254 species, 1293 specimens.

R denotes taxon not found in Ross’ checklist 1975-77.

N denotes taxon not previously reported from Tuxtlas.

S denotes definitive sight record; all others are collected specimens.

PA denotes specimen collected at Playa Azui, near Catemaco.

Taxon

Study Sites: Laauna Encantada

20 vii - 31 viii 1985

HESPERIOIDEA

HESPERIIDAE

Pyrginae

Epargyreus exadeus (Cramer 1 779) 2

LJrbanus proteus (Linnaeus 1 758) 2

Urbanus dorantes (Stoll 1 790 2

LJrbanus procne (Plotz 1 880) 1

Astraptes anaphus (Cramer 1 777) 1

Astraptes fulgerator (Welch 1 775) 4

Autochton neis (Geyer 1 832) 3

Aethilla lavocrea Butler 1 872 1 PA

Achalarus toxeus (Plotz 1882) 1 PA

Cogia calchas (Herrich-Schaeffer 1 869) 2

Nisoniades rubescens(Mosch\er 1 876) 3

Nisoniades ephora (Herrich-Schaeffer 1 870) 1

Carrhenes canescens (R. Felder 1 869) 1

Xenophanes trixus (Stoll 1 780) 2

Achlyodes thraso (Hubner 1 807) 1

Achlyodes busirus (Stoll 1 782) 2

Timochares trifasciata (Hewitson 1 868) 2

Gesta gesta invisus (Butler & H. Druce 1 892) 1

Ephyriades brunneus floridalis (Bell & W. Comstock 1 948) 2
Pyrgus communis (Grote 1 872) 2

Pyrgus oileus (Linnaeus 1 767) 3

Heliopetes macaira (Reakirt 1866) 1 PA

Heliopetes arsalte (Linnaeus 1 758) 1 PA

Quadrus cerealis (Stoll 1 782) 3

Quadrus lugubris (R. Felder 1 869) 1

Nascus phocus (Cramer 1 777) 1

Theogenes lactifera (Butler & Druce 1 872) 1

Antigonus nearchus (Latreille 1 824) 1

Hesperiinae

Calpodes ethlius (Stoll 1 782) 1

Cymaenes odilia (Burmeister 1 878) 3

Ancylocypha arene (Edwards 1871) 1

Pompeius pompeius (Latreille 1 824) 3

Talides sergestus (Cramer 1 775) 1

Vettius fantasos anaca (Evans 1 955) 2

Cynea cynea (Hewitson 1 876) 1

Parphorus decorum (Herrich-Schaeffer 1 869) 1

Moerus striga stroma (Evans 1 955) 1

Anthoptes epictelus (Fabricius 1 793) 1

Papias subcostulata Integra (Mabille 1891) 1

EBITROLOTU

i 1985- Vi 1987

128

J. Res. Lepid.

Megathyminae

Agathymus rethon (Dyar 1913) 1 (+1 S) N

PAPILIONOIDEA

PAPIUONIDAE

Papilioninae

Parides photinus (Doubleday 1844) 6

Parides erithalion polyzelus (R. Felder 1 865) 3 5

Parides lycimenes septentrionalis (J. de la Maza

and Diaz 1978) - 8 R, N

Parides iphidamas (Fabricius 1793) - 7

Parides eurimedes mylotes (Bates 1 856) - 8

Parides sesostris zestos (Gray 1 852) - 2

Battus belus varus (Kollar 1 850) - 5

Papilio polyxenes asterias (Stoll 1 782) 1 S, PA

Papilio cresphontes (Cramer 1 777) - 1 R

Papilio thoas autocles (Rothschild & Jordan 1 906) 6 1

Papilio androgeus epidaurus (God man & Salvin 1890) IS 2

Pyrrhosticta victorious (Doubleday 1 844) - 2

Priamides anchisiades idaeus (Fabricius 1793) 2 7

Eurytides marchandi (Boisduval 1 836) - 3 R, N

Eurytides phaon (Boisduval 1 836) 1 PA 3

Eurytides belesis (Bates 1 834) 1 7

Eurytides branchus (Doubleday 1846) 1

Eurytides philolaus (Boisduval 1 846) 1

Eurytides epidaus (Doubleday 1846) 2

PIERIDAE

Dismorphiinae

Dismorphia amphiona praxinoe (Doubleday 1 844) 2 2

Dismorphia theucarilla fortunata (Lucas 1854) - 3

Pierinae

Ascia monuste (Linnaeus 1764) 9 1

Melete isandra (Linnaeus 1764) 2

Itaballia pisonis kicaha (Reakirt 1 863) - 5

Catasticta nimbice ochracea (Boisduval 1 836) - 1

Pieriballia viardi (Boisduval 1 836) - 5

Gluthophrissa drusilla poeyi (Butler 1 872) - 5

Coliadinae

Zerene cesonia (Stoll 1791) 2 1

Anteos clorinde nivifera (Fruhstorfer 1 907) 5

Phoebis sennae marcellina (Cramer 1 777) 10 3

Phoebis philea (Linnaeus 1763) 8 2

Phoebis argante (Fabricius 1775) 4 2

Phoebis agarithe (Boisduval 1 836) 2 1

Rhabdodryas trite (Linnaeus 1 758) - 3

Pyrisitia proterpia (Fabricius 1775) 5 2

Pyrisitia lisa (Boisduval & LeConte 1 833) - 3

Pyrisitia nise nelphe (R. Felder 1864) 8 4

Pyrisitia dina westwoodi (Boisduval 1836) 6

Eurema albula celata (R. Felder 1 869) 7 6

Eurema daira eugenia (Wallengren 1860) 6 4

Eurema xanthochlora (Kollar 1850) 1 6

Eurema nicippe (Cramer 1 780) 1

29(1-2): 105-133, 1990(91)

129

LYCAENIDAE

Theclinae

Pseudolycaena damo (Druce 1875) 2 5

“Theda” theocritus (Fabricius 1793) 1 PA, R 1

“Theda” hesperitis (Butler and Druce 1872) 1 0

“Theda” sp. aff. plusios Godman and Salvin 1887 - 2

“Theda” barajo (Reakirt 1866) 1 3

“Theda” tephraeus (Gey er 1837) 1

“Theda” haldones (Butler and Druce 1872) 1

“Theda” 7 unidentified species - totals

Rekoa marius (Lucas 1 857) 6

Rekoa meton (Cramer 1 782) IS 1

Eumaeus toxea (Hubner 1806) 2 PA 4

Brangas cocdneifrons (Godman & Salvin 1 887) - 1

Atlides polybe (Linnaeus 1 758) 1

Ministrymon arola Hewitson 1 868 2

Tmolus echion (Linnaeus 1 758) 2

Tmolus azia (Hewitson 1 873) 1 PA

Tmolus 3 unident, species. - total 3

Oenomaus ortygnus (Cramer 1 782) 1

Calycopis isobeon (Butler & Druce 1 872) 8 6

Calycopis 3 unident, species. - total 7

Cyanophrys 5 unident, species. - total 5

Cyanophrys miserabilis Clench 1 946 3

Cyanophrys herodotus (Fabricius 1793) 1

Arawacus sito (Boisduval 1836) 3 3

Arawacus togarna { Hewitson 1863) - 8

Panthiades bathis (Fabricius 1781) 1 1

Panthiades bitias (Cramer 1 777) - 1

Strymon sp. - 1

Strymon columella Reakirt 1 866 1 PA

Strymon rufofusca (Hewitson 1 877) 2 PA

Electrostrymon sp. aff. cyphara (Hewitson 1874) 3

Areas cypria (Geyer 1 837) - 1

Theritas sp. - 1

Chalybs sp. - 3

Polyommatinae

Hemiargus isola (Reakirt 1866) - 1

Hemiargus ceraunus zachaeina (Butler and Druce 1872) 2

Leptotes cassius striata (Edwards 1 887) 2 3

Everes comyntas texanus (R. Chermock 1944) 3 6

Celastrina ladon gozora (R. Chermock 1 944) - 1

Zizula tulliola (Godman & Salvin 1887) - 1 R, N

RIODINIDAE

Euselasiinae

Euselasia sergia (Godman & Salvin 1 885) - 6

Hades noctula (Westwood 1851) 1 8

Riodininae

Perophthalma tullius lasus Westwood 1851 - 6

Leucochimona vestalis (Bates 1 865) - 4

Leucochimona lepida nivalis (Godman & Salvin 1885) - 5

Mesosemia gemina J. de la Maza and R. de la Maza 1 980 - 2

Mesosemia telegone (Boisduval 1836) 5 R

Eurybia elvina (Stichel 1910) 3R 1

130

J. Res . Lepid.

Calospila sudias (Hewitson 1856) 5

Napaea umbra (Boisduval 1 870) - 1

Napaea euchariia picina Stichel 1910 - 3 R, N

Cremna thasus subrutila Stichel 1910 - 3 R, N

Charis velutina (Godman & Salvin 1878) - 6

Calephelis sp. 1 2 4

Calephelis sp. 2 2 -

Melanis pixe (Boisduval 1 836) 4 5

Charmona gynaea zama (Bates 1868) 7

Lasaia agesilas callaina Clench 1 972 1 S, R

Lasaia unident. sp. - 2

Mesene croceela Bates 1865 - 2

Symmachia tricolor hedemanni (R. Felder 1869) - 1

Sarota myrtea (Godman & Salvin 1 886) - 4

Sarota chrysus (Cramer 1 782) - 6 R, N

Calydna lusca venusta (Godman and Salvin 1886) 1

Calydna sturnula hegias (R. Felder 1869) 1 PA, R, N

Emesis mandana (Cramer 1 780) - 2

Emesis vulpina (Godman & Salvin 1886) - 2 R, N

Emesis lucinda saturata Godman and Salvin 1 886 9 R

Emesis lupina (Godman and Salvin 1886) 4

Thisbe lycorias (Hewitson 1853) 5 1 R

Lemonias agave (Godman & Salvin 1886) - 2

Juditha molpe (Hubner 1808) 8 6

Theope 2 unident. sp. - total 3

Theope virgilius Fabricius 1793 3 R

Theope eupoiis Schaus 1 890 1 R

Theope bacenis Schaus 1 890 1 R

Pandemos godmanii DeWitz 1 877 - 1 R

Menander menander purpurata Godman & Salvin 1878 - 5

NYMPHALIDAE

Heliconiinae

Agraulis vanillae incarnata (Riley 1 926) 3 3

Dione juno huascuma (Reakirt 1 866) 1 4

Dryas iulia moderata (Stichel 1 926) 5 6

E ueides isabella eva (Cramer 1775) 16 4

Eueides lineata (Salvin 1 868) 5 2

Eueides aliphera gracilis Stichel 1 903 - 1

Heliconius ismenius telchina (Doubleday 1 847) 5 1

Heliconius erato petiverana (Doubleday 1 847) 11 4

Heliconius hortense (Guerin-Meneville 1829) 1 1

Heliconius charitonius vazquezae

(Comstock & Brown 1 950) 5 3

Heliconius doris transiens (Staudinger 1 896) - 1

Philaethria dido diatonica (Fruhstorfer 1912) - 1

Dryadula phaetusa (Linnaeus 1 758) - 2

Nymphalinae

Hypanartia lethe (Fabricius 1793) 4

Siproeta epaphus (latreille 1882) 2 1

Siproeta stelenes biplagiata (Fruhstorfer 1 907) 4 2

Anartia fatima (Godart 1 793) 4 5

Anartia jatrophae (Godart 1 820) 4 1

Junonia evarete (Cramer 1 782) 4 1

29(1-2):105-133, 1990(91)

131

Argynninae

Euptoieta hegesia hoffmanni (Comstock 1 944) 2 4

Melitaeinae

Chlosyne janais (Drury 1 782) 1 PA 3

Chlosyne erodyle (Bates 1864) - 1

Chlosyne lacinia (Geyer 1 837) 5 1

Chlosyne hippodrome (Geyer 1 837) 1 (3 PA)

Thessalia theona (Menetries 4855) 4

Phyciodes vesta (W. H. Edwards 1869) 2

Anthanassa tulcis (Bates 1 864) 1 PA

Anthanassa ptolyca (Bates 1 864) 2 R

Eresia phillyra (Hewitson 1852) 2 2

Castilia myia (Hewitson 1 864) 2 7

Tegosa guatemalena (Bates 1 864) 3

Limenitidinae

Adelpha melanthe (Bates 1866) 4 3

Adelpha leuceria Druce 1 879 - 9

Adelpha celerio diademata (Fruhstorfer 1915) 4 R 5

Adelpha phylaca (Bates 1 860) - 2

Adelpha ixia leucas (Fruhstorfer 1 91 5) - 1R

Adelpha felderi (Boisduval 1870) - 1 R

Adelpha milled Beutelspacher 1 975 - 1 R

Adelpha diazi Beutelspacher 1 975 - 1 R

Adelpha leucerioides Beutelspacher 1 975 - 1

Adelpha iphiclus (Linnaeus 1758) 5

Adelpha sp. - 1

Myscelia ethusa (Doyere 1840) 4

Myscelia cyariiris (Doubleday 1848) 1 2

Dynamine postverta mexicana (d’ Almeida 1952) 9 1

Dynamine dyonis (Geyer 1 837) 3

Eunica alcmena Doubleday & Hewitson 1850 - 1

Eunica monima (Cramer 1 782) 1

Diaethria anna (Guerin-Meneville 1844) 1 4

Diaethria astala (Guerin-Meneville 1844) 2

Callicore lyca (Doubleday & Hewitson 1847) - 1

Callicore texa titania (Salvin 1 869) - 1

Nessaea aglaura (Doubleday 1848) - 5

Biblis hyped a aganisa (Boisduval 1 836) 5

Mestra amymone (Menetries 1 857) 3 PA

Hamadryas februa ferentina (Godart 1 824) 7 2

Hamadryas feronia farinulenta (Fruhstorfer 1916) 8

Hamadryas guatemalena marmarice (Fruhstorfer 1 91 6) 5

Hamadryas amphinome mexicana (Lucas 1 853) 7 4

Hamadryas iphthime (Bates 1864) - 1

Colobura dirce (Linnaeus 1 758) 5 3

Historis odius (Fabricius 1775) IS 1

Smyrna blomfildia datis (Fruhstorfer 1 908) 8 2

Coea acheronta (Fabricius 1775) - 4

Marpesia chiron marius (Cramer 1 780 ) 3 3

Marpesia petreus tethys (Fabricius 1777) IS 1

Marpesia harmonia (Doubleday & Hewitson 1847) - 8

Temenis laothoe hondurensis (Fruhstorfer 1907) 1 4

Epiphile adrasta { Hewitson 1861) 4 2

Nica flavilla canthara (Doubleday 1849) 3 1

Pyrrhogyra neaerea hypsenor (Godman & Salvin 1884) 1 1

132

J. Res. Lepid.

Pyrrhogyra otolais (Bates 1 864) 1 5

Catonephele mexicana (Jenkins & De la Maza 1 985) 5 9

Catonephele numilia esite (R. Felder 1869) - 4

Apaturinae

Doxocopa laure (Drury 1773) - 1

Charaxinae

Prepona omphale octavia Fruhstorfer 1 904 - 3

Archaeoprepona demophon centralis (Fruhstorfer 1905) 2 2

Archaeoprepona demophoon ssp. nov. 1 6

Archaeoprepona amphimachus amphiktion

(Fruhstorfer 1916) 3 2

Zaretis callidryas (Felder 1869) 2 1

Zaretis itys (Cramer 1 777) - 1

Anaea aidea (Guerin-Meneville 1844) 9

Consul fabius cecrops (Doubleday 1 849) 4 2

Consul electra (Westwood 1 850) 5 2

Memphis pithyusa (R. Felder 1869) - 15

Memphis morvus boisduvali (W. Comstock 1 961 ) 1 5

Memphis oenomaus (Boisduval 1 870) 1 6 R

Memphis neidhoeferi (Rotger 1965) - 4 R, N

Memphis forreri (God man and Salvin 1884) - 3 R

Memphis xenica (Bates 1864) - 5 R

Memphis proserpina (Salvin 1 869) - 4

Memphis xenocles (Westwood 1 850) - 1 R, N

Fountainea eurypile confusa (Hall 1 929) - 5

Fountainea ryphea (Cramer 1 775) - 1 R

Siderone marthesia (Cramer 1 777) - 1

Satyrinae

Pierella luna heracles (Boisduval 1870) 6 5

Taygetis andromeda inconspicua (Draudt 1 931 ) 5 2

Taygetis virgilia rufomarginata (Staudinger 1 888) 1

Pareuptychia ocirrhoe (Sulzer 1 779) 1 5

Pareuptychia metaleuca (Boisduval 1 870) 3 4

Cissia usitata pieria (Butler 1 866) 11 12

Cissia iabe (Butler 1 869) - 2

Cissia renata disaffecta (Butler 1 874) - 3 R, N

Cissia libye (Linnaeus 1 797) 2 6

Hermeuptychia hermes sosybius (Fabricius 1 793) 6 7

Euptychia westwoodi (Butler 1866) 4

Euptychia jesia (Cramer 1 869) 1

Megisto rubricata anabelae Miller 1 976 - 7 R, N

Brassolinae

Opsiphanes tamarindi sikyon Fruhstorfer 1912 2 PA, R 1

Opsiphanes cassina fabricii (Boisduval 1 870) 2 PA 6

Opsiphanes quiteria quirinus Godman and Salvin 1881 - 2 R

Caligo memnon (C. Felder & R. Felder 1865) 4 3

Caligo uranus (Herrich-Schaffer 1853) - 5

Morphinae

Morpho polyphemus luna (Butler 1 872) 1 S 3

Morpho peieides montezuma {G uenee 1859) 6 12

Acraeinae

Altinote ozomene nox (Bates 1 864) 1 4

29(1-2):105-133, 1990(91) 133

Danainae

Danaus plexippus (Linnaeus 1 758) IS 1

Danaus gilippus thersippus (Bates 1863) 5 6

Danaus erisimus montezuma (Talbot 1943) 1 PA

Lycorea cleobaea atergatis (Doubleday & Hewitson 1 847) 3 4

Ithomiinae

Tithorea harmonia hippothous (Godman and Salvin 1879) 1 PA 4

Melinaea lilis imitata (Bates 1 864) 3 4

Meehan itis polymnia lycidice (Bates 1 864) 6 10

Meehan itis menapis (Bates 1 864) 3 6

Oleria pauia (Weymer 1 883) 1 3

Aeria pacifica (Godman & Salvin 1879) 3 3

Dircenna klugii (Geyer 1 837) 1 6

Pteronymia cotytto (Guerin-Meneville 1 844) 3 14

Greta oto (Hewitson 1 855) 5 1

Greta nero (Hewitson 1855) - 5

Hyposcada virginiana (Hewitson 1 855) - 5

Napeogenes tolosa (Hewitson 1855) - 4

Ithomia patilla (Hewitson 1852) - 1

Hypoleria cassotis (Bates 1 864) 1

Appendix 3. Butterfly species associated with lower montane rainforest, found at
EBITROLOTLJ but absent at Laguna Encantada.

Papilionidae

Pieridae

Nymphalidae:

Parides iphidimas
Parides eurimedes

* Dismorphia theucarila

* Itaballia pisonis
Pieriballia viardi
Caligo uranus
Cissia labe
Greta nero
Heliconius doris
Ithomia patilla
Mechanitis lysimnia
Napeogenes tolosa

* Nessaea aglaura

Habitat specialist sensu DeVries 1987

Journal of Research on the Lepidoptera

29(1-2):134-142, 1990(91)

Territorial Hilltopping Behavior of Three
Swallowtail Butterflies (Lepidoptera, Papilionidae)
in Western Brazil.

Carlos E. G. Pinheiro

Departamento de Ecologia, Funda5ao Universidade de Brasilia, 70910 Brasilia DF, Brazil

Abstract. This paper investigates differences in the territorial
hilltopping behavior of Papilio thoas, Battus polydamas and Eury tides
orthosilaus. Seven behavioral activities of hill topping males are de-
scribed and measured, including inter and intraspecific aggression,
territory establishment, and choice of preferred sites for territories. P.
thoas and B. polydamas shared many behavioral characteristics and
preferred the same territorial site, near the highest point of the hill.

Most interspecific aggression occurred between these species. P. thoas ,
the dominant species, won aggressive encounters and established its
territories first. B. polydamas set up territories in neighboring areas
placed around the territories ofP. thoas. Marked behavioral differences
were found between E. orthosilaus and the other species; its territories
were established several meters above the largest trees in the area and
did not overlap with those of P. thoas and B. polydamas. Behavioral
activities in these butterflies occurred in a sequential pattern each day.
Aggressive acts predominated in the early hours and decreased as
territory owners become established and newly arriving butterflies
become rare.

Introduction

Many butterflies and other insects fly to hilltops in order to mate
(Shields, 1967; Baker, 1983). Hilltopping is in many aspects similar to lek
behavior shown by many birds and other vertebrate species. It is believed
to increase the mating efficiency of individuals of a species reducing the
area in which sexual encounters can occur (Poulton, 1904).

According to Alcock (1987) it is not obvious why hilltops would become
“rendezvous points” for any species. It has been noted that these sites
generally do not contain food resources for adult butterflies, nor are they
locations where females emerge, oviposit or even rest. There are only
males waiting to mate.

Hilltopping butterflies exhibit a considerable diversity with respect to
the way that males aggregate and form territories (Courtney & Ander-
son, 1986). In some species males form clusters of several individuals
with little or no aggressive behavior among them (see examples in
Alcock, 1985; Baughman et al., 1988). In the other extreme males do not
tolerate each other, and form mating territories as defined by Brown &
Orians (1970); such is the case for many swallowtail butterflies (Shields,
1967; Lederhouse, 1982; Rutowski et ah, 1989).

29(1-2):134-142, 1990(91)

135

Although much information has been obtained on hilltopping behavior,
it is still poorly documented which species exhibit such behavior, espe-
cially in the tropics. In this paper I examine some aspects of the territorial
hilltopping of males of three large swallowtail butterflies that use the
same place for hilltopping: Papilio thoas brasiliensis Roths. & Jord.,
1906, Battus polydamas polydamas (L., 1758) and Eury tides orthosilaus
(Weymer, 1899). I describe their preferences of sites for territories and
use a behavioral method to estimate and compare the frequencies of
different kinds of behaviors associated with hilltopping, including inter
and intraspecific encounters and other behaviors associated with terri-
tory establishment. The order of occurrence of determined behavioral
activities of males is also reported.

Study Site and Methods

Field work was done in the region of the Rio Manso, Mato Grosso State,
western Brazil (55°50'W, 14°52'S) on sunny days between May 16-21 and
October 13-16, 1988; and April 22-26, 1989. Observations and data
collecting were concentrated on the top of the highest hill in the area (405
m elevation), composed of a dome measuring 120 x 40 m covered with a
combination of cerrado vegetation, grasses and bare rocks. F orest vegeta-
tion is found 150 m below at the bottom of two small valleys around the

Table 1 . Behavioral categories performed by males of P. thoas, B. polydamas
and E. orthosilaus during territorial hilltopping and their characteristcs (see also

the text).

Category

Species

Characteristics

Intraspecific Aggression

All

Includes chasing, grappling in
the air and falling to the ground

Interspecific Aggression

All

Only chasing, usually faster than
intraspecific aggression

Extensive Patrolling

All

An exploratory flight over a large
area of the hilltop

Restricted Patrolling

All

Flying about inside a small and
well defined area

Perching All but E. orthosilaus

Resting on some vantage point
the male darts out at every
passing butterfly

Hovering

All

Hovering or gliding in place

Flying in circles

Only P. thoas

Flying in perfect circles within a
diameter of 2 m

136

J. Res. Lepid.

i — __ — 1 50 m

I, II, III, IV, V

study areas

o

P. thoas: territories

o

B. polydamas : territories

□

highest point

trees

■p

E. orthosilaus (above trees)

Fig 1 . Preferred sites for territories. (A) Territories of P. thoas, near the highest point,
and of B. polydamas, when both species occur together. (B) Territories of
B. polydamas, near the highest point, after remotion of males of P. thoas.
Territory boundaries fit displacements done during “restricted patrolling”.

north side of the hill. Several other species of butterflies, skippers and
bees were also observed using this place for hilltopping.

Behavioral activities were categorized based on aggressivity and indi-
vidual displacement over the area (Table 1). Records of behavioral
categories were taken at the end of the rainy season (April). A scan
method (Altmann, 1974) was used for quantifying frequencies of behav-
ioral categories. The instantaneous activities of butterflies were recorded
every 5 minutes throughout the hilltopping period.

29(1-2): 134- 142, 1990(91)

137

During scan intervals butterfly displacements were also recorded by
six previously trained observers placed at special points of the hilltop.
With this larger number of observers it was possible to follow butterflies
individually for long periods of time. Preferred sites for territories were
determinated by the number of inter and intraspecific aggressive en-
counters in different areas of the study site (Fig. 1).

Results

HILLTOPPING

Hilltopping is concentrated in the warmer periods of the day. Males
generally appear on the hilltop between 0900-1130 AM depending on the
time of the year and temperature of the day and can remain there for
more than three hours. Territorial hilltopping was only observed after
the rainy season (April and May) when there was a relatively greater
number of butterflies on the hilltop and aggressive encounters among
individuals were frequent. In October 13-16 the number of butterflies
present on the hilltop was small and males did not form territories.

BEHAVIORAL DIFFERENCES AMONG SPECIES

Differences in the kinds and frequencies of behavioral categories
performed by males of each species during territorial hilltopping are
shown in Table 2. The maximum number of butterflies on the hilltop
recorded during scan sampling and the number of established territories
was 3/2 for P. thoas, 5/4 for B. polydamas and 5/3 for E. orthosilaus.

P. thoas and B. polydamas shared a larger number of behavioral
categories and showed a more diversified behavior than E. orthosilaus ,
who do not use “perching” and “flying in circles” behaviors. The latter was
also not observed in B. polydamas. The most common behaviors shown
by P. thoas and E. orthosilaus were extensive and restricted patrolling,
and hovering by B. polydamas.

Nearly 25% of all behavioral records of P. thoas and B. polydamas and
15% of E. orthosilaus were related to aggressive encounters. In all cases
these encounters were more frequent among conspecific males than
between species (see Table 3).

More subtle differences among species also occurred within each
behavioral category. Aggressive acts among conspecific males of E.
orthosilaus can follow a standard sequence, always starting with hori-
zontal or vertical chasing among two or more males and can last several
minutes, sometimes ending very far from the hilltop. Chasing generally
stops when one of the males flies away from the other or from the
hilltopping arena. Chases usually determine territory owners. However,
in some cases chasing can be followed by grappling in the air and falling
to the ground. Depending upon the height and the place of the fall, serious
wing damage can result. The butterflies hold each other with the legs and
even with the clasper used by males to hold females during copulation.
Grappling and falling were also seen in P. thoas , but more rarely than in
E. orthosilaus , and never in B. polydamas.

138

J. Res. Lepid.

Table 2. Relative frequency of behavioral categories performed by males P.
thoas, B. polydamas and E. orthosilaus throughtout a territorial hilltopping day in
April, in the Rio Manso, Western Brazil (records were taken at five minute
intervals; n= number of records)

Behavioral categories

Papilio thoas Battus polydamas Euritides
(n^62) (n=61 ) orthosilaus (n=47)

Intraspecific Aggression

19.35

21.31

14.29

Interspecific Aggression

4.83

3.28

—

Extensive Patrolling

25.81

9.84

38.78

Restricted Patrolling

25.81

24.59

28.57

Perching

6.45

6.56

—

Hovering

9.68

34.43

18.37

Flying in circles

8.06

—

—

Table 3. Number of intra and interspecific aggressive encounters (AS and AO
respectively) of P. thoas, B. polydamas and E. orthosilaus per area, during
territorial hilltopping in the Rio Manso, Western Brazil

AREA

P.

AS

thoas

AO

B. polydamas

AS AO

E. orthosilaus

AS AO

l,V

II

III

IV

11

3*

2 2*

3

2

5

* Interspecific encounters were between P. thoas and B. polydamas, except one
case against a passing pierid butterfly

Extensive and restricted patrolling of males ofP. thoas and B. polydamas
occurred near the ground. Restricted patrolling is done over a very
defined area that the animals cover repeatedly, following the same route.
E. orthosilaus patrolled at higher elevations, near 5 m above the canopy
of the large trees, and the routes followed were not repeated as in the
other two species.

Activities related to feeding or oviposition by adult butterflies were
never observed during any period of the year. Furthermore, hostplants
of these species (JPiper sp, forP. thoas, Aristolochia vines for B. polydamas
and probably a Lauraceae used by E. orthosilaus ) were not found among
the local plants.

29(1-2):134-142, 1990(91)

139

Mating success of these species seemed to be small given that matings
were not observed. However, sometimes males of E. orthosilaus , when
pursuing butterflies, possibly females, flew away from the hilltop with-
out returning afterwards.

SEQUENTIAL PATTERNS OF BEHAVIORAL ACTIVITIES

The sequence in which males of all species showed each behavior and
formed territories reveals that certain activities predominate at different
times during hilltopping. Extensive patrolling is more common in newly
arrived butterflies and before leaving the hilltop. During initial explor-
atory displacements other males were frequently encountered and ag-
gressive acts occurred, characterizing an aggressive phase in the early
hours. As territory owners became established and newly arriving
butterflies became rare aggressive acts decreased.

After choosing a site for territory, males performed restricted patrol-
ling for several minutes. Inside the area delimited by such behavior,
which can be considered as the territory limits, males started to hover
and perch (except for E, orthosilaus ). Finally P. thoas began to fly in
circles whereas B . polydamas andE. orthosilaus predominantly hovered.

TERRITORIAL SITES

The most desirable area for territory of each species as indicated by the
number of aggressive encounters is shown in Table 3. P. thoas and B.
polydamas preferred the same area (III) near the highest point of the
hilltop. Most interspecific aggressive encounters occurred between these
species, leading to the formation of nonoverlapping interspecific territo-
ries. P. thoas , the dominant species, often won these encounters and
established its territory first. Under these conditions B. polydamas set
up territories in neighboring areas placed around the territories of P.
thoas (Fig. 1A). An aditional experiment done afterwards showed that
when males of P. thoas were removed from the hilltop, territories of B.
polydamas were also set up in area III (Fig. IB). Landmarks like the
highest point, trees and small plants were used by both species as
reference points for territory boundaries.

Territories of E. orthosilaus did not overlap with the other species.
They were set up to form a barrier to newly arriving butterflies at the
hilltop. These butterflies hovered or patrolled continuously several
meters above the canopy of the larger trees, which were used as central
reference points in their territories.

In all cases territories were set up on the north side of the hill, and only
pursuing males were observed flying to the south side (Fig. 1, area V). As
newly arriving butterflies were always observed coming from the north,
it is possible that the formation of territories on this side of the hilltop can
be related to the proximity of the forests below the hill, and/or to solar
movements.

140

J. Res. Lepid.

Discussion

Results support the idea that these swallowtail butterflies are indeed
territorial hilltoppers, with territoriality defined in the sense of Brown &
Orians (1970). The assembly of males on the hilltop is also similar to the
lek behavior of many vertebrate species; see also DeVries (1980) for a
Pierine butterfly.

The defense of landmark territories seems to be a common mate-
locating tactic among Papilionid butterflies. However, many differences
related to site tenacity and aggression toward intruder males have been
shown in some well studied species. Males of Papilio zelicaon and P.
polyxenes actively defende hilltop territories to which they return over
days or even weeks (Shields, 1967; Lederhouse, 1982). Otherwise males
of Battus philenor perch on peaks and/or patrol routes along ridgetops
but show little site tenacity and interact only briefly with conspecific
males (Rutowski et al., 1989). The more aggressive behavior in the genus
Papilio when compared to Battus species was also observed in this work.
Territorial preferences and behavioral activities were so similar between
P. thoas and B. polydamas as to lead to interspecific aggression and
nonoverlapping territories. P. thoas often won aggressive encounters and
established its territories in the most disputed sites while B. polydamas
was compelled to use neighbour areas. Such observations may suggest
that interspecific competition for territories played an important role in
the evolution of different mate-locating tactics among swallowtail but-
terflies and even among other hilltopping species, since similarities in
male preferences for territories have been shown for many taxa includ-
ing distantly related species like butterflies and wasps (Alcock & ONeill,
1987).

Aggressive encounters in adult butterflies, which do not bear any
specialised organ or appendix to injure other animals, are often charac-
terized by ritualized flights and chases that rarely lead to contact among
individuals as those observed inP. thoas and especially in E. orthosilaus.
This kind of behavior seems to be common in other Eury tides species like
E . protesilaus (Cr., 1707) also present in the region and frequenty misled
with E. orthosilaus when observed at distance.

Territorial hilltopping seems to be density-dependent in swallowtail
butterflies since it was never observed when few individuals were
present on the hilltop (see Brown et al. 1981 for population dynamics of
B. polydamas and other troidine swallowtails). The most common
behaviors observed in solitary males on the hilltop were extensive
patrolling and/or perching near the highest point. In this situation
butterflies frequently left early or appearred on the hilltop intermit-
tently.

“Restricted patrolling” and “perching” in these butterflies could be
considered as true territorial behaviors. In the case of P. thoas and B.
polydamas , restricted patrolling was more frequently observed when
two or more males were together in the area. In two cases, when a male

29(1-2):134-142, 1990(91)

141

of P. thoas was removed from the territorial site its conspecific neighbor
increased its flight area, returning to extensive patrolling again. “Perch-
ing” has been also observed in other nonterritorial butterflies and thus
cannot be necessarily considered a conspicuous territorial behavior.
According to Scott (1986) it occurs because butterfly eyes are better at
detecting fluttering movements than detailed shape and color pattern.
Butterflies approach animals to smell and see them in order to find out
if they be females or males of the same species. On the other hand, in the
case ofP. thoas and B .polydamas , aggressive acts always ensue when the
passing butterfly is a conspecific male. Under such conditions many
authors regard this behavior as territorial (see revision of Fitzpatrick &
Wellington, 1988).

The greater behavioral differences found between E. orthosilaus and
the other species are probably due to the ability of the '‘kite” swallowtail
E. orthosilaus to use the wind for gliding or hovering motionless for
relatively long periods of time without strong wing movements; therefore
perching may not be necessary. P. thoas and B. polydamas must alter-
nate behaviors like hovering and restricted patrolling with resting on
some projection, since these behaviors are more expensive energetically
due to the strong wing movements required.

Mating success of hilltopping males seemed to be small. Matings were
not observed during several full day observation sessions and did not
occur during scan sampling periods. A relatively low mating frequency
has been also reported for other hilltopping invertebrates (Alcock, 1981;
1987). According to Shields (1967) matings are rarely observed because
females of many species of hilltopping butterflies generally mate only
once or a few times. For at least one hilltopping swallowtail butterfly, P.
polyxenes , it was observed that most matings occurred in territories
placed in small elevations in the area (Lederhouse, 1982).

Acknowledgments. I thank J.V.C. Ortiz, O.A. Shibatta, S.O. Ferreira, N.F. Martins,
G.C. Dalton and B.F. Dias for assistance in the field. K.S. Brown Jr., P.J. DeVries
and an “anonymous” reviewer read drafts of the manuscript and furnished many
helpful comments. Field facilities were provided by Eletronorte.

Literature Cited

Alcock, J., 1981. Ridgetop rendezvous. Nat. Hist. 93(l):42-47.

, 1985. Hilltopping in the Nymphalid butterfly Chlosyne californica

(Lepidoptera). Arner. Midi. Nat. 113(l):69-75.

, 1987. Leks and hilltopping in insects. J. Nat. Hist. 21:319-328.

Alcook, J. & O’Neill, K.M. , 1987 . Territory Preferences and Intensity of Competition
in the Grey Hairstreak Strymon melinus (Lepidoptera, Lycaenidae) and the
Tarantula Hawk Wasp Hemipepsis ustulata (Hymenoptera, Pompilidae). Am.
Midi. Nat. 118:128-138.

Altmann, J., 1974. Observational study of behavior: sampling methods. Behaviour
49(3/4):227-267.

Baker, R.R., 1983. Insect territoriality. Ann. Rev, Entomol. 28:65-89.

142

J. Res. Lepid.

Baughman, J.F., Murphy, D.D. & Ehrlich, P.R., 1988. Population structure of a
hilltopping butterfly. Oecologia (Berl.) 75:593-600.

Brown, J.L. & Orians, G., 1970. Spacing patterns in mobile animals. Ann. Rev.
Ecol. Syst. 1:239-262.

Brown, K.S.Jr., Damman, A.J. &Feeny, P., 1981. Troidine swallowtails (Lepidoptera:
Papilionidae) in southeastern Brasil: natural History and foodplant
relationships. J. Res. Lepid. 19:199-226.

Courtney, S.P. & Anderson, K, 1986. Behaviour around encounter sites. Behav.
Ecol. Sociobiol. 19:241-248.

DeVries, P.J., 1980. Observations on the Apparent Lek Behavior in Costa Rican
Rainforest Perrhybris pyrrha Cramer (Pieridae). J. Res. Lepid. 17:142-144.
Fitzpatrick, S.M. & Wellington, W.G., 1983. Insect territoriality. Can. J. Zool.
61:471-486.

Lederhouse , R. C . , 1 982 . Territorial defense and lek behavior of the black swallowtail
butterfly Papilio polyxenes. Behav. Ecol. Sociobiol. 10:109-118.

Poulton, E.B., 1904. A possible explanation for insect swarms on mountain- tops.

Entomol. Soc. London Trans. 1904:XXIII-XXVL
Rutowski, R.L., Alcock, J. & Carey, M., 1989. Hilltopping in the Pipevine
Swallowtail Butterfly ( Battus philenor ). Ethology 82:244-254.

Scott, J.A., 1986. The butterflies of North America: a natural history and field
guide. Stanford IJniv. Press. Stanford. XTV + 583pp.

Shields, O., 1967. Hilltopping. J. Res. Lepid. 6:69-178.

Journal of Research on the Lepidoptera

29(1-2):143-156, 1990(91)

Ball Mountain Revisited: Anomalous Species
Richness of a Montane Barrier Zone

Arthur M. Shapiro

Department of Zoology, University of California, Davis, California 95616

Abstract. Ball Mountain, located in eastern Siskiyou County, north-
ern California, has one of the richest documented butterfly faunas of
any area its size in temperate North America, now numbering 101
entities (species and subspecies) recorded in 18 collecting trips over 8
years. It has a clear faunistic relationship with the Warner Mountains,
some 225 km distant. Ball Mountain separates two different climates,
and its faunal richness is enhanced by the presence of vicariant species
or subspecies pairs in several complexes on its eastern and western
slopes. The possible role of “mass effect” in generating butterfly species
richness is discussed.

What determines species richness, the number of species of a given
taxonomic group in an area? A vast literature has grown up in ecology
attempting to answer this question, as any introductory textbook in that
discipline attests; the subject has gained in urgency when applied to
conservation problems. Particular efforts have been devoted to compari-
sons of species richness along various gradients, such as between tropical
and temperate regions. Anomalously low or high species counts within a
region may be useful in identifying important factors associated with
those sites which affect species richness.

Ball Mountain is a 2330 m basaltic volcano of Pliocene age located in
eastern Siskiyou County, northern California, 47 km N of Mount Shasta.
Its geographic and vegetational characteristics are described in Shapiro
(1986), along with a summary of its principal butterfly habitats. A
butterfly faunal survey of Ball Mountain was initiated in 1983, triggered
by its intriguing location as an isolated high point in a biogeographically
complex region. An initial visit demonstrated extreme telescoping of the
altitudinal vegetation belts, important slope- and exposure-mediated
vegetational discontinuities, and the presence of numerous rare or
disjunctly-distributed plant species. (Although Goosenest is a larger
volcano at 2524 m, it is geologically younger and much less well-
vegetated.) As of late 1985 the accessible areas above 1500 m, encom-
passing representatives of all identified butterfly habitats, had been
collected seven times: Shapiro (1986) reported 68 species, a rich fauna
for an area less than 150 km2 even given the pronounced vegetational
zonation.

Since then Ball Mountain has been collected 11 more times, raising the
total taxa recorded there to 101. Many of the additional taxa were found
to be intensely localized in sites on the eastern and southern sides of the

144

J. Res. Lepid.

mountain. Although rigorous comparisons are not available, it is evident
that this must stand as one of the richest butterfly faunas, for its area,
in temperate North America. Shapiro, Palm and Wcislo (1981) recorded
115 species in the Trinity Alps, encompassing much more topographic
and vegetational diversity in an area of 2030 km2. No 150-km2 segment
of the Trinities is likely to contain 100 species. The richest butterfly
fauna recorded in California (and perhaps temperate North America) is
that of Donner Pass in Placer and Nevada Counties (Sierra Nevada, 2100
m), with roughly 115 species in some 50 km2 (Shapiro, unpublished data,
1972-1990). However, Donner is embedded in a large and topographi-
cally diverse mountain range, while Ball Mountain is one of only three
mountain “islands” above 2270 m between Mount Shasta and central
Oregon, and most of the surrounding country is below 1300 m. Thus the
resident portion of its fauna, if montane-endemic, must represent either
historic (relict) or recent colonization from substantial distances away.
Shapiro (1986) pointed out a distinct endemic element in the Ball
Mountain fauna, as well as an apparent connection to the fauna of the
Warner Mountains in Modoc County, 225 km to the E.

Subsequent collecting has revealed an even more pronounced Warner
Mountain connection in the fauna. The complete collection data for 18
days afield are given in Table 1. Comments follow on some especially
interesting additions to the fauna. They should be read in conjunction
with Shapiro (1986) and the most relevant major faunistic treatment,
Dornfeld (1980) for Oregon. A revised faunal breakdown by families
appears in Table 2.

Pieris napi L. — The “parent colony” of the single previous record has
been located in the Shovel Creek (E slope) drainage. Only a spring brood
has been found, but the phenotype is indistinguishable from the hitherto
unique Warner Mountain one. The host plant has not been identified.
Anthocharis sara Lucas. — Previous records were based only on males
and were characterized as the nominate subspecies sara. Now that
females are in hand it appears that they should be assigned to ssp. thoosa
Scudder or flora Wright, with yellowish upperside in the female. What-
ever the applicable name, they resemble Warner Mountains ones.
Anthocharis lanceolata Bdv. — Uncommon, but present on both the W
and E slopes at 1650-1800 m, primarily in canyons.

Coenonympha tullia Mueller complex. — West-slope specimens fall
comfortably into the taxon eryngii Henry Edwards. East-slope ones are
highly variable, with some approaching elko W. H, Edwards. Populations
along Prather Ranch Road at the E base of the mountain are especially
dense. This may be an intergrade zone. Porter and Geiger (1988) treat
several nearby populations as such. Montague (Shasta Valley, west of
Ball Mountain) is treated as eryngii , while MacArthur and Adin, to the
SE, and Goose Lake, just W of the Warners (and nearly due E) are treated
as intergrade zones. Specimens from the far W of Butte Valley (Sams
Neck-Meiss Lake Road) are mostly very similar to Goose Lake ones.

29(1-2):143-156, 1990(91)

145

Oeneis nevadensis Feld. & Feld. — Expected, but first collected only in
1988 as it flies only in even-numbered years; widespread and fairly
common.

Cercyonis pegala Fabr. complex. — Occasional specimens of ssp. boopis
Behr occur on the W slope up to 1850 m. On the E slope the situation is
more complicated. The species is very abundant in the lower E-slope
canyons and especially at the foot of the mountain, along Sams Neck-
Meiss Lake Road, whence it extends into marshy grasslands in Butte
Valley. These populations are extremely variable and seem to represent
intergrades from boopis to nearly typical ariane Bdv., with all pheno-
types represented in a series of 26 collected 22. VII. 1990. They also occur
in a range of habitats from open conifer forest and shrubland to grassy
swales with no woody vegetation. Boopis is common along streams in
Shasta Valley W of Ball Mountain and even occurs in alfalfa fields there.
Euphydryas chalcedona near wallacensis Gunder. — Common at mid-
elevation on both slopes, and indistinguishable from a long series from
Cedar Pass in the Warner Mountains. In 1990, under severe drought
conditions, this species flew a full month early on Ball Mountain. In June
1991 it reached outbreak abundance, flying by thousands.

Speyeria zerene Bdv. — Abundant. In the Warner Mountains, E -slope
populations are of the pale ssp. gunderi Comstock and W-slope popula-
tions show a complete and very confusing array of phenotypes intergrading
from gunderi to conchyliatus Comstock. With about 60 Ball Mountain
specimens on hand it is evident that about two-thirds of this variation
exists on Ball Mt.; the most gunderi -like phenotypes are missing, but
some females are rather clay-colored above and olivaceous, vaguely
greenish-brown below and such specimens are not known elsewhere in
the region except in the Warners. One possible explanation of this
situation is an historic presence of gunderi here, later swamped out in
pure form through hybridization with conchyliatus. Such a scenario is
consistent with a Hypsithermal Warner-Ball Mountain connection,
discussed below.

Speyeria mormonia Bdv. — Comparison of our series (including six 1990
captures) with specimens from Bidwell Peak in the Warner Mts.
(21.VII.1985, leg. C. Hageman) shows Ball Mt. specimens consistently
lacking the greenish VHW basal flush usually present in the Warners. In
this regard they match northern Sierran material more closely, though
on other characters they are essentially intermediate.

Satyrium fuliginosum W. H. Edwards. — - Three Ball Mountain speci-
mens: two from Little Shasta Meadow, 21. VI. 1987 and 22. VII. 1990 and
one from the mid-W slope, 24.VII. 1987. What is confusing is that the two
Little Shasta specimens appear subspecifically different from the other.
The W-slope animal resembles Trinity-Eddy ones, while the Little
Shasta ones are smaller, intense black on the upper surface, and with the
heavily-spotted high-desert facies below. This problem must remain
unresolved until breeding can be established, but at least the Little
Shasta situation implies residence.

146

J. Res. Lepid.

Table 1 . Occurrence of butterfly species on Ball Mountain on 18 collecting days,

1 983-1 991 . On all occasions the mountain was sampled comparably except on
4.IX.1987, when increasing cloudiness terminated butterfly activity in early
afternoon. Colias hybrids are not counted separately in the totals, but named
subspecies of polytypic complexes ( Cercyonis pegala, Lycaena editha/
xanthoides) are. For further information, see text. (Note that 1 990 phenology
was unusual; low snow pack produced a full month of acceleration at the
beginning of the flight season, followed by cold, wet weather in late spring leading
to a normalization to a slight deceleration in midsummer.)

Papilio zelicaon
Papilio rutulus
Papilio eurymedon

Parnassius clodius
Neophasia menapia

Pontia sisymbrii
Pontia beckerii
Pontia occidentalis
Pieris napi "Warner"

Pieris rapae

Colias eurytheipe

Colias philodice eriphyle

(Colias x hybrids)
Anthocharis sara sara (?)
Anthocharis lanceolata
F.uehloe ausonides
Coenonynpha tullia nr . eryngii
Coenonympha tullia nr . elko
Oeneis nevadensis
Cercyonis pegala nr. ariane
Cercyonis pegala nr. boopis
Cercyonis silvestris

Danaus plexippus
Limenitis lorquini
Adelpha bredowii californica

Vanessa virginiensis

Vanessa cardui

Vanessa annabella

Vanessa atalanta

Precis coenia

Wymphalis californica

Kymphalis rcilberti f urcillata

Hymphalis antiopa
Polygonia faunus rusticus

Polygonia oreas silenus

Polygonia zephyrus
Polygonia satyrus
Phyciodes campestris

Phyciodes mylitta

Chip syne hof fmanni segregata

Chlosyne palla

Euphydryas chalcedona

Euphydryas editha ssp.

Boloria epithore
Speyeria coronis nr. simaetha
Speyeria zerene nr ■ conchy liatus
Speyeria callippe nr . rupestris
Speyeria egleis nr . oweni
Speyeria atlantis ssp .

XXX XX

X XXX X

X XX X

X X X X XX

XXX XXXXX X

XXX

X XX

xxxxxxxxxxx

XXX XX XX X

XX XX X XX

X X

X XX

x X X X X X X

Xx XX xxxx

X

XXX
X X

X

X XX

Xx X

XXXXXXXXX

X XX

xx XX XX X

X XX XX

X X

X X

XXX X

x X XXX

XXX XX X

X
X

XXX X XXX

X

XXXX X X
XXX XX XXX

X

XX XX

XXXXX
X

XXXX X X

XXX

XXX xxxx

X X X X X X X

XXXX XXXX

xxxx xxxx

X

X X X X X X X

X xxxx
X X X X X X X

X X XX

XXX XXX

X X XX

X XX

X X X X X X

X

X X

X X XX

X X

X X XX

X XX

XX X

X XXX

X X XX

XX X

X

X X
X

XX XXX

XXX XXX

X XX

X X X X X X X

XXX XXX

X X X X X X X

29(1-2):143-156, 1990(91)

147

mi r« vo in

o ^ 00 oo oo oo

O' oi m o O' o

O' 0\ i-l i-l i-( i-l

O 0O S OJ OJ

• M M M M W

>>>>>>

Speyeria cybele leto
Speyeria mormonia ssp.

Speyeria hydaspe nr .purpurascens XXX

Strymon melinus X

Satyrium saepium X

Satyrium sylvinus X X

Satyrium californica X X

Satyrium tetra X

Satyrium auretorum X

Satyrium fuliginosum X

Mitoura nelsoni X x X X X X

Mitoura spinetorum

Incisalia mossii ssp. X X

Incisalia eryphon XXX XX

Incisalia iroides X X

Lycaena arota X

Lycaena heteronea gravenotata
Lycaena rubidus duofacies
Lycaena xanthoides/editha

Lycaena editha X

Lycaena gorgon X X

Lycaena helloides X X X X X X

Lycaena nivalis "Warner" X X X X

Plebeius Idas nr . anna/ricei X X

Plebeius melissa "Warner" X XX

Plebeius saepiolus X X x X

Plebeius icarioides X X x X X

Plebeius acmon XX XX

Plebeius lupini XXX

Agriades "glandon" ssp. X X

Everes comyntas X

Everes amyntula X X X X X X X

Glaucopsyche lygdamus Columbia XX X

Glaucopsyche piasus X X x

Celastrina argiolus echo X X X X X

Apodemia mormo nr . mormo

Ochlodes sylvanoides X X

Polites sonora XX XX

Polites sabuleti
Atalopedes campestris

Hesperia "comma complex" X

Hesperia lindseyi X X

Hesperia juba X X X X

Hesperia Columbia X x

Amblyscirtes vialis X

Thorybes mexicana aemilia X

Pyrgus ruralis X x X

Pyrgus communis X X

Epargyreus clarus X

Erynnis icelus X

Erynnis persius XX XX

Erynnis propertius X x X X X X

38 33 36 35 59 32 27 26 50 30 25 24 17 32 9 34 32 17

148

J. Res. Lepid.

Table 2. Composition by family of some regional butterfly faunas, revised slightly
from Shapiro, 1 986. The two Cercyonis and Coenonympha subspecies and
Lycaena editha and L. editha/ xanthoides intergrades are counted as separate
species, but Colias hybrids are not counted. Increases in the Ball Mountain fauna
since 1986 are concentrated in Lycaenidae (1 1 species), Hesperiidae (7) and
Nymphalidae (5), but the largest proportional increase was in Satyridae (5).

Additions to other faunas since 1986, not otherwise documented, are not

included.

Family

Trinity Alps

Mount Eddy

Crater Lake

Ball Mountain

Papilionidae

7

5

4

4

Pieridae

14

7

9

11

Satyridae

4

3

4

6

Danaidae

1

1

1

1

Nymphalidae

31

25

25

29

Riodinidae

1

1

0

1

Lycaenidae

32

27

26

33

Hesperiidae

25

11

11

16

Totals

115

80

80

101

Lycaena editha Mead. — Phenotypically normal editha , indistinguish-
able from Warner Mountain ones, occur from Little Shasta Meadow
down the E slope to Sams Neck-Meiss Lake Road, but at low density.
Apparent intergrades to L. xanthoides Bdv. occur on the W slope from
Ruck’s Cabin down, the E-most and highest-elevation intergrades yet
discovered. Such populations are abundant in alfalfa fields in Shasta
Valley (as at Montague), feeding on weedy docks. Apparent intergrades
thus occur within 4.5 km of apparently pure editha. A male taken at
Little Shasta Meadow on 22. VII. 1990 is as large as an intergrade but
phenotypically editha.

Lycaena rubidus Behr. — Occurs in sedgy swales in Butte Valley
(seemingly in the same ecological role as intergrading editha-xanthoides
in Shasta Valley), apparently straying upslope in the E-slope canyons on
Ball Mt. The possibility of competitive exclusion among these entities,
with dominance possibly mediated by climate, deserves further study.
Lycaena heteronea gravenotata Riots. — Although Shapiro (1986) specu-
lated that this entity might have a different host plant than normal L. h.
heteronea Bdv., two females were observed ovipositing on Eriogonum
umbellatum Torr. (Polygonaceae), the usual heteronea host in northern
California, at the quarry on the mid-W slope on 24. VII. 1987 and another
at Little Shasta Meadow, 22. VII. 90. The very heavily spotted pheno-
types figured by Shapiro ( 1986) have been found in every specimen of this
entity seen or collected throughout this study. Four local breeding
colonies are known on Ball Mt.

29(1-2):143-156, 1990(91)

149

Plebeius melissa W. H. Edwards. — The highly variable populations of
this complex in the Warners remain taxonomically unresolved at this
writing (S. Mattoon, pers. comm.). All our Ball Mt. melissa are of
“Warner” phenotypes and are from the E slope above 1550 m, except a
female taken 1. IX. 1990 on Haplopappus flowers on the S slope. There is
a P. idas anna Edwards - ricei Cross population at Martin Dairy and
Little Shasta Meadow, 6-8 km away. This population is quite distinct and
shows no sign of phenotypic intergradation. Its host plant has not been
determined, and it persists at unusually low density for the taxon.
Everes comyntas Godt. — - This record is based on a single female taken
on the lower W slope on 10. V. 1990. Apparent comyntas have also been
taken at Gazelle in Shasta Valley. Wright’s taxon sissona , from Sisson
(now Mt. Shasta City), was classified by Miller and Brown (1981) as a
synonym of comyntas , but reclassified by Ferris (1989) under amyntula
Bdv. The type, if extant, should be examined as both species apparently
do occur in far-northern California.

Plebeius saepiolus Bdv. - — On 10. V. 1990 a very dense population of this
species was found unexpectedly on the lower W slope of Ball Mountain,
in juniper-Oregon oak-bunchgrass steppe, closely associated with and
seen ovipositing on Trifolium macrocephalum (Pursh.) Poir. This site
was completely dry in July. A series of 26 ? , 4 / (all fresh) reveals a
remarkable degree of variability in the underside pattern, tending
toward obsolescence but with a few extremely heavily-spotted individu-
als. The average wing size is the largest I have seen in this species. The
existence of this low-altitude (1650-1800 m) ecotype, apparently adapted
to a vernal-ephemeral host in a novel dry habitat, demonstrates nicely
how little we do know about the northern California butterfly fauna!
Hesperia Columbia Scudder. — Apparently resident on the W slope
(only?), but only the first brood has been collected. (In Scott Valley
double-brooded, as further south in the Coast Range.)

Hesperia juba Scudder. — Taken visiting dandelions pushing through
mushy snow below the summit on 10. V. 1990; the only other butterfly
present was Pontia occidentalis Reakirt (extreme vernal phenotypes,
newly eclosed). As elsewhere in northern California, this species has a
perfectly clear early spring-autumn bivoltinism.

Hesperia comma L. complex. — West slope animals, collected at low
density mainly from thistles, are extremely variable phenotypically.
East-slope animals and those taken on the south slope with Polites
sabuleti Bdv. are apparently pure harpalus W. H. Edwards in phenotype.
They are at much higher density. They are indistinguishable from Cedar
Pass (Warner Mt.) series.

Hesperia lindseyi Holland. — - Locally common on the W slope only,
mostly in juniper-Oregon oak but straying up into mixed-conifer vegeta-
tion along roadsides, on Asclepias and Apocynum flowers. The darkest
lindseyi are phenotypically similar to light E-slope harpalus , but easily
separated from them by geography and flight season.

150

J. Res. Lepid.

Polites sabuleti Bdv. — Totally absent on the W slope, but common to
abundant on the E and S slopes mainly at Haplopappus flowers in
September (as at Jess Valley in the Warners). Apparently strictly single-
brooded, and phenotypically identical to both Warner and N slope Mt.
Shasta specimens.

Discussion

Shmida and Wilson (1985) defined a factor contributing to species
richness (or “diversity,” usually imprecisely used), which they called
“mass effect”: the establishment of species in sites where they cannot be
self-maintaining, but are regularly replenished from nearby permanent
populations. Their studies of plant communities in desert washes
revealed unstable populations of upland populations whose persistence
depends on seed rain from above; such populations appear deceptively
“persistent.” More or less the same mechanism was also invoked by Heck
(1979). Stevens ( 1989) claimed that, due to the allegedly narrower niches
of tropical as against temperate organisms, “mass effect” generated by
variance in microhabitats would contribute to high species richness in
many tropical environments.

“Mass effect” has been known to community ecologists and amateur
naturalists alike for many years, even if not formally named. Almost all
regional faunas and floras contain significant numbers of “stray” species
which breed occasionally or even fairly regularly— -enough to be
unsurprising as records, but without being truly permanent members of
the biota. Donner Pass, astride the mountain chain separating the
montane-Mediterranean Sierran W slope from the much more continen-
tal Great Basin, is a natural collector of dispersing butterflies, a fact
which accounts for a sizeable chunk of its high species count. Can
something of the sort be recognized at Ball Mountain? If so, whence come
the dispersers?

Roughly a fifth of the Ball Mountain species list is based on single
collections (one or more individuals in one location on one date alone).
This superficially suggests Ball Mountain as a collector of dispersers, but
the data must be interpreted cautiously. Four of these species ( Vanessa
atalanta L. , Speyeria cybele leto Behr. — females in late summer ,Atalopedes
campestris Bdv., Epargyreus clarus Cramer) are well-known for their
dispersal ability or tendency to “stray.” Three more are almost always
rare, and little is known of their population biology: Polygonia oreas
silenus W. H. Edwards, Mitoura spinetorum Hew., and Incisalia mossii
H. Edwards. The following species are more or less common nearby at
low elevations and stray upslope more or less frequently: from the W: C.
pegala boopis, Polygonia satyrus W. H. Edwards, Satyrium tetra W. H.
Edwards, S. auretorum Bdv.; from the E: C.p. cf.ariane,L. rubidus; from
both sides, Coenonympha, as well as much commoner species such as
both Colias. Thorybes mexicana aemilia Skinner is mysterious; its hosts
( Trifolium spp. are abundant and suitable habitats (moist meadows)

29(1-2):143-156, 1990(91)

151

likewise. Pontia sisymbrii Bdv. seems out of place and its status is
uncertain. It could conceivably breed on the grassy lava flow near the top
of the mountain, but no host has been found.

Among the common species which are probably not permanent resi-
dents high on the mountain, but depend on regular recolonization from
below, are such “weedy” taxa as Precis coenia Hbn., Plebeius acmon
Westwood & Hewitson;, Pontia beckerii W. H. Edwards, and perhaps
Adelpha bredowii Geyer, as well as the well-documented directional
seasonal migrants Danaus plexippus L. and Vanessa cardui L.

Is Ball Mountain a collector of strays? Below the Juniper-Oregon Oak
zone on the W slope the vegetation is disturbed weedy (annual) grass-
land, with abundant nectar sources in summer and fall (especially
Centaurea solstitialis L., Compositae, established within the past 15 yr)
and narrow corridors of riparian vegetation. To the E in Butte Valley is
agricultural land (alfalfa, potatoes), interspersed with shrub-steppe and
moist, sedgy swales. Montane forest butterflies dispersing through these
habitats would not lack for nectar sources, but would presumably keep
moving until they encountered a more congenial (forested) habitat. This
is less improbable than appears at first glance, if anecdotal evidence is
worth anything. For example, the only P. oreas silenus recorded in a 5-
yr study of the Trinities and Eddies (Shapiro, Palm and Wcislo 1981 —
plus 8 further yr of occasional collecting, AMS, unpublished) was taken
in an alfalfa field some 2 km from coniferous forest. Shapiro (1982)
reported surprisingly frequent visits to a suburban butterfly garden by
species whose breeding habitats were several km away. But the species
most likely to move among upland “islands” of forest, e. g., various
Speyeria, especially female zerene and coronis Behr, are common on Ball
Mt. and give no hint of being dependent on “mass effect” for their
continued existence; in the case of S. zerene, the phenotypic variability
argues quite otherwise unless we assume regular exchange of specimens
with the Warners, 225 km away.

A significant component of the faunal richness at Ball Mountain
consists of species with Warner Mountains affinities, and this is telling.
These are extremely unlikely to represent either recent or frequent
colonizations. In terms of dispersal capability and pattern of spatial
distribution, they are a mixed lot. Euphydryas chalcedona is a good
disperser and widespread; E. editha is highly colonial but fairly vagile;
S. mormonia is extremely colonial in the N state, with only three
populations known (Mount Eddy, Warners, Ball Mt., all phenotypically
different); the various Lycaenids and Hesperiids vary among them-
selves. Pieris napi and Lycaena nivalis appear to be very compelling
Warner-Ball disjuncts. It is unclear how L. heteronea fits into this. The
gravenotata phenotype does not occur in the Warners, but has a spotty
and very local distribution in the N state, strongly suggestive of
relictualism; it never seems to co-occur with normal heteronea.

152

J. Res. Lepid.

The essence of both Croizatian panbiogeography and vicariance bioge-
ography is the premise that congruent distributions in different taxa
should be interpreted as reflecting a common historical process. Shapiro
(1970) did this with the distributions of sedge-feeding marsh butterflies
in the eastern Nearctic, pointing out that a few seemingly similar
patterns ( e . g. , for Pyrgus communis Grote) were demonstrably irrel-
evant. As Humphries and Parenti (1986, p. 87) declare with reference to
the biogeography of the Southern Hemisphere, “The finding of a pattern
corroborated in whole or in part by numerous other groups of plants and
animals invites a general explanation...” The pattern ofWarner-Ball Mt.
disjunctions leads to a prediction that further floristic and faunistic
affinities are to be expected. The high diversity of ecological adaptation
and dispersal capability among the taxa so far identified, and the fact
that some are very poor candidates for dispersal, reinforce the appear-
ance of real pattern. So does climatology: strong easterly flow, which
would carry the Warner biota across Butte Valley to the Ball Mt.-Willow
Creek Mt. massif, is very rare, and the topography is overwhelmingly
aligned against it.

The most probable explanation of our putative pattern is not recent
dispersal and colonization, but a relictual distribution. During the
Xerothermic or Hypsithermal interval some 5-6000 yr BP (Axelrod,
1977), warm summers and reduced rainfall caused the elevational life-
zone gradients in the northern Sierra to migrate upslope. The subse-
quent return to somewhat cooler, moister conditions left isolated relict
populations of various plants, such as Foxtail Pine ( Pinus balfouriana
Grev. & Balf.) (Mastrogiuseppe, 1972), in northern California and in the
high central and southern Sierra Nevada, as the intervening low moun-
tains were unable to sustain populations of them. Similarly, Shapiro
(1970) inferred that the distribution of Pleheius melissa in the eastern
United States was of Xerothermic relict origin. In summary, the
butterfly connection between Ball Mountain and the Warners suggests
an ancient expansion of the more arid Warner climate, which allowed
penetration of eastern Siskiyou County by elements from farther east
which largely died out as the Xerothermic yielded to more modern
conditions. On Ball Mountain very local microclimates or soil conditions
may have permitted the persistence of species or (in S. zerene ) relict genes
in populations. Because this scenario makes specific paleoclimatic
predictions, it is potentially testable.

A Comment on Barriers

The relatively low massif of Ball and Willow Creek Mountain separates
two different climates and vegetation zones. Table 3 presents some
relevant climatological data. Mount Hebron Ranger Station (NOAA
#5941, 41°47’N, 122°02W, 1290 m) has much colder winters (and cooler
summer night) than Montague/Siskiyou County Airport in Shasta Val-

Table 3. Climatic comparisons between Mount Hebron Ranger Station (Butte Valley) and Yreka (west side Shasta Valley). Mount
Hebron data are frequently unavailable in winter, but data available from local ranchers indicate much colderr winters than at Yreka

with occasional minima to -34°C.

29(1-2):143-156, 1990(91)

153

o 03
^ CO

-r-: rV

CO 'St

O 00
00 h-
05

E z

3*

m

I o

E

■S CO

eg

'5.

"o

£ <

Q.

75

I ^

CO

TJ

E

CO ^

J>*

-C

■#*»

c

o

E 2
c5

03

>,

6 <
oo
o
00

05 __

CM CO
Uf> T~

evi cd

CM O
CO CO

o oa

05 O

d

eg

co

CD

05

C

©

cc

c

o

L.

JD

03

X

O CO
LO 00
00 CO

T—

^ ID LO

O oo lo
cxi K
m y'~ r~

ffiOr-
0 CO T-
E od 05

1 — f—

m

0 < LO

5 Z S

2 S

co

eg

CO

w.

0

O)

c

CO

OC

c

o

k_

n

©

x

c

©

©

s

©

X)

to

>.

CO

a

05

CO

m 1

© o
1

CO

© _

■E C

x .£

© c

sN
I I t

© 3

co .»=

CD

Cl lo

o

T-

E

CD

o

o

evi

04

©

T”

i<

©

T—

©

_>*

©

CD

CD

sz

o

CO

CO

05

CO

05

©

cm

1“

L.

©

CD

CO

©

i.

2

T“

**r

T—

CM

LO

05

co

A

T™

00

CD

00

o

OJ

CM

05

d

LL

-t—

r-

LO

CO

CD

LO

00

cxi

■T

CD

00

CM

T—

<

o

O

LO

z

o

CO

LO

’T

00

LO

<

T-.

LO

CO

z

T”

CO

05

T—

o

o

o

Q, h-

£ <0

m ‘

CO

05

00

<

Z

T“

CO

o

CD

T“

T—

CM

h-

X

<

00

Z

T“

X

X

d

od

CM

X

X

05

cd

T6™

<

X

Z

05

X

<

00

z

>

>

d

d

CO

co

>

>

05

d

CM

co

>

>

d

co

>

=

05

CM

CM

T“

>

=

LO

LO

cd _
©
CO

©

O)

c

CO

X

c

o

_Q

©

X

©

E ©

cd

C ©

-3s£

ZS

ZJ

©

o £

o ®

>

2 >-

2 >

co

c

CO

©

E

o'

>, °? 5

S £

03 ©

T™ >

©

5 I

6 J,
^ z

154

J Res. Lepid.

ley, or Yreka (NOAA #9866, 41°43', 122°38'. 800 m) across the valley.
This is reflected in the presence E of Ball Mt. of elements of the N Great
Basin fauna including C. p. ariane and L. ruhidus. (Other elements, e.
g. , Pontia beckerii , Colias philodice Godt., Plebeius melissa , etc. enter
Shasta and Scott Valleys as well.) Similarly, C. p. boopis , the L.
xant ho ides -edit ha intergrades, and H. lindseyi and Columbia behave as
W-slope entitites . Mitoura nelsoni Bdv. as it occurs on Ball Mt. may be
biologically two different entities, but neither Kurt Johnson (pers.
comm.) nor I can differentiate them except by ecological context. On the
W slope it feeds on Calocedrus decurrens Torr.(Cupressaceae) as else-
where. On the lava flow near the top, and on the lower E slopes and the
fringe of Butte Valley, it feeds on juniper. The juniper-feeding popula-
tions are on average slightly purpler in habitus than the incense cedar-
feeding ones. They may or may not be the same as the very abundant
juniper feeders farther E in Modoc Co.; perhaps only biochemical genet-
ics will finally clarify these issues.

The title of this paper calls Ball Mountain a “montane barrier zone.” In
what sense might it be considered a barrier? There is no evidence that
butterflies cannot fly across Ball Mountain. It is neither very high nor
capped with permanent ice or snow, and during butterfly season it is
seldom cloudy before midday. Yet, on a smaller scale, Ball Mountain
appears analogous to the Sierran crest as a faunal divide. At Bonner
Pass (2100 m) both Coenonympha tullia California Westwood and C. t.
ampelos Edwards occur as strays, ampelos at least breeds occasionally,
but neither persists and no gene flow is likely to occur between them
there. Farther N they intergrade (Porter and Geiger, 1988). Similar
phenomena occur at Bonner in the Plesperia u comma complex,” in Polites
sabuleti, and in the Phyciodes campestris Behr fmontana Behr complex
(Shapiro and Geiger, in preparation). At Ball Mountain we find the
H. ” comma complex,” C. tullia , C. pegala , and Lycaena editha/xanthoides
at least behaving in the same ways.

The actual nature of species (or subspecies!) borders - the lines beyond
which biological entities no longer persist - is very seldom understood.
“Intrinsic barriers to dispersal” (Ehrlich, 1961) tells us little. In butter-
flies such limitations may be essentially physiological ( Colias philodice ,
Shapiro, unpublished data), harking back to C. H. Merriam’s “laws” of
biogeography (Merriam, 1894). The role of Ball Mt. as a “barrier” is more
likely to entail its presence between two contrasting climates, than its
physically obstructing access from one to the other. The montane
coniferous forest that dominates Ball Mountain, rather than trapping
wandering strays, seems to serve as a filter permeable to some species
and not to others. It is very remarkable to observe the precise localization
of Polites sabuleti or Hesperia “comma9* of the harpalus phenotype to
well-defined areas on the E and S slopes of Ball Mt. when their favored
nectar source, Haplopappus sp., lines the roads and is in bloom simulta-

29(1-2):143-156, 1990(91)

155

neously all over the mountain. It would be more remarkable to under-
stand it.

The presence of so many species in so small an area seems to represent
in this case the conjunction of four factors: proximity of two faunas;
attraction of dispersing montane butterflies; the presence of a relict
Warner Mountain fauna; and an unusually telescoped vegetation gradi-
ent which presents many habitats over short distances. When the
richness of the Ball Mountain fauna was first discovered, and the large
number of rare species noted, it was initially assumed “mass effect” and
collection of strays would account for much of the phenomenon. That
possibility seems increasingly remote.

Note added in proof

Since this paper went to press, a collecting trip 6.VII.91 recorded 48
species on Ball Mt., three of them new to the site: Papilio multicaudatus
(Kirby) (Papilionidae), Euphilotes enoptes (Bdv.) ssp. (Lycaenidae), and
Carterocephalus palaemon (Pallas) (Hesperiidae), all in the Shovel
Creek drainage on the E side. This raises the total taxa to 104. This trip
also turned up a colony of Hesperia lindseyi at Shovel Creek (first record
on the E) and established a host plant of P. anna/ricei there {Lotus
oblongifolius (Benth.) Greene, Leguminosae). The complete species list
follows: P. zelicaon, P. eurymedon, P. rutulus, P. multicaudatus, P.
clodius, P. beckeri, P. occidentalis, C. eury theme, C.philodice,A. lanceolata,
E. ausonides, C. tulli nr. eryngii, C. p. boopis, C. siluestris, D. plexippus,
L. lorquini, V. cardui, N. californica, N. antiopa., P . zephyrus, P. campestris,
P. mylitta, C. h. segregata, C. palla, E. chalcedona, B. epithore, S. coronis,
S. z . conchyliatus, S. c. nr. rupestris, S. h. purpurascens, M. nelsoni, I.
iroides, L. xanthoides/editha, L. helloides, L. nivalis, P. i. anna/ricei, P.
saepiolus, P. icarioides, P. acmon, P. lupini, E. amyntula, G. piasus, C. a.
echo, E. enoptes, P. communis, E. persius, C. palaemon, H. lindseyi.

Acknowledgments. This research has been supported by California Agricultural
Experiment Station Project CA-D*-AZO-3994-H, “Climatic Range Limitation of
Phytophagous Lepidopterans.” I thank Adam Porter, John Wagoner, Rick and
Carol Kareofelas, Lee Simon, and Bill Overton for companionship on trips since
1986 and Dave Olson, Linda Farley and Katie Giberson for additional data. Bill
Overton first introduced me to Ball Mountain. In his defense, he didn’t know it was
loaded.

Literature Cited

Axelrod, D. I. 1977. Outline history of California vegetation, pp. 139-194. in M.
G. Barbour and J. Major, eds., Terrestrial Vegetation of California. Wiley-
Inter science, New York.

Dornfeld, E. J. 1980. The Butterflies of Oregon. Timber Press, Forest Grove,
Oregon. 276 pp.

156

J. Res. Lepid.

Ehrlich, P. R. 1961. Intrinsic barriers to dispersal in the checkerspot butterfly,
Euphydryas editha. Science 134: 108-109.

Ferris, C. D. 1989. A supplement to A Catalogue/Checklist of the Butterflies of
America North of Mexico. Lepid. Soc., Memoir #3.

Heck, K. L. 1979. Some determinants of the composition and abundance of motile
macroinvertebrate species in tropical and temperate turtlegrass ( Thalassia
testudinum ) meadows. J. Biogeogr. 6: 183-200.

Humphries, C. J. and L. R. Parenti. 1986. Cladistic Biogeography. Oxford
Monographs on Biogeography #2. Oxford University Press. 98 pp.

Mastrogiuseppe, R. J. 1972. Geographic variation in foxtail pine {Pinus balfouriana
Grev. & Balf.). MS Thesis, Humboldt State University, Areata, CA.

Merriam,C. H. 1894. Laws of temperature control of the geographic distribution
of terrestrial animals and plants. National Geographic Magazine 6: 229-238.

Miller, L. D. and F. M. Brown. 1981. A Catalogue/Checklist of the Butterflies of
America North of Mexico. Lepid. Soc., Memoir #2.

Porter, A. H. and H. J. Geiger. 1988. Genetic and phenotypic population structure
of the Coenonympha tullia complex (Lepidoptera: Nymphalidae: Satyrinae) in
California: no evidence for species boundaries. Can. J. Zool. 66: 2751-2765.

Shapiro, A. M. 1970. Postglacial biogeography and the distribution of Poanes viator
and other marsh butterflies. J. Res. Lepid. 9: 125-155.

1982. Within-range butterfly dispersal: an urban garden as a detector.

Atala 8: 46-49.

1986. Montane insular butterfly biogeography: fauna of Ball Mountain,

Siskiyou County, California. Great Basin Nat. 46: 334-347.

Shapiro, A. M., C. A. Palm and K. L. Wcislo. 1981. The ecology and biogeography
of the butterflies of the Trinity Alps and Mount Eddy, northern California. J.
Res. Lepid. 18: 69-152.

Shmida, A. and M. V. Wilson. 1985. Biological determinants of species diversity.
J. Biogeogr. 12: 1-20.

Stevens, G. C. 1989. The latitudinal gradient in geographical range: how so many
species coexist in the tropics. Am. Nat. 133: 240-256.

Journal of Research on the Lepidoptera

29(1-2):157-160, 1990(91)

Butterflies from Nagpur City, Central India
(Lepidoptera: Rhopalocera)

T. N. Pandharipande

Department of Zoology, Nagpur University, Nagpur 440 010, iNDIA

Abstract. Butterflies recorded at Nagpur city in central India from
1.1. 1987 to 31. XII. 1989 are listed and compared with the fauna of south
India. A total of 61 species have been recorded. The butterfly fauna of
Nagpur differs from fauna of Nilgiri (Blue) Mountains of south India
and this might be due to differences in climate between the regions.

Introduction

The butterfly fauna of India is quite well known (Winter-Blyth, 1957;
Laithwaite et al., 1975; Smart, 1975; Larsen, 1987a), but local lists of
species at regional or city levels are very few (Antram, 1924; Larsen,
1987b, c, d; Wynter-Blyth, 1943, 1944, 1946, 1947; Yates, 1935, 1946).
Most of these studies deal with south India fauna and there are no lists
of species from central India. Nagpur (Fig. 1) is located at the center of
India, 20° 9'N and 79° 9'E. This paper is an introduction to its butterfly
fauna.

Methods

Butterflies were collected weekly from 1.1.1987 to 31.XII.1989 at various sites
within Nagpur city (Fig. 1). At other times, specimens were collected and
identified if new species were observed. Identification was made from published
literature (Evans, 1932; Laithwaite et al., 1975 and Smart, 1975). Some of the
specimens were sent to Dr. Jun Mitsuhashi, Tokyo, Japan and Dr. C. D. Ferris,
Laramie, Wyoming, USA for identification. Genitalic examination was carried
out to confirm identification of some species.

Results and Discussion

The 61 species of butterflies I recorded in Nagpur are listed in Table 1.
The butterfly fauna of Nilgiri (Blue) Mountains of south India is rich and
very interesting with 300 species (Larsen, 1988). Some of the largest
Papilionidae, such as Troides minus , Papilio helenus daksha, P.
polymnester and P. paris tamilana are recorded from Nilgiri (Larsen,
1987a); however, none of these was recorded from Nagpur.

Papilionidae, Lycaenidae, Nymphalidae and Danaidae recorded here
also occur in south India (Larsen, 1987b-d). The members of Pieridae
and Satyridae recorded in this investigation are noted from south India
except for Pieris napi, Anthocharis cardamines, Mycalesis malsarida,
M. nicotia and Callerebia scanda, which occur in Nagpur.

The common evening brown, Melanitis leda, occurs in two very differ-
ent seasonal forms. The wet season form occurs during monsoon (Jun-

158

J. Res. Lepid.

Oct) and is replaced by camouflaged dry season form which continues
throughout the winter and spring (Nov-May). Seasonal variations in
this species have been studied earlier (Brakefield and Larsen, 1984).

The only member of Acraeidae in India, Acraea violae (Smart, 1975)
has been recorded from both Nagpur and south India (Larsen, 1987d). A
total of eight Hesperiidae species were recorded in this study, also.

South India has a rich butterfly fauna probably due to climatic
conditions which differ substantially from the climate of Nagpur (central
India). The climate of the Blue Mountains is mild with low temperatures
and very high rainfall throughout the year (Lengerke, 1977) due to direct
exposure to the southwest monsoon. Secondly, the thick evergreen
forests of the Blue Mts. provide favorable conditions for a rich butterfly
fauna. The climate of Nagpur differs from that of the Blue Mts. with very
high temperatures and unusual rainfall. The variations in butterfly
fauna between these two regions are associated with climatic differences
of the regions. The butterfly fauna of Nagpur city is comprised of 61
species (Table 1), almost all of which were recorded at all the investigated
sites (Fig. 1). Maxima in butterfly populations were observed during late
July to January and in summer populations declined.

Fig. 1 Survey sites in Nagpur City.

29(1-2):157-160, 1990(91) 159

Table 1. Butterflies recorded in Nagpur city during 1987-1989

Papilionidae

Pachliopta aristolochiae (Fab.)

P. hector { L.)

Graph ium agamemnon{L.)

G. nomius (Esp.)

Papiiio polytes (L.)

P. demoleus (L.)

Pieridae

Anthocharis cardamines (L.)

Satyridae

Callerebia scanda (Moore)

Melanitis led a (L.)

Mycalesis malsarida (Btlr.)

M. nicotia (Flew.)

M. visala (Moore)

Orsotriaena medus (Fab.)

Ypthima huebneri (Kirby)

Y. philomela (Job.)

Catopsilla pomona (Fab.)

G. pyranthe (L.)

C. floret la (Fab.)

Gepora nerissa (Fab.)

Colotis danae (Fab.)

Delias each ar is (Dry.)

Pieris napi (L.)

Terias hecabe (L.)

T. laeta (Bdv.)

Danaidae

Danaus chrysippus{L.)

D. plexippus (L.)

Lycaenidae

Castalius rosimon (Fab.)

Catochrysops Strabo (Fab.)

Deudorix epijarbas (Moore)

Euchrysops pandava (Horsf.)

Freyeria trochilus (Freyer)

Jamides alecto (Felder)

Syntarucus plinius (Fab.)

Tajuria cippus (Fab.)

Tarucus alteratus (Moore)

Vacciniina optilete (Knoch.)

Zizina otis indica (Murray)

D. limniace (Cramer)

Euploea core (Cramer)

Acraeidae

Acraea violae (Fab.)

Nymphalidae

Apatura parisatis (Westw.)

Argyronome laodice (Moore)

Byblla ilithyia (Drury)

Euthalia nais (Forster)

Hypolimnas bolina (L.)

H. misippus (L.)

Junonia aim an a (L.)

J. lemon las (L.)

J. orithya (L.)

Neptis sappho (Pallas)

Phalanta phalanta (Drury)

Symbrenthia hypselis (Godt.)

Hesperiidae

Bi basis harisa (Moore)

Gelaenorrhinus ratna pulomaya( Moore)
Goladenia dan (Fab.)
lambrix salsala (Moore)

Odontoptilum angulata (Feld.)
Pelopidas mathias (Fab.)

Spialia galba (Fab.)

Taractrocera maevius (Fab.)

160

J. Res. Lepid.

Acknowledgements. The author is grateful to Dr. Jun Mitsuhashi, Tokyo Univer-
sity of Agriculture and Technology, Tokyo, Japan and Dr. C. D. Ferris, University
of Wyoming, USA for identifying the specimens. The author wishes to thank Dr.
T. R. New, La Trobe University, Bundoora, Australia for encouragement and
helpful suggestions and Dr. S. K. Raina, Dr. A. M. Khurad, Dr. M. M. Rai and Mr.
M. K. Rathod for much needed help during this study. The typing of the manuscript
by Mr. P. S. Mahulikar is gratefully acknowledged.

Literature Cited

Antram, C. B. 1924. Butterflies of India. Calcutta and Simla.

Brakefield, P. and T. B. Larsen. 1984. The evolutionary significance of dry and wet
season forms in polyphenic tropical Satyrinae. Biol. J. Linn. Soc. Lond. 22: 1-
12.

Evans, W. H. 1932. The identification of. Indian butterflies. Bombay Natural
History Society, Bombay.

Laithwaite, E., A. Watson and E. S. P. Whalley. 1975. The dictionary of butterflies
and moths in colour. Michael Joseph, London.

Larsen, T. B. 1987a. Butterflies in India. Oxford University Press, New Delhi.

. 1987b. The butterflies of the Nilgiri Mountains of southern India (Lepidoptera:

Rhopalocera). J. Bombay Nat. Hist. Soc. 84 (1): 26-54.

. 1987c. The butterflies of the Nilgiri Mountains of southern India (Lepidoptera:

Rhopalocera). J. Bombay Nat. Hist. Soc. 84 (2): 291-316.

. 1987 d. The butterflies of the Nilgiri Mountains of southern India (Lepidoptera:

Rhopalocera). J. Bombay Nat. Hist. Soc. 84 (3): 560-584.

. 1988. The butterflies of the Nilgiri Mountains of southern India (Lepidoptera:

Rhopalocera). J. Bombay Nat. Hist. Soc. 85 (1): 26-43.

Lengerke, J. J. von. 1977. The Nilgiris — weather and climate of a mountain area
in south India. Steiner Verlag, Wiesbaden.

Smart, P. F. 1975. The illustrated encyclopedia of the butterfly world. Salamander
Book, London.

Wynter-Blyth, M. A. 1943. Note on the Curetis species at Kallar. J. Bombay Nat.
Hist. Soc. 43: 671-672.

. 1944. The butterflies of the Nilgiris. J. Bombay Nat. Hist. Soc. 44: 536-549.

. 1946. Addenda and corrigenda to the butterflies of the Nilgiris. J. Bombay

Nat. Hist. Soc. 45: 613-615.

. 1947. Additions to the butterflies of the Nilgiris. J. Bombay Nat. Hist. Soc.

46: 736.

. 1957. Butterflies of the Indian region. Bombay Nat. Hist. Soc., Bombay.

Yates, J. A. 1935. The butterflies of the Nilgiri District. J. Bombay Nat. Hist. Soc.
38: 330-340.

. 1946. The butterflies of the Nilgiris — a supplementary note. J. Bombay Nat.

Hist. Soc. 46: 197-198.

Journal of Research on the Lepidoptera

29(1-2):161-171, 1990(91)

Interaction between Papilio hectorides
(Papilionidae) and four host plants (Piperaceae,
Rutaceae) in a southern Brazilian population.

Penz, C.M.1

Curso de Pos-Graduagao em Genetica, Dep. Genetica, Universidade Federal do Rio Grande
do Sul, Caixa Postal 1953, 90.001 Porto Alegre, RS, Brasil.

Araujo, A.M.

Dep. Genetica, Universidade Federal do Rio Grande do Sul, Caixa Postal 1953, 90.001 Porto
Alegre, RS, Brasil.

Abstract Larval development of Papilio hectorides , including time to
pupation, pupal mass and mortality, was measured on four different
plant species (two in Piperaceae and two in Rutaceae). The results were
compared with oviposition observations carried out during the spring-
summer period 1986 in one southern Brazilian population from a study
site in Sapucaia, RS, Brazil. The rank order for plant suitability
differed from the rank order of egg distribution in the field. The most
heavily used plant in the field ( Zanthoxylum hiemale) resulted in the
highest growth rates in the laboratory, but a infrequently used species
{Piper amalago ) was equally suitable for larval growth. Piper xylosteoides
produced a large proportion of males that had one extra molt; an extra
molt also occurred at low frequency in Zanthoxylum hiemale and Z.
rhoifolium.

Introduction

Comparisons of larval development on different host plants and egg
distribution in the field have not always demonstrated concordance
between rank order of female choice and plant suitability (see Singer,
1984 for a review). Larvae can feed on a wider range of plants than
females actually use for oviposition {e.g., Wiklund, 1973; Berenbaum,
1978) and they can grow well even on introduced plants that have been
incorporated into the diet {e.g. Scriber and Finke, 1978).

Larval development and survivorship in the field depend on a number
of factors. The quality of the food source, determined by nutrient avail-
ability and presence of secondary compounds, affects the time needed by
larvae for the completion of development as well as the mass achieved
at the end of larval stage (see Scriber & Slansky, 1981 for a review).
Because larvae are vulnerable to parasitoids and predators (Feeny et al. ,
1985) and fecundity is related to female body mass in a number of species
{e.g., Papilio polyxenes , Lederhouse, 1981; Lederhouse et al., 1982), host

1 Present address: Department of Zoology, University of Texas, Austin, Texas 78712, USA.

162

J. Res. Lepid.

plant utilization should evolve towards the production of the heaviest
possible pupae in the shortest possible time. Therefore, both time to
pupation and pupal mass are good measures of the suitability of a
particular host species for its herbivores. Physical characteristics of the
environment, such as solar radiation exposure (Grossmueller and
Lederhouse, 1985), also influence larval growth. Consequently, larval
developmental time, final mass and survival will depend on the ovipos-
iting females’ choice of the host plant and microhabitat.

According to this scenario, we expect selection to favor the utilization
of the most suitable plants, balancing between developmental and
survival optima. In order to investigate this question, we measured
larval growth rates of Papilio hectorides Esper, 1794 on four host species
found at a site in Sapucaia, RS Brazil. These measurements were used
to ask whether rank order of suitability of these plants is related to the
rank order in which females oviposit on them in the field.

In southern Brazil, Papilio hectorides is oligophagous on Piperaceae
and Rutaceae (see Biezanko, 1959a, b; Biezanko et al. , 1974 for host plant
records; eggs were also found on Piper mikanianum in Maquine, RS and
Moreira, Gramado, RS, pers. obs.). These butterflies lay small clusters of
1-8 round, orange eggs on the underside of the host plant leaves.

Methods and Materials

STUDY SITE AND HOST PLANTS

Field observations were conducted in a second growth area adjacent to Parque
Zoologico de Sapucaia, Sapucaia do Sul, state of Rio Grande do Sul, Brazil
(30°S,51°W) in the spring-summer period, 1986. The area had been deforested
and subsequently reforested with Eucalyptus trees. It was characterized by a
patchy secondary native flora mixed with Eucalyptus trees. The potential native
hosts present were Zanthoxylum hiemale, Z. rhoifolium, (Rutaceae); Piper
amalago , P. gaudichaudianum and P. xylosteoides (Piperaceae). Introduced
potential hosts were Citrus limon and C. reticulata (Rutaceae).

In order to evaluate the distribution of eggs in different host plants, we selected
three 400 m transects that included a representative sample of sunny and shady,
open and dense habitats. All species of Piperaceae and Rutaceae present in the
transects were considered potential hosts, individual plants were counted and
inspected for eggs. The eggs found were collected and transferred to the labora-
tory for measurements of larval growth rates. Each cluster was assigned to a
single plant species. Plant species frequency in the transects was estimated by
the percentage of individuals of a particular species relative to the others.

LABORATORY TESTS

In order to evaluate host plant suitability, we reared a total of 344 P. hectorides
larvae on the four native plant species on which eggs or larvae were found in the
field: P . xylosteoides (N=106), P. amalago (N=66), Z. hiemale (N=90) and Z.
rhoifolium (N=82). Measurements of (1) time to pupation, (2) mass and (3)
mortality in each instar and immature stages were taken. Larvae were placed
in 100ml plastic containers and kept at 25°C under constant light conditions.
Additional eggs for the trials were obtained from captive females, and these were

29(1-2):161-171, 1990(91)

163

distributed equally among the four plants tested. Food was changed daily, from
a store of leaves that was kept at 8°C for a period of up to one week. Larvae were
weighed on a Bosch precision balance (accuracy: lpg) within 24h of each molt and
pupae were weighed three days after pupation.

In order to determine the leaf water content of each host tested, 10 leaves of
each plant species were collected and weighed together fresh and again after
seven days at 45°C in a convection oven. The loss in water due to storage was
evaluated in a similar way, 10 leaves were weighed together fresh and again after
a week in our 8°C storage chamber.

Results

HOST PLANTS AND OVIPOSXTION RECORDS
Host plants of Papilio hectorides differed in growth form, leaf shape
and texture, and frequency in the transects at the study site (Table 1).
The Piper species differed greatly in frequency; P. xylosteoides and P.
gaudichaudianum were particularly abundant both in sunny and shady
sites. However, the nature of the leaves ofP. gaudichaudianum , highly
pilose and lignified, probably prevented its utilization by the ovipositing
females. Zanthoxylum rhoifolium (Rutaceae) individuals sampled in the
study area were almost always smaller thanZ. hiemale and grew mainly
in sunny sites along the transects, whereas Z. hiemale occured in both
sunny and shady conditions. The introduced Citrus limon and C. reticulata
were rare in the study area; only one well developed individual of C. limon
could be sampled in a shady site.

Except for Z. rhoifolium , leaf water content was similar among the
plants and the small differences observed between the species were not
likely to account for differences in nutrition efficiency (Table 1; see also
Scriber, 1979b). In addition, the water loss during storage was small (P.

Table 1. Comparison between the native potential hosts P. xylosteoides (PX), P.
amalago (PA), P. gaudichaudianum (PG), Z. hiemale (ZH), Z. rhoifolium (ZR)
concerning growth form, leaf shape and texture, leaf water content (%), relative
frequency (%), number of clusters, number of eggs and mean number of eggs per

cluster.

Plant

Growth

Form

Leaf Shape
and Texture

Water
Content (%)

Freq.

(%)

N

Clusters

N Eggs

eggs/

Cluster

ZH

tree

compound,

glabrous

76.8

4.5

15

90

6.0

PA

shrub

single,

glabrous

70.0

3.9

1

6

—

PX

herb

single,

glabrous

78.7

51.8

13

56

4.3

ZR

tree

compound,

pilose

57.8

2.0

0

—

PG

shrub

single,

pilose

__

37.8

0

164

J. Res. Lepid.

xylosteoides : 4.38%; P. amalago : 1.43%; Z. hiemale : 2.74%; Z. rhoifolium :
3.42%) and these fluctuations probably did not influence the nutrition
efficiency of the plants (Scriber, 1979b).

At the study site, eggs were found on three of seven potential hosts (see
Table 1 for information on number of eggs found on each host). No eggs
or larvae were observed on Piper gaudichaudianum. Larvae (N=3) were
found on Zanthoxylum rhoifolium. One larva and one egg (sterile) were
found on the introduced Citrus limon, but no utilization was observed for
C. reticulata in the study site. In addition, most of the eggs were found
in shady sites.

If we calculate a utilization index for each plant by dividing the number
of egg clusters found on it by its frequency in the transects, the following
rank order is obtained: Z. hiemale >P. xylosteoides >P. amalago >Z.
rhoifolium (on which larvae were observed). The distribution of egg
clusters in the field was different from the expected if the clusters were
laid randomly on the four effectively used native plants according to the
frequency that they were sampled (%2 test; yf - 91.04, %24[0001]= 18.467).
However, the data available is insufficient to ask if the differences were
due to growth form or size of the plant. Zanthoxylum hiemale received
more eggs than P. xylosteoides as a result of a difference in the number
of eggs per cluster (Table 1) (t-test; t=2.09, P=0.0233). We had no evidence
of missing eggs in the clusters collected in the field.

DEVELOPMENT AND GROWTH RATES

The quality of the host plant affected time to pupation; because larval
stage is a vulnerable period, the plant species that resulted in the fastest
development was considered the most suitable host (Table 2). Male and
female growth was pooled because their larval development time did not
differ significantly (Table 3). Time to pupation did not differ between Z.
hiemale andP. amalago , but we found that (1) larvae grew significantly
faster onP. amalago andZ. hiemale comparared toP. xylosteoides and (2)
on any of the three other species compared to Z. rhoifolium. Significant
differences were also found when the congeneric pairs P. xylosteoides vs.
P. amalago and Z. hiemale vs. Z. rhoifolium were compared (Table 4).
Thus, the rank order for host suitability based on time to pupation was
Z. hiemale - P. amalago > P. xylosteoides > Z. rhoifolium.

Final pupal mass differed among host plant species and was influenced
by sex (Table 2). However, in the sample raised onP. xylosteoides larvae
that experienced extra instars were heavier than the larvae that passed
through the normal number of instars (see Instar Number section).
Insects raised on Z. hiemale andP. amalago formed significantly heavier
pupae than those raised on P. xylosteoides and Z. rhoifolium , females
weighing more than males (Table 4, Table 5). We ranked the host plants
based on mean pupal weight as: for males, Z. hiemale > P. amalago > P.
xylosteoides > Z. rhoifolium ; and for females, Z. hiemale > P. amalago >
Z. rhoifolium > P. xylosteoides. However, significant differences between
means were found in just four out of 12 comparisons (Table 4).

29(1-2):161-171, 1990(91)

165

Table 2. Growth comparison of P. hectorides larvae raised on four host plants, t
represents the the calculated value for the t-test (two-tailed; * significant at a 0.05
level); L, larvae; P, pupae; T, total. The abreviations of the food plants are the same
as in Table 1 , PX5i and PX6i refer to individuals that went through five and six
instars, respectively, among the larvae raised on P. xylosteoides.

time to pupation (days)

pupal mass (mg)

mortality (%)

six instars

occurrence

plant

N

?

/

t

9

/

? /

L

P

T

9

/

? /

%

ZH

61

21.8

23.2

1.73 ns

22.4

746.8

835.9

13.1

15.2

27.1

1

0

1

1.6

PA

50

22.5

22.7

0.32 ns

22.6

721.0

763.2

11.9

11.9

22.4

0

0

0

—

PX

65

27.4

28.4

1 .30 ns

27.8

698.1

681.5

27.4

18.5

41.6

15

4

19

29.2

PX5i

46

26.2

27.8

1 .86 ns

27.0

673.6

650.3

PX6i

19

29.2

31.5

1.15*

29.7

744.2

853.3

ZR

44

28.8

31.2

1.22 ns

29.8

640.5

703.1

24.4

15.3

35.9

1

0

1

2.3

Table 3. Comparison of time to pupation obtained on four host plants (one-factor
ANOVA). Data on P. xylosteoides (PX) pool five instar larvae (5i) and six instar

larvae (6i).

Source

DF

SS

MS

F

between plants

3

2159.944

719.981

51.344 P = 0.0001

within plants

216

3028.892

14.023

total

219

5188.836

Table 4. Tukey multiple comparison tests of the means of the time to pupation (Tp), and
male and female pupal mass (Mp) on the four host plants. The abreviations of the food
plants are the same as in Table 1 . Asterisks represent significance at 0.05 (*), 0.01 (**)

and 0.001 (***) levels.

Comparison

Tp ?/

Mp ?

Mp /

PX vs. PA

10.442***

1.301

3.612

PX vs. ZH

11.441***

2.793

7.987***

PX vs. ZR

3.868*

3.050

0.791

PA vs. ZH

0.396

1.412

3.226

PA vs. ZR

13.163***

4.104*

2.476

ZH vs. ZR

14.122***

5.439***

6.061***

166

J. Res . Lepid.

Table 5. Comparison of male and female pupal masses obtained on the four host
plants tested (two-factor ANOVA). Data on P. xyiosteoides (PX) pool five instar
larvae (5i) and six instar larvae (6i).

Source DF

SS

MS

F

P

host plant 3

478987.36

159962.46

14.70

0.0001

sex 1

101613.50

101613.50

9.35

0.0025

plant vs. sex 3

91702.29

30567.43

2.81

0.0403

error 212

2303115.89

10863.57

Table 6. Comparison of development time and pupal mass between individuals
that passed through five (5i) or six (6i) instars in the sample raised on P.
xyiosteoides. Comparisons between 5i and 6i larvae on time spent in the first two
instars refer to the total number of larvae raised on this plant; for the comparison
of the time to pupation all the pupae that eclosed were used (? and /); and for
the comparison of the final mass, just males were used. Comparisons between
larvae raised on P. xyiosteoides and Z rhoifolium refer to males only. Time is
measured in days and mass in mg. t represents the the calculated value for the t-
test (two-tailed); asterisks, the significance at 0.05 (*), 0.01 (**) and 0.001 (***)
levels. The abbreviations of the food plants are the same as in Table 1 .

comparison

5i

N 0!

N

t

PX vs. PX

time spent in the 1st instar

3.9

58 4.5

21

2.67*

time spent in the 2nd instar

4.2

58 5.0

21

2.60*

time to pupation ? /

27.1

46 29.7

19

3.42 ***

mass ?

667.8

25 744.2

15

2.26*

comparison

PX6i

N ZR5i

N

t

PX vs. ZR

time to pupation ? /

29.2

15 28.8

25

0.29

mass ?

744.2

15 640.5

25

3.69 ***

In the larval stage, P. xyiosteoides and Z. rhoifolium caused twice the
mortality observed on P. amalago and Z. hiemale; pupal mortality was
similar on all host plants (Table 2).

P. hectorides diapauses in the pupal stage during the dry season. Most
of our test insects did not diapause, eclosing 18-17 days of pupation
(Table 7). Nevertheless, about 20% of the individuals in our experiment
spent longer periods in the pupal stage (the longest was 158 days). We
cannot determine the cause of the lengthening of the pupal stage because
the larvae were kept under controlled temperature and light. However,

29(1-2):161-171, 1990(91)

167

Table 1. Comparison of mean larval duration and median pupal duration (days)
between the four plants. The abreviations of the food plant are the same as in

Table 1 .

plant

larval instars

pupa

1st

2nd

3rd

4th

5th

6th

ZH

4.0

3.4

3.5

4.2

7.7

—

17

PA

3.3

3.1

3.8

4.5

8.0

—

17

PX 5i

3.9

4.2

4.5

5.7

9.2

—

17

PX 01

4.5

5.0

4.2

4.2

4.8

7.7

17

ZR

4.1

3.5

4.5

5.9

12.0

—

10

Table 8. Comparison of mean larval mass one day after each molt (mg) between
larvae on the four plants. Mean egg weight equals 1 .3 mg (N - 266). The
abbreviations of the food plant are the same as in Table 1 .

plant

instars

2nd

3rd

4th

5th

0th

ZH

8.7

34.4

110.1

477.8

—

PA

9.1

37.0

129.9

554.9

—

PX 5i

7.5

27.5

100.1

511.2

___

PX 6i

5.9

21.3

04.4

214.7

710.0

ZR

9.2

31.2

115.0

430.3

—

this diapause was not correlated with host plant species, sex, number of
instars in the larval period or relatedness of offspring from captive
females.

A mean of the time spent in each instar and larval mass after each molt
is presented in Tables 7 and 8.

INSTAR NUMBER

Although the majority of the larvae passed through five instars, some
individuals passed through an extra molt. The proportion of six-instar
larvae differed among food plants and sexes (Table 2); P. xylosteoides was
the plant which had the highest proportion of six-instar individuals and
the occurrence of one extra instar was more common among males.

The occurrence of one extra instar increased the time to pupation of the
individuals that fed on P. xylosteoides ; the increase in final mass placed
six-instar individuals raised on this plant species near to that on P.
amalago and Z. hiemale. Since first and second instars were significantly

168

J. Res. Lepid.

longer in six-instar larvae than in five-instar ones (Table 6), total time to
pupation was also extended (Table 6). This elongation in the larval stage
may explain why males with an extra instar were significantly heavier
than five instar-ones (Table 6), leading to a significant interaction
between food and sex in the comparison of mean pupal weight (Table 5).

In attempt to sort out the effects of the number of molts and time to
pupation on the final mass, we compared six-instar males raised on P.
xylosteoides with five-instar males raised onZ. rhoifolium. The duration
of the larval stage, that time when larvae can feed, was similar between
the two groups; mass accquisition, however, differed (Table 6). In this
particular case, the time that larvae had the opportunity to feed was not
the limiting factor in terms of mass accquisition. However, it was not
possible to separate the effects of the number of molts from the difference
in food quality, since the two groups under comparison fed on different
plant species.

Discussion

Growth rate, pupal mass, fecundity and larval survival all seem to
positively correlate with efficiency of host use (Wiklund, 1973; Courtney,
1981; Lederhouse et al ., 1982). Therefore, it is expected that natural
selection will favor insects that oviposit on potential hosts in the rank
order of their suitability (Singer, 1971; Jaenike, 1978; Rausher, 1985).
However, there are examples of field observations which fail to show this
expected ranking (Chew, 1977; Courtney, 1981, 1982).

In the population investigated, the rank orders for plant suitability and
number of eggs found differed. Larvae developed equally well on a
heavily used host (Z. hiemale) and on a plant that received few eggs (P.
amalago). The abundant P. xylosteoides , the third-ranked host in terms
of larval development, was also heavily used.

Most of the eggs were found on two unrelated plants that differed in
growth form, leaf shape and abundance and a preference for shady sites
was observed. Since Z. hiemale andP. amalago were similarly abundant
and distributed both in sunny and shady sites, the difference in utiliza-
tion between these two plant species cannot be attributed to unequal
probability of encounter. Their distinct growth form, leaf size and shape,
and, probably, secondary chemicals (see Feeny et al ., 1983; Baldwin and
Schultz, 1988) are stronger discrimination cues for the ovipositing
females and may cause differences in acceptance. P. xylosteoides , espe-
cially abundant in shady sites, was the second ranked host in terms of
number of eggs found. The difference in number of clusters between P.
xylosteoides and Z. hiemale was not as pronounced as the difference in
absolute number of eggs received, which resulted from a significant
difference in the average number of eggs per cluster. The cause(s) of the
difference in mean cluster size between these two host species remains
to be investigated, but it is possibly a response to intrinsic qualities of
each of the plants (see Pilson and Rausher, 1989 for a comparison).

29(1-2): 161- 171, 1990(91)

169

Zanthoxylum rhoifolium was mostly distributed in sunny sites, which
may have an influence on its low utilization, assuming the females to
have a preference for shady sites.

The estimates of growth rate demonstrated that unrelated plants can
be equally suitable for larval development ofP. hectorides. Zanthoxylum
hiemale and P. amalago resulted in the fastest development, heaviest
pupae and lowest mortality among the four plants tested. Contrasts
between congeneric pairs showed remarkable differences in suitability.
The differences in the leaf water content and texture may have some
influence in the differential performance observed between the species of
Zanthoxylum, but the same explanation is not valid for the two Piper
species.

In those plants that supported less efficient growth, P. xylosteoides and
Z. rhoifolium, the time needed by a larva for the completion of its
development was longer and the final mass was smaller than the
equivalent values obtained for Z. hiemale and P. amalago. Thus, in
general, additional time in the larval stage could not counterbalance the
poorer quality of the two low ranking hosts in terms of larval develop-
ment. However, the occurrence of an extra instar, with a consequent
increase on the time that larvae spent feeding, increased the final mass
of the individuals raised on P. xylosteoides to a value comparable to the
ones obtained for the most suitable hosts, Z. hiemale and P. amalago.

The occurrence of extra molts was related to host plant species and
butterfly sex. Larvae raised on P. xylosteoides that underwent an extra
instar (about 30% of the total) showed a significant increase in mass. The
physiological event of passing through an extra instar seems to be acting
in conjunction with the elongation of the larval period, increasing the
pupal mass as a result. Interestingly, almost all the six-instar individu-
als obtained in this experiment were males. Because the extra molt tends
to increase the final mass, and also because body mass is related to
fecundity (Lederhouse et al., 1982), we would expect females to be more
frequent among individuals that passed through six instars. Scriber
(1979a) reported the occurrence of extra molts in larvae that experienced
a poor diet, but he did not include individual records on sex, time to
pupation, or pupal mass.

The observations reported here suggest that diet breadth in the
population studied is constrained by females oviposition behavior; the
larvae can perform well on a plant that has been underexploited by
females in the field. However, growth rate estimates were done under
controlled conditions in the laboratory, preventing us from evaluating
other ecological factors that can play an important role in larval success
and survival in the field {e.g., differential predation of eggs or larvae
according to their location or host, Dempster, 1984; Grossmueller and
Lederhouse, 1985). The observed micro habitat preferences for oviposi-
tion, the non-utilization of a potentially suitable plant and the observed
differences in the number of eggs per cluster between the two most

170

J. Res. Lepid.

heavily used plants in the field are points that demand deeper investiga-
tion.

Acknowledgements. Our thanks go to Phil DeVries, Larry Gilbert, Janet Lanza,
Bob Lederhouse, Mike Singer and Bob Srygley and for valuable criticism and
suggestions that very much improved the manuscript. Thanks also to Keith Brown
Jr. for criticism on C.M.P. Master Thesis. CNPq fellowship to C.M.P. made this
research possible.

Literature Cited

Baldwin, I.T. & J.C. Schultz, 1988. Phylogeny and the patterns of leaf phenolics
in gap- and forest-adapted Piper and Miconia understory shrubs. Oecologia,
75:105-109.

Berenbaum, M. 1978. Toxicity of a furanocoumarin to armyworms: a case of
' biosynthetic escape from insect herbivores. Science, 201:532-534.

Biezanko, C.M. 1959a. Contribuifao ao conhecimento da fisiografia do Rio
Grande do Sul, la - Papilionidae da zona missioneira. Arquivos de Entomologia,
ser.B, 12 pp.

1959b. Contribuifao ao conhecimento da fisiografia do Rio Grande do Sul,

la - Papilionidae da zona sudeste do Rio Grande do Sul. Arquivos de
Entomologia, ser.B, 17 pp.

Biezanko, C.M., A. Rufinelli & D. Link 1974. Plantas y otras sustancias
alimenticias de las orugas de los lepidopteros uruguayos. Revista do Centro
de Ciencias Rurais, 4(2):107-148.

Chew, F. 1977. Coevolution of Pierid butterflies and their cruciferous foodplants.

2- The distribution of eggs in potential foodplants. Evolution, 31:568-579.
Courtney, S.P. 1981. Coevolution of Pierid butterflies and their cruciferous
foodplants. 3- Anthocaris cardamines (L.) survival, development and
oviposition on different host plants. Oecologia, 51:91-96.

1982. Coevolution of Pierid butterflies and their cruciferous foodplants. 4-

Crucifer apparency andAnthocaris cardamines oviposition. Oecologia, 52:258-
265.

Dempster, J.P. 1984. The natural enemies of butterflies, pp. 97-104 In: Vane-
Wright R.I. & P.R. Ackery (Eds.) The Biology of Butterflies. Academic Press,
Orlando.

Feeny, P., W.S. Blau & P.M. Kareiva. 1985. Larval growth and survivorship of
the black swallowtail butterfly (Papilio polyxenes) in central New York
(USA). Ecological Monographs, 55:167-188.

Feeny, P., L. Rosenberry & M. Carter. 1983. Chemical aspects of oviposition
behavior in butterflies, pp. 27-76 In: Ahmad, S. (Ed.) Herbivore insects: host-
seeking behavior and mechanisms. Academic Press, New York.
Grossmueller, D.W. & R.C. Lederhouse. 1985. Oviposition site selection: an aid
to rapid growth and development in the tiger swallowtail butterfly Papilio
glaucus. Oecologia, 66:68-73.

Jaenike, J. 1978. On optimal oviposition behavior in phytophagous insects.

Theoretical Population Biology, 14:350-356.

Lederhouse, R.C. 1981. The effect of female mating frequency on egg fertility in
the black swallowtail, Papilio polyxenes asterius Stoll (Papilionidae). Journal
of the Lepidopterist’s Society, 32:145-159.

29(1-2). 161- 171, 1990(91)

171

Lederhouse, R.C., MLB. Finke & J.M. Scriber. 1982. The contributions of larval
growth and pupal duration to protandry in the black swallowtail butterfly,
Papilio polyxenes. Oecologia, 53:296-301.

Pilson, D. & ML!.), Rausher. 1989. Clutch size adjustment by a swallowtail
butterfly. Nature, 333:361-363.

Rausher, M.D. 1985. Variability for host preference in insect populations:
mechanistic and evolutionary models. Journal of Insect Physiology, 31:873-
889.

Scriber, J.M. 1979a. The effects of sequentially switching foodplants upon
biomass and nitrogen utilization by polyphagous and stenophagous Papilio
larvae. Entomologia Experimentalis et Applicata, 25:240-252.

1979b. Effects of leaf-water suplementation upon post-ingestive nutritional

indices of forb-, shrub-, vine-, and tree feeding lepidoptera. Entomologia
Experimentalis et Applicata, 25:240-252.

Scriber, J.M. & P. Feeny. 1979. Growth of herbivorous caterpillars in relation to
feeding specialization and the growth form of their food plants. Ecology,
60:829-850.

Scriber, J.M. & M. Finke. 1978. New food plant and oviposition records for the
eastern black swallowtail, Papilio polyxenes , on an introduced and native
umbellifer. Journal of the Lepidopterist’s Society, 32:236-238.

Scriber, J.M. & F. Slansky. 1981. The nutritional ecology of immature insects.
Annual Review of Entomology, 26:183-211.

Singer, M.C. 1971. Evolution of food plant preferences in the buttefly Euphydryas
editha. Evolution, 25:383-389.

1984. Rutterfly-hostplant relationships: host quality, adult choice and

larval success, pp. 81-87 In: V ane- Wright R.L & P.R. Ackery (Eds.) The
Biology of Butterflies. Academic Press, Orlando.

Wiklund, C. 1973. Host plant suitability and the mechanism of host selection
in larvae of Papilio machaon. Entomologia Experimentalis et Applicata,
16:232-242.

Journal of Research on the Lepidoptera

29(1-2):172, 1990(91)

Book Review

RAINFORESTS: A GUIDE TO RESEARCH AND TOURIST FACILITIES AT
SELECTED TROPICAL FOREST SITES IN CENTRAL AND SOUTH AMERICA.
James L. Castner. 1990. 380+xxxiv pages. Feline Press, P. O. Box 7219,
Gainesville, Florida 32605. $21.95 plus 1.50 shipping in U. S. ISBN 0-9625150-
2-7.

Although I am a confirmed resident of temperate regions, the tropics command
a great deal of concern for me. The horror of tropical forest destruction in the
name of greed and accommodation to the exploding world population is a familiar
theme. Comes now a charming little paperback guide, a sort of “Rainforests on
$5.00 a Day” for biologists. Although actually more than $5 a day, the ambiance
is incomparable for adventure. The reading is delightful and the book gives a
great deal of detail on how to get around to the most friendly and fruitful places
across the bulk of tropical America. I only wish Castner had provided this work
prior to my first venture into the jungles, which was more of a bumble than an
adventure. Every one of our readers with the slightest curiosity for the area
covered should have this volume.

Peter Raven, the eminent tropical conservation botanist, writes the foreword
in which he urges everyone to make the trip while there is something left and for
the privilege of observing the world’s richest palette of biodiversity before most
of it becomes a mere memory. Ironically, our children may be unable to make the
trip, not because of the loss of the forests themselves, but because the results of
deforestation will probably magnify the upheavals in our global economic
systems by way of climatic effects.

On the other hand you do not need to take the tropical grand tour in person, as
Castner vicariously provides a convincing sense of the places which would be on
the itinerary. The facilities and sites of Peru, Ecuador, French Guiana, Venezu-
ela, Trinidad, Costa Rica and Panama are described in detail. Costs, generally
cheap, are given as well as all the contacts you need in order to make arrange-
ments. Natural history notes are sparse, the bulk of discussion being concerned
with the details of travel and accommodations. There is a very complete list and
description of books and journals of the area, and all relevant organizations are
listed and described. The latter include important activist groups.

For the highest adventure, of course, you just take a plane to places that sound
relevant to your interests and start conversations. Most natives, from Kabul to
Manaos, are friendly to travellers and will provide anecdotes that flavor the trip.
Such was the case for a collector friend who quickly learned there is more than
one kind of “butterfly” in Spanish. Upon asking “<(,D6nde estan las mariposas?”
he was told to walk two blocks down* one block over, and up the stairs. These
“painted ladies” awaited customers.

One fault with the book is a rather steep tab, but if sales are limited, this pricing
is necessary.

R. H. T. Mattoni, 9620 Heather Road, Beverly Hills, CA 90210, USA.

INSTRUCTIONS TO AUTHORS

Manuscript format; Two copies must be submitted, double-spaced, typed, with wide
margins. Number all pages consecutively. If possible italicize rather than underline scientific
names and emphasized words. Footnotes are discouraged. Do not hyphenate words at the right
margin. All measurements must be metric. Time must be cited on a 24-hour basis, standard
time. Abbreviations must follow common usage. Dates should be cited as: day- Arabic numeral;
month-Roman numeral; year- Arabic numeral (ex. 6. IF. 1384). Numerals must be used for ten
and greater e.g. nine butterflies, 12 moths.

Electronic submission: The Journal is now being produced via desktop publishing, allowing
much shorter publication times. Although typewritten manuscripts are acceptable, those
submitted on computer disc are highly preferred. After being notified of your paper's
acceptance, submit either a Macintosh or IBM disc version. Include on your disc both the fully
formatted copy from your word processing program and a text-only (ASCII) copy. The two most
preferred formats are Microsoft Word for the Macintosh and either Microsoft Word or Word
Perfect for the IBM, although translation utilities will allow conversion from most formats. Put
returns only at the ends of paragraphs, not at the end of each line. Use one tab to indent each
paragraph. Even if your printer is incapable of outputting italics, please specify italics rather
than underline in your disc copy. Please note any special characters that are used in either the
body of the text or the tables (e.g. e, ii, °, §, p, 6 , 9 ).

Title material; All papers must include the title, author's name, author's address, and any
titular reference and institutional approval reference. A family citation must be given in
parenthesis (Lepidoptera: Hesperiidae) for referencing.

Abstracts and Short Papers; All papers exceeding three typed pages must be accompanied
by an abstract of no more than 300 words. Neither an additional summary nor key words are
required.

Name citations and Systematic Works; The first mention of any organism should include
the full scientific name with unabbreviated author and year of description. There must be
conformity to the current International Code of Zoological Nomenclature. We strongly urge
depositing of types in major museums, all type depositories must be cited.

References; All citations in the text must be alphabetically listed under Literature Cited in
the format given in recent issues. Abbreviations must conform to the World List of Scientific
Periodicals. Do not underline or italicize periodicals. If four or less references are cited, please
cite in body of text not in Literature Cited. For multiple citations by the same author(s), use
six hyphens rather than repeating the author's name.

Tables; When formulating tables, keep in mind that the final table will fill a maximum space
of 11.5 by 19 cm either horizontally or vertically oriented. Number tables with Arabic
numerals. When submitting tables on disc, use tabs between columns rather than multiple
spaces.

Illustrations; Color can be submitted as either a transparency or print, the quality of which
is critical. Black and white photographs should be submitted on glossy paper, and, as with line
drawings, must be mounted on stiff white cardboard. Authors must plan on illustrations for
reduction to page size. Allowance should be made for legends beneath, unless many consecutive
pages are used. Drawings should be in India ink. Include a metric scale. Each figure should
be cited and explained as such. Each illustration must be identified by author and title on the
back. Indicate whether you want the illustration returned at your expense. Retain original
illustrations until paper is accepted. Legends should be separately typed on pages entitled
"Explanation of Figures." Number legends consecutively with separate paragraph for each
page of illustration.

Review; All papers will be read by the editor(s) & submitted for formal review to two referees.

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

Volume 29 Number 1-2 Spring 1990(1991)

IN THIS ISSUE

Date of Publication: December 31, 1991

Hilltopping by the Red Admiral Butterfly: Mate Searching Alongside Congeners 1
William D. Brown and John Alcock

Electrophoretic Studies in the Genus Melanargia Meigen, 1828 11

(Lepidoptera: Satyridae)

Paola Mensi, Aldo Lattes, Luigi Cassulo, RobertaCinti, Emilio Balletto

Hybridization of the Brazilian Papilio (Pyrrhosticta) (Section V) with the North 21
American Papilio (Pterourus) (Section III)

J. Mark Scriber, Robert C. Lederhouse, and Keith S. Brown Jr.

Significant Additions to the Butterflies of the Trinity Alps andMount Eddy, 33

Northern California

Arthur M. Shapiro

Heritable Color Variants in Automeris io (Saturniidae) 37

Thomas R. Manley

Patterns of geographic variation and evolution in polytypic butterflies 54

Paul C. Hammond

Cryptic larval polychromatism in Rekoa marius Lucas and R. palegon Cramer 77

(Lycaenidae: Theclinae)

Ricardo Ferreira Monteiro

On Pieris ( Artogeia ) marginalis macdunnoughii Remington (Pieridae) 85

S. R. Bowden

An Annotated List of Lepidopterological Journals 92

Gerardo Lamas

The Butterflies (Lepidoptera) of the Tuxtlas Mts., Veracruz, Mexico, 105

Revisited: Species-Richness and Habitat Disturbance
Robert A. Raguso and Jorge Llorente-Bousquets

Territorial Hilltopping Behavior of Three Swallowtail Butterflies 134

(Lepidoptera, Papilionidae) in Western Brazil
Carlos E. G. Pinheiro

Ball Mountain Revisited: Anomalous Species Richness of a Montane Barrier Zone 143
Arthur M. Shapiro

Butterflies from Nagpur City, Central India (Lepidoptera: Rhopalocera) 157

T. N. Pandharipande

Interaction between Papilio hectorides (Papilionidae) and four host plants 161

(Piperaceae, Rutaceae) in a southern Brazilian population
C. M. Penz and A. M. Araujo

Book Review 172

Cover Illustration: Photograph by J. Mark Scriber.

IE JOURNAL
RESEARCH

ON THE LEPIDOPTERA

Volume 29
Number 3
Fall 1990(1991)

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

ISSN 0022 4324

Published By:

The Lepidoptera Research Foundation, Inc.
9620 Heather Road
Beverly Hills, California 90210
(213) 274-1052

Founder:

William Hovanitz

Editorial Staff:

Rudolf H. T. Mattoni, Editor
Scott E. Miller, Assistant Editor

Associate Editors:

Emilio Balletto, Italy

Henri Descimon, France
Philip DeVries, U.S.A.

Thomas Emmel, U.S.A.
Lawrence Gall, U.S.A.
Hansjuerg Geiger, Switzerland
Otakar Kudrna, Germany
Arthur Shapiro, U.S.A.

Atuhiro Sibatani, Japan
Karel Spitzer, Czechoslovakia

Manuscripts and Notices Material may be sent to the Editor at:
9620 Heather Road, Beverly Hills, CA 90210 (213) 274-1052

The JOURNAL is sent to all members of the FOUNDATION.

Classes of Membership:

$ 20.00 year (vol.)

Regular (Individual)
Contributing

Student/Retired-Worldwide
Subscription Rate/Institutions
Life

$ 30.00 or more, year (vol.)

$ 18.00 year (vol.)

$ 28.00 year (vol.)
$ 250.00

STATEMENT OF OWNERSHIP AND MANAGEMENT

THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA is published four times a year by THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The office of publication and the general business
office are located at 9620 Heather Road, Beverly Hills, California 90210. The publisher is THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The Editor is R. H. T. Mattoni at the above address.
The Secretary-Treasurer is Leona Mattoni at the general business office. All matters pertaining to
membership, dues, and subscriptions should be addressed to her, including inquiry concerning mailing,
missing issues, and change of address. The owner is THE LEPIDOPTERA RESEARCH FOUNDA-
TION, INC. , a non-profit organization incorporated under the laws of the State of California in 1965. The
President is R. H. T. Mattoni, the Vice President is John Emmel, the Secretary-Treasurer is Leona
Mattoni. The Board of Directors is comprised of Barbara Jean Hovanitz, Leona Mattoni, and R. H. T.
Mattoni. There are no bond holders, mortgages, or other security holders.

Printed by Art Printing Works Sdn. Bhd., 29 Jalan Riong, 59100 Kuala Lumpur, Malaysia.

Journal of Research on the Lepidoptera

29(3):173-194, 1990(91)

Evolutionary Ecology of Sympatric Catocala Moths
(Lepidoptera: Noctuidae)

I. Experiments on Larval Foodplant Specificity

Lawrence F. Gall

Entomology Division, Peabody Museum of Natural History, Yale University, New Haven,
CT 06511 USA

Abstract. Larval foodplant specificities are presented for 31 species of
Nearctic Catocala moths whose larvae feed on trees in the plant
families Juglandaceae, Fagaceae, and Salicaceae. Arena tests deter-
mined preferences when larvae could choose on which plants to feed,
and field rearing experiments determined survival and growth rates in
no-choice situations. In arenas, larvae of the Juglandaceae, Fagaceae
and Salicaceae feeders ate only foods from their own plant family.
Taxonomic divisions within each plant family defined further bound-
aries to foodplant acceptability. Thus, among the Juglandaceae feed-
ers, 2 species ate walnuts ( Juglans ), 2 ate pecans (section §Apocarya of
Cary a), and the remaining 14 ate hickories sensu strictu (section
§Eucarya of Carya). All the Carya section §Eucarya feeders preferred
shagbark hickory, Carya ovata , and 6 of the 14 species specialized on
it. Among the Salicaceae feeders, 1 species ate willows ( Salix ) and 6 ate
poplars (Pop ulus), Foodplant preferences were not crisply defined in
the 5 Fagaceae feeders. Larvae usually selected the favored foodplant
as a resting site. Preferences broadened as larvae matured, and the
Juglandaceae and Salicaceae feeders preferred young leaves over old
leaves. In nearly all cases, survival was highest and growth rates most
rapid in the field on the foodplants preferred in arena tests.

Introduction

The Holarctic noctuid moth genus Catocala Schrank contains slightly
over 200 species, distributed equally across the Nearctic and Palearctic
temperate zones. The genus is most atypical for noctuids in showing
great species diversity coupled with highly restricted larval foodplant
use — larvae of individual Catocala species eat only one or several closely
related plant genera, and the recorded foodplant breadth of the genus
spans but nine dicotyledonous plant families. Even this liberally esti-
mates foodplant breadth among Catocala, as 137 of the 161 species whose
life histories are currently known feed only on plants in the families
Juglandaceae, Fagaceae, Salicaceae, or Rosacea e, There are large arrays
of sympatric, synchronic, closely related species feeding in each of these
four plant families, and hence the Catocala are well suited for studying
the phylogenetic history of foodplant exploitation (note that the reported
high species diversity is not in question, since survey collecting for the
moths and the species-level taxonomy of the genus have been amply
studied: see Hampson, 1912; Barnes & McBunnough, 1918; Forbes,
1954; Sargent, 1976, 1977).

174

J. Res. Lepid.

This is the first of three baseline data papers addressing the foodplant-
linked ecologies of Catocala moths, and covers experiments on larval
foodplant specificity of 31 Nearctic Catocala species whose larvae feed on
Juglandaceae, Fagaceae, Salicaceae, and Rosaceae. The following two
articles in this journal issue examine distributions of wild larvae among
their foodplants and the oviposition preferences of the female moths; a
related paper has covered larval feeding habits within individual foodplant
trees (Gall, 1987). The data in these articles help circumscribe the limits
of potential foodplant acceptability for the major groups of Catocala, and
form the backdrop against which questions about foodplant use in
ecological and evolutionary time by the same Catocala species can be
appropriately framed.

Because analysis of the phylogenetic history of Catocala foodplant use
is made easier by simultaneously comparing results from all these
studies, I have extracted parts of what would traditionally comprise the
Discussion sections from the three articles in this series, and will be
integrating these into a comparative treatise on the subject. Toward that
same goal, a separate series of concurrent articles is addressing taxo-
nomic and nomenclatural issues in the genus (see Gall & Hawks, 1990;
our taxonomic monograph of all the Nearctic Catocala species is also in
the late stages of preparation, to be published in a forthcoming “Moths
of North America North of Mexico” fascicle).

Materials and Methods

A BIOLOGICAL PROFILE OF CATOCALA

All Catocala are univoltine, with adults flying in mid to late summer and
hibernal diapause as eggs. The adults have cryptic fore wings which conceal
boldly patterned hindwings; the interactions of the adults with avian predators
have been the topic of many previous studies by Sargent and his coworkers (e.g.,
Sargent, 1976, 1977; Schlenoff, 1985). Eggs are laid singly or in clumps under
exfoliating bark, or in crevices on tree trunks and branches. In northeastern
North America, most species commence feeding as larvae during May. Egg hatch
can be synchronous or highly staggered, but is constant, within any given
Catocala species, the degree of staggering being closely linked to the foliating
schedules of the different larval foodplants (Schweitzer, 1982; Schweitzer and
Gall, in preparation).

The larvae . are semiloopers, usually with highly cryptic morphologies that
match the twigs and bark which serve as resting sites during the daytime. Larvae
generally feed at night. Prior to this research, the wild larvae had been presumed
to be mildly oligophagous among related plant genera within single plant
families. Foodplant breakdowns for the 108 Nearctic species which I currently
plan to recognize in the forthcoming taxonomic monograph are as follows:
Fagacaeae, 26; Juglandaceae, 25; Salicaceae, 23; Rosaceae, 15; Leguminoseae, 7;
Ericaceae, 4; Myricaceae, 4; Tiliaceae, 1; not yet recorded, 3 (which with virtually
no question will prove to be 1 Fagaceae, 1 Juglandaceae, and 1 Salicaceae).

FOODPLANT ARENA PROTOCOL

The technique employed here was the “foodplant arena,” in which both
potentially suitable and presumably unsuitable foods are offered to larvae, which

29(3): 173-194, 1990(91)

175

then must make a choice for feeding (see Gall, 1987). Small branches from young
(2-5 cm dbh) foodplant trees are collected late each afternoon, ca. 1-3 weeks after
bud burst in spring. Near dusk (the time at which Catocala larvae become active
and begin feeding), pieces of leaves of different plant species are clipped from
their petioles and taped by one edge to a paper strip (food choice items being
ordered at random along the strip), and the strip coiled and placed flush against
the wall of a small glass jar. Several young larvae are then introduced to the
bottom of the jar (variation in foodplant preference among individual larvae of
the same Catocala species is not pronounced, and the total amount of leaf
material in each choice item is more than the maximum that the several larvae
can potentially consume; Gall, 1987). Strips are removed the following morning,
and larval resting locations and the amount of different food items consumed are
used to calculate preference.

Two types of control foods were offered in each arena. Type I controls were
plant species from a Catocala foodplant group other than the group being
examined: red/scrub oak or lombardy poplar for the Juglandaeeae-feeding
Catocala ; shagbark/pignut hickory or lombardy poplar for the Fagaceae feeding
Catocala ; and red/scrub oak or shagbark/pignut hickory for the Salieaceae-
feeding Catocala . Type II controls were “entirely inedible” plant species — ones
not closely related to present foodplants of any Catocala. The Type II control in
arenas was typically plantain ( Plantago major L.).

For each arena test, the area of food consumed per choice item was expressed
as a percentage of the total area consumed, preserving relative between-choice
differences while assigning unit weight to each arena test. For first instars, the
amounts eaten were measured using a dissecting microscope fitted with an
ocular grid. For partially grown larvae, a clear plastic ruled sheet was used.

INTERVENING FACTORS IN FOODPLANT ARENAS

Problems typically confronting foodplant preference studies include, among
others: (1) the geometry of how to present food choices; (2) how to minimize the
effects of geo- and phototaxis; and (3) how to insure that larvae are aware of and/
or have “tested” all available food items as they are beginning to feed. For many
lepidopteran larvae, these present serious methodological obstacles, and, in
general, the more polyphagous and/or sedentary a species one tests, the more
significant these obstacles are likely to be. A brief summary of these variables
as they relate to Catocala is relevant.

In all instars, Catocala larvae are extremely active — within moments of
introducing them to a foodplant arena, they typically traverse the entire perim-
eter of the jar. When placed in arenas during the daytime, the larvae eventually
settle down after a few minutes, on or near a food item, and rest. When introduced
at dusk, this initial period of rapid movement also includes mandibular contact
and/or momentary chewing on leaves encountered. Again, after a period of
usually less than five minutes, the larvae have passed several times around the
arena, and start to settle and feed (rather than rest). The distribution of larvae
in arenas remains largely stable from that settling period on through the next
morning, when arenas are disassembled.

Most Catocala larvae are positively phototactic to some degree. I have always
found mature larvae to be somewhat less active during the day than first instars
(see Wojtusiak, 1979, for diel locomotor actograms of Catocala fraxini L. larvae
of varying ages). Starting arena tests at dusk or night, in dark rooms, largely
eliminated problems with phototaxis (all laboratory preference tests were con-

176

J. Res. Lepid.

ducted under uncontrolled, but essentially natural photoperiod). Note that
Catocala larvae become active and commence feeding at true dusk even in fully
lit rooms with no windows.

In initial trials, the geometry of how foods were presented had no discernible
effect on larval foodplant preference. To test this, leaves were presented hanging,
as above, or flat on the bottom of the arena. Foodplant preferences were the same
in three such paired tests (n=4 arenas per test) for the Salicaceae-feeding
Catocala cara Guenee, and the Juglandaceae -feeding Catocala neogama Smith
and Catocala retecta Grote (p>.25 for each by ANOVA). Foods were presented
hanging in all subsequent trials.

FIELD REARING PROCEDURES

To assay Survival and development rates on foodplants in the field, I confined
larvae to large rearing exclosures (“sleeves”) on branches of different foodplants.
Sleeves were made of either lightweight muslin or fine nylon netting, and varied
in size depending upon the number and age of larvae confined in them. Equal
numbers of newly hatched, unfed larvae were placed in sleeves on suitable
foodplants, the genetic background of larvae for each rearing experiment being
controlled by splitting the progeny of a single female among the sleeves.
Environmental differences were controlled, insofar as possible, by selecting
sleeving sites at which all foodplants for an experiment were growing in
immediate proximity. Larval instar was used as the growth index, and was
assayed by using both head capsule width and the instar-specific morphological
patterns of Catocala larvae. Each instar was further subdivided into an early and
late stage, which was assayed using body size.

The primary research site for larval rearing was West Rock Park, three miles
west of the Yale University campus. This is an open-canopied, xeric to mesic
igneous trap-rock ridgetop habitat (Gall, 1987). West Rock is near the northern
limit for many elements of more southerly biota and in addition contains the
usual fauna and flora of southern New England broadleaf forests. The canopy is
dominated by oak, hickory, and ash, with a sparse understory of rosaceous,
ericaceous, and other shrubs.

It should be noted that when sleeve rearings were initiated in 1980, 1 began
measuring development rate as the number of days elapsed from first sleeving
until pupation. In 1980, and especially 1981, the essentially total defoliation of
the West Rock study sites by the introduced pest gypsy moth ( Lymantria dispar
L.) terminated most sleeve experiments before the Catocala larvae could success-
fully pupate, and re-sleeving was impossible. Hence, the measure of develop-
ment rate was soon switched to larval instar at the time of sleeve census.

CATOCALA AND FOODPLANT SPECIES STUDIED

Arena and rearing tests were carried out for 19 J uglandaceae-feeding Catocala
species: angusi Guenee, dejecta Strecker, flebilis Grote, epione Drury, habilis
Grote, insolabilis Guenee, judith Strecker, lacrymosa Guenee, luctuosa Hulst,
nebulosa Edwards, neogama, obscura Strecker, palaeogama Guenee, piatrix
Grote, residua Strecker, retecta Grote, subnata Grote, ulalume Strecker, and
vidua Smith. Hickories ( Carya Nutt.) and walnuts ( Juglans L.) are the foodplants
of these Catocala. Because the generic abbreviation (“C.”) is the same for both
Catocala and Carya, by convention I drop the abbreviation when referring to the
moths, and retain it for the foodplants. Among the four hickories offered in arena

29(3):173-194, 1990(91)

177

tests were three Carya from section §Eucarya DC. : shagbark hickory, Carya
ouata (Mill.) K. Koch; pignut hickory, Carya glabra (Mill.) Sweet; and mockernut
hickory, Carya tomentosa Nutt. The fourth was a Carya from section §Apocarya
DC.: bitternut hickory, Carya cordiformis (Wang.) K. Koch. Two species of
Juglans were offered to larvae in arena tests: black walnut, J uglans nigra L. ; and
butternut, J. cinerea L.

Arena tests were also conducted with 12 other Catocala species. Included were
7 Salicaceae feeders: amatrix Hubner, cara, concumbens Walker, meskei Grote,
parta Guenee, relicta Walker, and unijuga Walker. Food choices offered to these
species included: black willow, Salix nigra Marsh.; pussy willow, S. discolor
Muhl.; cottonwood, Populus deltoides Bartr.; quaking aspen, P. tremuloides
Michx.; bigtooth aspen, P. grandidentata Michx.; and lombardy poplar, P. nigra
var. italica Muench. Preferences were also determined for 5 oak-feeding
Catocala : arnica Hubner, coccinata Grote, ilia Cramer, lineella Grote, and
micronympha Guenee. Foods offered in the red oak subgenus (. Erythrobalanus
Spach) included: red oak, Quercus borealis Michx.; scarlet oak, Q. coccinea
Muench.; and scrub oak, Q. ilicifolia Wangenh. Foods offered in the white oak
subgenus (. Lepidobalanus Endl.) included: white oak, Q. alba L.; chestnut oak,
Q. prinus L.; and post oak, Q. stellata Wangenh.

STATISTICAL PROCEDURES

Data analysis was done on either an IBM 3083 mainframe or an IB M/PS 2
Model 70 microcomputer, using SAS Version 5.16 and SAS/PC Version 6.03,
respectively (S AS Institute Inc., 1985a, 1985b). The GLM procedure was used for
analysis of variance, with a posteriori comparisons of means done by Tukey’s
HSD (results using Tukey tests were comparable statistically to those from
Bonferroni and Duncan multiple-range tests, and each returned the same
biological conclusions). Percentage data were arcsine square root transformed
prior to analysis. Most nonparametric analyses used the NPAR1WAY and FREQ
procedures, but in instances where expected cell sizes were five or less, and the
number of cells was also small, Fisher’s exact tests and/or exact binomial
probabilities were calculated. For these, I used my own analysis of frequency
routines programmed in REXX and FORTRAN following the methods of Sokal &
Rohlf (1981).

Results

ARENA TESTS: FIRST INSTARS

1. General Patterns

Reactions of larvae to Type I and Type II controls were quite different.
That the larvae of all Catocala species readily discriminated between
potentially suitable and unsuitable foods is shown by the minuscule
amounts of Type II controls eaten; similarly, that larvae identified their
own foodplant group correctly is shown by the low amounts of Type I
controls eaten (Figures 1-3).

Table 1 shows the (Spearman rank) correlations between amounts of
food eaten and larval resting sites in arenas. While these resting sites
tended to be on the same leaves at which larval feeding occurred, it was
apparent from the outset of this research that resting site was an
insensitive index of preference. The correlations for the Juglandaceae

178

J. Res. Lepid.

Table 1. Mean correlation between resting positions on foods in arenas and
amounts of same foods eaten by first instar Catocala larvae. R = Spearman
rank correlation coefficient. N = total number of arenas. Correlations
calculated for each arena, then pooled by species.

Species

r

n

Species

r

n

habiiis

0.77

4

amatrix

0.95

4

judith

0.75

8

cara

0.72

5

obscura

0.83

9

concumbens

0.55

6

residua

0.83

8

meskei

0.54

3

parta

0.56

8

epione

0.65

11

relicta

0.65

14

palaeogama

0.70

11

unijuga

0.65

5

retecta

0.64

10

vidua

0.64

7

arnica

0.59

3

coccinata

0.86

8

subnata

0.96

11

ilia

0.42

6

lineella

0.76

4

neogama

0.70

4

micronympha

0.62

5

feeders are greater than those for either the Salicaceae or Fagaceae
feeders (pc. 01 in each case, by Mann-Whitney U-tests using data from
individual arenas). The low correlations for several oak feeders are in
part due to small sample sizes, which are in turn due to the habits of those
larvae to rest away from the food choices (on arena walls, and so forth).
This clearly reflects real biological differences in preferred natural
resting sites (e.g., off the trees, as in ilia). The lower correlations for the
Salicaceae feeders are not traceable to lower sample sizes as with the oak
feeders. Again, some of these Salicaceae feeders doubtless tend to rest off
their foodplants, but their native larval biologies remain less thoroughly
investigated.

2. JUGLANDACEAE-FEEDING CATOCALA

Figure 1 shows foodplant preferences for 18 Juglandaceae-feeding
Catocala species (data are from a total of 134 arenas, involving larval
broods from 45 different female moths).

Only neogama and piatrix preferred J uglans over all other foods, each
eating black walnut (< J . nigra ) and butternut ( J. cinerea ) nearly equally.
Unfed, newly hatched larvae of five neogama broods refused to feed on
either Cary a ovata or Cary a cordiformis, even when no other food was
offered. Although piatrix larvae ate more J uglans than Cary a, they
clearly tolerated hickories better than did larvae of neogama.

Larvae of subnata and nebulosa preferred bitternut hickory, Carya
cordiformis. Their distantly second-ranking food choice next to C.
cordiformis was J uglans nigra. Unfed first instars of three subnata and
two nebulosa broods refused to eat any Carya , and barely nibbled on
J uglans nigra when only it was offered as food; larvae of each species also

29(3):173-194, 1990(91)

179

I

<8

R

angusi

n=5, m=U, pf i-1 .05

O. (3. 0.

O

O 0

0.0.

habilis

rv=8, m=162, pf 1=1 .12

i 0. 0

O

0O

0.0

iudith

n=11, m=l79, pf 1=1 .06

d 0. 0

O

0 0

00

obscura

rv=8, iw=246, pf i = 1.95

d 0. 0.

0

00

00

residua

n=8, m=417, pfi=2.69

• 00.

0

00.

0,0

deiecta

n=4, ro=Z74, pf ■ =0.81

0.0.0

O

O0

0.0.

epione

n=11, ra=748, pf i=0.47

0. 0. 0.

0

0.0.

0.0.

f lebilis

n=5. m=66, pf i =0.76

0 0 0.

O

O 0

0.0.

insolabilis

n=3, m=65, pfi=0.83

0 0. 0.

O

O0

0.0.

laerymosa

n=6 , m=67, pf i =0. 66

0 0 0.

0

o©

©o.

palaeoaama

n=13, m=358. pfi=0.B1

0. 0 0

0

0.0.

00

reteeta

n=17, m=226, pfi=0.93

00.0

0

00.

0 0.

ulalume

n=2, m=219. pfi=0.72

0 0 0

O

O 0

0.0.

vidua

n=8, iii=273, pf i =0 . 68

0.00

o

0.0.

0.0.

Fig. 1 . Foodplant arena results with previously unfed first instar larvae of Juglandaceae-
feeding Catocala . Areas of pies proportional to amounts (percent) of food
eaten; dot in center = food not offered. Pies having the same subscripted
letters are not significantly different at the 0.05 level (Tukey HSD tests). N
= total number of arena tests. M = mean amount of food eaten per arena
in square mm. PFI = preferred foodplant index, here measuring
preference of C. ovata relative to other Carya foodplants (see text for
elaboration; the greater the PFI the greater the preference).

180

J. Res. Lepid.

Figure 1 continued

-i o

O M

h u

i i i

52

53

0 u

BO

«M m

3® H 2

Wfi ©6

P»M S3 H

►?lo|

M

H M

« m

g g

~ o. o. o

•

0.0.

0.0.

“ 0.0.0

4

0.0.

0.0.

™ ©. O. ©

O

0.0.

0.0.

~ 0. 0. 0.

O

0.0.

0.0.

accepted pecan, Carya illinoiensis (Wang.) K. Koch of section §Apocarya ,
but this foodplant was not used in arenas due to lack of material (pecan
is not native to Connecticut).

The other 14 Catocala species preferred Carya to other foodplant
genera (see below for responses to Carya cordiformis, which was usually
omitted as a food choice in their arenas). All but epione consumed more
C. ovata than any other Carya , and larvae of 5 of the 14 species — angusi ,
hahilis, Judith , ohscura, and residua — showed quite pronounced
preferences for C. ovata. This penchant for C. ovata is quantified in
Figure 1 by a “preferred foodplant index” (PFI), which is here defined as
the amount of C. ovata consumed divided by the sum of all Carya
consumed. The PFI is higher for the set of five species including angusi,
habilis Judith, obscura, and residua than for the remaining nine Carya
feeders (u=0, nl=5, n2=9, pc. 05 by Mann-Whitney IJ-test).

It is notable that only epione did not eat substantially more C. ovata
than other Carya. In contrast to the other 13 Carya feeders examined
(especially the taxa with larger numbers of arena tests), epione did not
consume Juglans as its second-ranking food item next to Carya. In fact,
epione was the only Juglandaceae feeder that ate significantly less
Juglans than it did its Type I Quercus/Populus controls (u=12, nl=8,
n2=8, pc. 05 by Mann-Whitney U-test; the amounts of control food eaten
were nevertheless meager).

Among the Carya §Eucary a Jeeding Catocala, the preference for C.
ovata is pronounced, and it is thus germane to ask: what preferences do
these larvae exhibit if C. ovata is unavailable in arenas? Several
additional arenas were run in which Carya cordiformis replaced C. ovata.

29(3):173“194, 1990(91)

181

In such arenas, test larvae of Judith , palaeogama, and residua each
increased their consumption of C. tomentosa and Juglans, while refusing
to eat Cary a cordiformis. Indeed, in other no-choice rearing situations,
unfed first instars of all North American Cary a §Eucarya -preferring
Catocala refuse to feed significantly on Carya §Apocarya when only it is
offered (my unpublished rearing and field data; see also below under
sleeve results). Larvae of neogama also refused to feed on Carya
cordiformis (see above); larvae of two piatrix broods accepted this plant,
although they did not feed as readily as when Juglans was offered.

Larvae of luctuosa (one arena test only) preferred C. ovata two to one
over all other foods, but the larvae ate much less leaf material in toto than
did larvae of all the other Catocala species tested. Moreover, luctuosa is
the only Carya £Ezmarya -preferring Juglandaceae-feeding Catocala that
showed poor (40-60 percent) survival as larvae when reared on C. ovata
indoors in the lab (the survival rate I obtain when rearing all other Carya
feeders indoors is never less than 75-80 percent). Initial field work with
luctuosa in Tennessee in 1987, and its geographic distribution in the
Nearctic indicate that this Catocala likely specializes on shellbark
hickory, Carya laciniosa (Michx.) Loud of section §Eucarya.

Test larvae of three of the Juglandaceae feeders were offspring from a
single mother: dejecta , nehulosa , and ulalume. Significant (p<.05 by
ANOVA) between-brood differences in preference existed for five of the
remaining species: epione , hahilis, ohscura , palaeogama , and retecta.
The significances for habilis and obscura were traceable to one evening
of tests in 1980, in which the arenas became waterlogged (the larvae
rested on and ate mostly Juglans). The between-brood differences were
more robust for epione , palaeogama , and retecta. These three species also
show the widest breadth in foodplant preference as both young and
mature larvae (see below). However, no consistent patterns emerged
when the known oviposition histories of the brood mothers were com-
pared to the foodplant preferences of their larvae (see Gall, 1991b; the
principal finding being that C. ovata best supports larval growth,
regardless of female oviposition history).

3. Salicaceae-feeding Catocala

Figure 2 shows foodplant preferences for the 7 Salicaceae-feeding
Catocala (data are from a total of 47 arena tests, involving offspring of 12
different females).

Only cara preferred willows over poplars. Larvae of car a did not
discriminate among the willows offered. The introduced lombardy
poplar, Populus nigra var. italica , was the first-ranked food item for
amatrix , meskei, and unijuga, but for only amatrix was this significantly
so (I regularly collect larvae of amatrix on this plant in the city of New
Haven). Larvae of concumbens , parta, and relicta all ate Populus
deltoides as their first-ranked food, with this pattern being significant for
concumbens and relicta. All Salicaceae-feeding species except for meskei

182

J. Res. Lepid.

amatriy

n=4, m=101

0. O I 0.

O0

0.0.

c ar a

n=6, m=241

0. 0 0. 0

0.0.

0.0.

concumbena

n=6, m=58?

0. 0. 0. 0.

0.0.

0.0.

meafcei

rv=3, m=267

0 ©. 0. 0.

©.©

©o.

parts

0=8, nv=442

0. 0 0. 0.

©0.

©o.

relicta

n=!4, m=296

0 0 0. 0.

©O.

0.0.

unijuqa

n=6, m=329

0. 0. 0, 0

0.0.

0.0.

Fig. 2. Foodplant arena results with previously unfed larvae of Salicaceae-feeding
Catocala. Areas of pies proportional to amounts (percent) of food eaten;
dot in center = food not offered. Pies having the same subscripted letters
are not significantly different at the 0.05 level (Tukey HSD tests). N - total
number of arena tests. M = mean amount of food eaten per arena in square
mm.

ate more of the Type I control ( Carya/Quercus ) than the Type II control,
and no species ate more Type I control than any salicaceous food.

Three of the Salicaceae feeders were represented by a single mother:
cara,concumbens, and meskei. Of the remaining four species, only relicta
showed significant (pc. 05 by ANOVA) between-brood differences in
preference. The oviposition histories of the relicta brood mothers were
not known.

4. Quercus- feeding Catocala

F igure 3 shows foodplant preferences for the 5 Quercus- feeding Catocala
(data are from a total of 36 arena tests, involving larvae from 12 different
females).

The foodplant preferences of the oak-feeding species were less sharply
defined than those of either the Juglandaceae or Salicaceae feeders.

29(3):173-194, 1990(91)

183

01

B

B

arnica

rr-i, m=65

O. 0. 0.

0. 0. 0.

O.O.

cocci nata

n=9, m=576

0 0. 0.

0. 0 0

0.0.

ilia

n=8, m=416

0. 0. 0

0 0 0.

0.0.

lineella

n=11, m=155

0 0 0.

0. 0 0

0.0.

micronympha

n=5, m=281

0. 0. 0.

0 0 0

0.0.

Fig. 3. Foodplant arena results with previously unfed larvae of Querci/s-feeding
Catocala. Areas of pies proportional to amounts (percent) of food eaten.
Pies having the same subscripted letters are not significantly different at
the 0.05 level (Tukey HSD tests). N = total number of arena tests. M =
mean amount of food eaten per arena in square mm. Quercus
( Lepidobalanus ) at left, Quercus ( Erythrobalanus ) at center.

Larvae of arnica consistently distinguished between the two oak subgen-
era, it being a Quercus (. Lepidobalanus ) feeder (but note the small
number of arenas). The cosmopolitan Nearctic ilia showed a tendency
toward Quercus (. Erythrobalanus ), eating Q. ilicifolia significantly more
than other foods. Larvae of coccinata , lineella , and micronympha did not
discriminate crisply between the two oak subgenera, although
micronympha did eat significantly more Q. ilicifolia than other foods. All
species ate more Type I control ( Carya/Populus ) than Type II, and none
ate more Type I control than any oak.

Each of the oak-feeding Catocala species was represented by offspring
of more than one mother. No species showed significant between-brood
differences in preference (p>.50 by ANOVA).

ARENA TESTS: PARTIALLY GROWN LARVAE

1. Effects of larval age

For 7 Juglandaceae and 3 Salicaceae feeders, foodplant arenas were
also run for third and fourth instar larvae. Because of the confounding
effects of leaf age and leaflet position (see Gall, 1987, and below), only
terminal leaflets were used as food items in arenas with mid instar

184

J. Res. Lepid.

Table 2. Foodplant preference in Catocala as a function of larvae age. PFI =
preferred foodplant index, here measuring preference for favorite food of
the first instar larvae compared to other foods (the greater the PFI the
greater the preference). N = total number of arenas. PFI declines between
first and middle instars, as foodplant preference broadens.

Species

Young

Larvae

PFI n

Mature

Larvae

PFI n

PFI

Change

Preferred First

Instar Food

epione

0.75

11

0.41

4

- 0.34

Cary a glabra

habilis

1.12

8

0.83

3

- 0.29

Carya ovata

judith

1.06

11

0.94

2

- 0.08

Carya ovata

obscura

1.95

8

2.13

2

+ 0.18

Carya ovata

palaeogama

0.81

13

0.37

6

- 0.44

Carya ovata

residua

2.69

8

2.49

2

- 0.20

Carya ovata

retecta

0.93

17

0.59

8

- 0.34

Carya ovata

amatrix

0.97

4

0.75

5

- 0.22

Populus italica

parta

0.60

8

0.36

3

- 0.24

Populus deltoides

relicta

0.60

14

0.54

4

- 0.06

Populus deltoides

Salicaceae-feeding Catocala. Both Juglans and Cary a cordiformis were
eventually eliminated from tests involving mid-instar Juglandaceae
feeders, since it proved impossible to assess objectively the age of leaflets
from these indeterminate foliating trees (responses of particular
Juglandaceae feeders varied, but most could be shifted back and forth at
will from eating large or small quantities of Juglans, depending upon the
leaflets selected).

With these caveats on leaf age in mind, Table 2 shows mid-instar
foodplant preferences for 7 Juglandaceae feeders ( epione , habilis Judith,
obscura ,palaeogama , residua , retecta) and 3 Salicaceae feeders ( amatrix ,
parta, relicta ). The foodplant use is expressed in Table 2 as a “preferred
foodplant index” (PFI), as described earlier, but with the numerator of
the index here being the foodplant species eaten most by the first instars.
The PFI decreased in 9 of 10 cases, indicating that foodplant preference
broadened as larvae aged (t=3, n=10, p<.05, Wilcoxon test).

2. Effects of leaf age

There is a natural dichotomy between the Salicaceae and Juglandaceae
in the timing of leaf appearance in the field. Plants in the Juglandaceae
are principally determinate foliators i.e., all leaves break from the bud at
the same time of year (true for all Cary a tested here; Juglans and Cary a
cordiformis are semi -determinate foliators). The salicaceous plants are

29(3):173-194, 1990(91)

185

indeterminate foliators, with new leaves being produced in most species
throughout at least the first half of the growing season.

To determine the influence of leaf age on mid instar larvae of the
Salicaceae feeders, arenas were constructed in which a young leaf was
used for only one food item, and old leaves for all others. Young leaves
were defined here as those immediately adjacent to the growing meris-
tem of branch termini, and old leaves as the most basal on the same
shoots. Several such tests were carried out on larvae of amatrix , cara,
and parta, with a different food item being “young” in each arena.

Young leaves were greatly preferred over old leaves (Table 3), in accord
with the findings of most other published studies on the influence of leaf
age on caterpillar feeding (e.g., Schweitzer, 1979). Leaf age nearly
completely overrides the usual foodplant preferences in each of these
three Catocala species (cf. Figure 2; note especially the shifting between
the two Salicaceae genera, Populus and Salix). Moreover, when pre-
sented with only old leaves in arenas, mid instar larvae of all three
Catocala either consumed vastly lower amounts of food, or rejected the
choice items entirely.

SLEEVE REAEINGS: SURVIVAL

1. J UGLANDACEAE-FEEDING CATOCALA

Field rearing experiments were conducted with 1 1 of the 19 Juglandaceae
feeders whose preferences were assayed in arenas: angusi , epione ,

Table 3. Arena preference tests for young versus old leaves, with Salicaceae-
feeding Catocala. Values in table are percents of food eaten. In each
arena, leaves of only one food item were young, all others old (as defined
in text). N = total number of arenas. Young leaves are preferred over old
leaves, and leaf age overrides the usual foodplant preferences of each
Catocala species (cf. Figure 2).

Species

n

Young Leaves:

Populus

deltoides

Populus

italica

Populus

tremuloides

Salix

nigra

amatrix

2

Populus deltoides

0.94

0.00

0.00

0.00

2

Populus italica

0.04

0.94

0.01

0.01

1

Populus tremuloides

0.14

0.10

0.75

0.01

1

Salix nigra

0.27

0.53

0.02

0.18

cara

2

Populus deltoides

0.92

0.01

0.00

0.07

3

Populus italica

0.00

0.89

0.00

0.11

1

Populus tremuloides

0.00

0.01

0.97

0.02

3

Salix nigra

0.00

0.03

0.16

0.81

parta

1

Populus deltoides

0.97

0.02

0.00

0.01

1

Populus italica

0.04

0.73

0.21

0.02

1

Salix nigra

0.02

0.02

0.07

0.89

186

J. Res. Lepid.

habilis, lacrymosa , neogama, obscura, palaeogama, piatrix, residua ,
retecta, and vidua. The numbers of different broods involved in these
rearings were 1,2, 1,1, 2, 2, 3, 1,3, 2, and 1, respectively. Two additional
sleeve experiments (broods habilis A79 and retecta A79) were halted
early in 1980 due to defoliation of the study sites by gypsy moth larvae.
Censuses were taken both early and late in the season for several sleeve
tests, affording insights into changes in survival and growth rates as a
function of time spent on particular foodplants.

Among the two Juglans feeders , piatrix survived equally as well in the
field on Juglans as on Cary a (Table 4). Two attempts were made at
rearing neogama brood B79 on C. ovata and Juglans nigra : experiment
B 79 1 was started with third instars that had previously been fed J uglans
nigra, while B792 was started with previously unfed first instars. Larvae
in each of these experiments refused to eat C. ovata, and died within
several days ; over half the larvae on J. nigra in each corresponding sleeve
pupated successfully (Sargent, 1982, found ca. 50 percent survival of
Massachusetts neogama on C. ovata in laboratory cage rearings). Larvae
of neogama broods X80, A85 , and A86 also refused to eat Cary a cordiformis
in the laboratory.

Among the Cary a feeders tested, angusi, habilis, obscura, residua,
retecta, and vidua survived best on C. ovata (Table 4). For each of these
species, survival was generally next best on C. glabra’, however, in none
was survival significantly higher on C. ovata than on the second ranked
food. Larvae of angusi, obscura, and residua died on C. tomentosa (all
deaths were in the first instar; note that these three species showed the
strongest arena preferences for C. ovata). Survival was moderate on
Juglans nigra for residua and vidua, and high for the other taxa.
Survival was uniformly high on all plants early in retecta brood E81, but
significant mortality occured between the two censuses on C. tomentosa
and C. glabra (pc. 05 for each by G-tests).

Of the remaining Carya feeders, epione survived equally well on all
juglandaceous foodplants; C. ovata ranked first and fourth in each of the
two broods, respectively. For palaeogama, survival was highest on C.
ovata only in brood E80. Survival on C. glabra ranked variously, with
larvae surviving poorly on it in brood A80, but well in E80. Broods A80
and E80 fared poorly on C. tomentosa, but brood B81 survived best on this
plant. Brood palaeogama A79 (not shown in Table 4; 25 larvae started
per sleeve) was the only 1980 sleeve experiment in which all larvae
pupated prior to heavy gypsy moth defoliation. Survival in the A79 brood
did not vary among C. ovata, C. glabra, and C. tomentosa (0.52, 0.72, and
0.48, respectively; p>.15 by G-test).

An intensive effort was made to force larvae of palaeogama broods A80
and E80 on Carya cordiformis and Quercus ilicifolia. These rearings
were done in the laboratory, under much more favorable conditions for
growth than in the field. All 25 larvae started from each brood refused
to feed on C. cordiformis, and died in the first instar. Of the 25 brood A80

29(3):173-194, 1990(91)

187

larvae started on Q. ilicifolia, 24 died in the first instar; a single
palaeogama A80 larva grew sluggishly on this oak through the first two
instars, only to die while molting into the third instar.

Because the life histories of subnata and nebulosa were unknown prior
to this work, all their broods were reared primarily in the laboratory on
the preferred foodplant, Cary a cordiformis , to obtain maximal series of
preserved immatures for later taxonomic work. This left few larvae with
which to attempt no-choice rearings, and I opted to place a few larvae of
subnata in the field on foodplants other than C. cordiformis. No subnata
brood B80 larvae survived past the first instar in sleeves on Cary a ovata ,
whereas 6 pupated successfully on J. nigra but gave malformed adults
(15 larvae started per sleeve). Of 10 brood A80 larvae started on J uglans
cinerea , only 5 survived through the first instar, and all died by the fourth
(implicit in the discussions here, of course, since there were no C.
cordiformis sleevings, is that this plant would have best supported larval
development of subnata in the field, as it does in the lab).

2. Other Catocala groups

The biology of the Juglandaceae feeders was emphasized for this larval
research, and hence comparable sleeve rearings for Catocala species in
other foodplant groups have yet to be conducted. Two broods were
started in 1981 with the oak-feeding ilia , but no larvae of either brood
survived beyond the third instar on any of the Quercus foodplants. I
suspect that these rearings were initiated too late in the season (8 May
1981), and that the leaf physiologies had already become unsuitable for
the larvae (egg hatch of this species in New Haven County is typically
during the last few days of April, and 1981 was an early spring). In 1983,
I confined larvae of parta brood A82 and relicta brood A82 in the
laboratory on several of their salicaceous foodplants. There were signifi-
cant differences in mortality among foodplants during the first instar
(Table 5; census at second instar). The patterns resembled the respective
first instar arena preferences (cf. Figure 2), each species surviving best
on Populus deltoides.

SLEEVE REARINGS: DEVELOPMENT RATES

There was a quite close correspondence between survival and growth
rate in the no-choice rearing studies. Among the Cary a feeders, growth
was most rapid on C. ovata in all broods except for palaeogama B81, and
significantly so in most (Table 4; growth in palaeogama B81 was fastest
on C. glabra). Growth in palaeogama A79 (not shown in Table 4) was
significantly faster on C. tomentosa and C. ovata compared to C. glabra
(median pupation dates of 4, 6, and 11 July 1980, respectively; pc. 05 for
difference of glabra against each other hickory, Mann-Whitney U-tests).

Growth was also most rapid on C. ovata in the two 1980 rearing
experiments that were halted due to gypsy moth defoliation ( habilis A79,
retecta A79). Of 25 retecta A79 larvae, all 20 survivors on C. ovata had

188

J. Res. Lepid.

pupated prior to 15 July, whereas only two of those on C. glabra and none
on C. tomentosa had done so. Five of 25 habilis A79 larvae had pupated
by this date; no larvae on C. glabra or C. tomentosa had yet done so, and
these were all at least one full instar behind the larvae on C. ovata.

The length of the larval period was shortest for epione (ca. 25-40 days),
longest for angusi, habilis , and obscura (each ca. 50-75 days), and
intermediate for the others (ca. 30-60 days). The egg of epione (and hence
its first instar larva) is considerably larger than those of the other Cary a
feeders treated here, and epione larvae also pass through only 5 instars
compared to 6-7 instars in the other species. These factors account for the
rapid development in epione (timing of egg hatch for all Cary a feeders in
southern Connecticut is the same; Schweitzer, 1982, and in preparation).
Larvae of angusi , habilis , and obscura simply grow more slowly than the
other species.

Lastly, among the Juglans- feeding species, growth was faster on
Juglans nigra than on J. cinerea for neogama X80 (Table 4). This was
also the case for piatrix , in which growth was much slower still on Cary a.

Summary

GENERAL TRENDS IN FOODPLANT SPECIFICITY

A principal finding here is that Catocala larval foodplant preferences
appear to be absolute along foodplant family lines. In choice situations
(arenas), larvae of 14 species of Juglandaceae, Salicaceae, and Fagaceae
feeders accurately discriminate against foods foreign to their own foodplant
array. In no-choice situations (sleeve and laboratory rearings), these
same Catocala larvae die if confined on foodplants from plant families
other than their own.

Limitations to larval foodplant acceptability also exist within each
foodplant family examined, but the crispness of such infra-familial
foodplant preferences is variable. Hence, preferences are least circum-
scribed for Fagaceae (two subgenera of Quercus tested), moderately so for
Salicaceae (two genera tested), and sharply delimited for Juglandaceae
(two genera, and two sections of Carya tested).

Age-specific influences on foodplant preferences were noted for both
the larvae and their foodplants. For 9 of 10 Catocala species studied,
larval foodplant preferences broadened from the first instar to the third/
fourth; and for Populus , Salix , Carya , and Juglans foodplants, young
leaves were preferred to mature leaves.

SPECIFICITY AMONG JUGLANDACEAOUS FOODPLANTS

The taxonomic divisions within Juglandaceae (. Juglans , the two sec-
tions of Carya) define boundaries to foodplant acceptability for the 19
Juglandaceae-feeding Catocala studied here. Two species prefer Juglans
C neogama , piatrix); two prefer section §Apocarya of Carya ( nebulosa ,
subnata ); and the remaining 15 prefer section §Eucarya of Carya ( angusi ,
dejecta , epione, flebilis, habilis , insolabilis, Judith, lacrymosa, luctuosa,
obscura, palaeogama, residua, retecta, ulalume, vidua).

29(3):173-194, 1990(91)

189

The capacity of larvae to cross these foodplant boundaries clearly
differs. The ability to feed largely successfully on Juglans, for example,
is shared by all 19 species examined. Among the two Juglans feeders,
larvae of piatrix feed successfully on Cary a, but those of neogama do not;
among the Cary a section §Apocarya feeders, larvae of both nebulosa and
subnata are unable to feed on Cary a section §Eucarya, and feed poorly at
best on Juglans; and no Cary a section §Eucarya feeder could feed
successfully on Cary a section §Apocarya. Hence, a foodplant shift from
either section of Cary a to Juglans appears easier than the converse. A
shift from Carya section §Eucarya to Carya section §Apocarya seems at
least as difficult as a shift from Juglans to Carya.

Of interest in this context is foodplant use (my unpublished rearing and
field data) by the remaining 7 Nearctic Juglandaceae-feeding Catocala
not treated in detail above: agrippina Strecker, atocala Brou, consors
Smith, maestosa Hulst, robinsoni Grote, sappho Strecker, and serena
Edwards. Throughout their ranges, each of consors, robinsoni, serena,
and sappho have been reliably recorded only on Carya, with robinsoni
and serena being limited to C. ouata. In the southcentral and southeast-
ern USA, maestosa feeds primarily on Carya illinoiensis and C. aquatica
(Mich.) Nutt, (both of Carya section §Apocarya); in southern New Jersey,
near this moth’s northeastern geographic limit, both Juglans and (planted)
Carya illinoiensis are used as foods. Larvae of agrippina use C. aquatica,
and rarely C. illinoiensis. Foodplant use by the recently described atocala
(a sibling of agrippina) remains unreported, but C. aquatica, C. illinoiensis,
and C. myristiciformis (Michx.) Nutt, (of Carya section §Apocarya) are
the only juglandaceous plants present where the moth flies in Arkansas
and Tennessee, and atocala will doubtless prove to use one of these.

Hence, only 6 of the 25 Nearctic Juglandacae-feeding Catocala prefer
foodplants other than Carya (7 of 26 if one includes atocala). The
remaining 19 prefer Carya , and each of these feeds heavily on shagbark
hickory, C. ovata, usually preferring it over other Carya. A subgroup of
7 of these 19 species — angusi, habilis, Judith, obscura, residua, robinsoni,
serena — is intimately associated with the biology of C. ovata. Their
larvae show the strongest arena preferences for C. ovata , and fare the
worst of all 19 Carya feeders on plants other than C. ovata in no-choice
sleeve rearings. Foodplant preference also probably broadens least in
this array of C. ovata specialist species as larvae mature (u=0, nl=4,
n2=3, p=0.05 by 1-tailed Wilcoxon test on data in Table 4).

Acknowledgements. I thank Richard Harrison, David Hawks, Charles Remington,
James Rodman, Dale Schweitzer, and Bruce Tiffney for many helpful discussions
about Catocala, and John Hartigan for statistical advice. Dale Schweitzer and
David Furth helped with foodplant and sleeving experiments, and the rearing of
innumerable larvae. Financial assistance was provided by the George D. Harris
Foundation, the E. Tappan Stannard Fund, a Sigma Xi RESA grant, and a Yale
University Prize Teaching Fellowship. This paper was drawn from a dissertation
submitted to Yale University in partial fulfillment for the degree of Ph.D.

190

J. Res. Lepid.

Table 4. No-choice field rearing results with duglandaceae-feeding Catocala larvae. Developmer
tal stage abbreviations: PU = pupa; PP = prepupa; LJ2 - late ultimate instar; U1 = earl
ultimate instar; P2 = late penultimate instar; PI = early penultimate instar; A2 = lat
antepenultimate instar; A1 = early antepenultimate instar; E2 = late ante-antepenultimat
instar; El = early ante-antepenultimate instar. Growth rates among sleeves analyze
by Kruskal-Wallis tests, survival among sleeves by G-tests (sleeves having same “Group
letter are not significantly different at the 0.05 level).

Developmental Stage at Census Growth Rate Survival

Species-brood

Foodplant

El

E2

A1

A2

PI

P2

U1

U2

PP

PU

Rank Group

N

% '

Grou

Carya glabra

1

3

7

9

7

0

0

0

0

0

3

b,c

27/50

0.54

b

angusi- A86

Carya ovata

0

0

0

0

11

10

0

0

0

0

1

a

21/25

0.84

a

Carya tomentosa

0

0

4

0

0

0

0

0

0

0

4

c

4/25

0.16

c

(late census)

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/25

0.00

d

Juglans nigra

0

0

2

2

11

1

0

0

0

0

2

a,b

16/25

0.64

a,

Carya glabra

0

0

0

0

3

1

13

8

0

0

3

a

25/30

0.83

a

epione- A81

Carya ovata

0

0

0

0

0

1

2

9

0

0

1

a

12/15

0.80

a

Carya tomentosa

0

0

0

0

0

0

0

7

0

0

2

a

13/15

0.86

a

(late census)

Carya cordiformis

Juglans nigra

0

0

0

0

11

3

0

0

0

0

4

b

14/15

0.93

a

Carya glabra

0

0

2

4

4

6

0

0

0

0

3

a,b

16/20

0.80

a

epione- A84

Carya ovata

0

0

0

1

2

7

0

0

0

0

1

a

10/10

1.00

a

Carya tomentosa

0

0

0

2

3

3

0

0

0

0

2

a,b

8/10

0.80

a

(early census)

Carya cordiformis

Juglans nigra

0

0

2

4

2

0

0

0

0

0

4

b

8/10

0.80

a

Carya glabra

0

0

0

0

0

0

0

9

0

0

4

b

15/20

0.75

a

epione- A84

Carya ovata

0

0

0

0

0

0

0

6

4

0

1

a

10/10 1.00

a

Carya tomentosa

0

0

0

0

0

0

0

7

1

0

2

a,b

8/10

0.80

a

(late census)

Carya cordiformis

Juglans nigra

0

0

0

0

0

0

2

5

0

0

3

b

7/10

0.70

a

Carya glabra

0

0

6

8

2

0

0

0

0

0

3

c

16/20

0.80

a

habilis- A84

Carya ovata

0

0

0

0

0

4

2

0

0

0

1

a

6/10

0.60

a

Carya tomentosa

0

0

0

0

0

0

0

0

0

0

0/10

0.00

b

(late census)

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/10

0.00

b

Juglans nigra

0

0

0

2

4

1

0

0

0

0

2

b

7/10

0.70

a

Carya glabra

0

0

5

3

6

0

0

0

0

0

4

a

14/15

0.93

a

iacrymosa- A87

Carya ovata

0

0

0

5

5

3

1

0

0

0

1

a

14/15

0.93

a

Carya tomentosa

0

0

0

2

7

1

0

0

0

0

2

a

10/15

0.67

a

(early census)

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/15

0.00

b

Juglans nigra

0

0

1

3

7

1

0

0

0

0

3

a

12/15

0.80

a

neogama-Bldl

K^aiyci yiavi a

Carya ovata

0

0

0

0

0

0

0

0

0

0

0/13

0.00

b

Carya cordiformis

(late census)

Juglans cinerea

Juglans nigra

0

0

0

0

0

0

0

0

0

7

1

a 7/13

0.54

a

Carya glabra

neogama- B792

Carya ovata

0

0

0

0

0

0

0

0

0

0

0/09

0.00

b

Carya cordiformis

(late census)

Juglans cinerea

Juglans nigra

0

0

0

0

1

3

3

2

0

0

1

a 9/09

1.00

a

29(3):173-194, 1990(91)

191

Developmental Stage at Census Growth Rate Survival

Species-brood

Foodplant

El

E2

A1

A2

PI

P2

U1

U2

pp

PU

Rank Group

N

% i

Group

Carya glabra

neogama-X&O

Carya ovata

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/25

0.00

b

(late census)

Juglans cinerea

0

0

0

0

2

9

2

0

0

0

2

b

13/25

0.52

a

Juglans nigra

0

0

0

0

0

2

10

7

0

0

1

a

19/25

0.76

a

Carya glabra

0

4

12

15

9

0

0

0

0

0

2

a

40/50

0.80

a

obscura- A83

Carya ovata

0

0

9

6

6

0

0

0

0

0

1

a

21/25

0.84

a

Carya tomentosa

0

0

0

0

0

0

0

0

0

0

0/25

0.00

c

(early census)

Carya cordiformis

Juglans nigra

4

3

6

0

0

0

0

0

0

0

3

b

13/25

0.52

b

Carya glabra

0

0

0

4

12

9

13

2

0

0

3

b

40/50

0.80

a

obscura- A83

Carya ovata

0

0

0

0

0

2

6

12

0

0

1

a

20/25

0.80

a

Carya tomentosa

0

0

0

0

0

0

0

0

0

0

0/25

0.00

c

(late census)

Carya cordiformis

Juglans nigra

0

0

0

0

2

7

4

0

0

0

2

b

13/25

0.52

b

Carya glabra

0

0

2

1

1

0

0

0

0

0

2

b

4/30

0.13

b

obscura-A83

Carya ovata

0

0

0

1

1

3

2

0

0

0

1

a

7/15

0.47

a

Carya tomentosa

0

0

0

0

0

0

0

0

0

0

0/15

0.00

b

(late census)

Carya cordiformis

Juglans nigra

Carya glabra

0

0

0

2

3

3

7

0

0

0

3

a

15/50

0.30

b

palaeogama- A80

Carya ovata

0

0

0

0

0

2

'5

6

0

0

1

a

13/25

0.52

a,b

Carya tomentosa

0

0

0

0

1

0

1

0

0

0

4

a

2/25

0.08

c

(late census)

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/25

0.00

c

Juglans nigra

0

0

0

0

0

4

6

4

0

0

2

a

14/25

0.56

a

Carya glabra

0

0

1

5

4

0

0

0

0

0

2

b

10/22

0.45

a

palaeogama- E80

Carya ovata

0

0

0

0

0

1

5

1

0

0

1

a

7/11

0.64

a

Carya tomentosa

0

0

2

2

0

0

0

0

0

0

3

b

4/11

0.36

a,b

(late census)

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/11

0.00

b

Juglans nigra

Carya glabra

0

1

10

18

15

2

0

0

0

0

1

a

46/50

0.92

a

palaeogama- B81

Carya ovata

0

0

1

20

2

0

0

0

0

0

2

a

23/25

0.92

a

Carya tomentosa

0

4

8

8

5

0

0

0

0

0

4

a

25/25

1.00

a

(late census)

Carya cordiformis

Juglans nigra

0

0

3

15

2

0

0

0

0

0

3

a

20/25

0.80

a

Carya giabra

0

0

0

2

7

0

0

0

0

0

4

c

9/10

0.90

a

piatrix- A82

Carya ovata

0

0

0

0

6

2

0

6

0

0

3

b,c

8/10

0.80

a

Carya cordiformis

(early census)

Juglans cinerea

0

0

0

0

3

5

0

0

0

0

2

a,b

8/10

0.80

a

Juglans nigra

0

0

0

0

0

2

4

2

0

0

1

a

8/10

0.80

a

Carya glabra

0

0

0

0

4

4

0

0

0

0

4

c

8/10

0.80

a

piatrix- A82

Carya ovata

0

0

0

0

0

2

6

0

0

0

3

b

8/10

0.80

a

Carya cordiformis

(late census)

Juglans cinerea

0

0

0

0

0

0

0

5

1

2

2

a

8/10

0.80

a

Juglans nigra

0

0

0

0

0

0

0

1

1

5

1

a

7/10

0.70

a

Carya glabra

0

0

0

3

21

2

0

0

0

0

2

b

26/40

0.65

a,b

residua- A83

Carya ovata

0

0

0

0

6

11

0

0

0

0

1

a

17/20

0.85

a

Carya tomentosa

0

2

1

5

0

0

0

0

0

0

3

c

8/20

0.40

b

(early census)

Carya cordiformis

Juglans nigra

0

3

7

4

0

0

0

0

0

0

4

c

14/20

0.70

a,b

192

J. Res. Lepid .

Developmental Stage at Census Growth Rate Survival

Species-brood

Foodplant

El

E2

A1

A2

PI

P2

U1

U2

pp

PU

Rank Group

N

% '

Group

Carya glabra

0

0

0

0

0

1

16

3

0

0

3

c

20/40

0.50

b

residua-A83

Carya ovata

0

0

0

0

0

0

0

5

11

0

1

a

16/20

0.80

a

Carya tomentosa

0

0

2

0

6

0

0

0

0

0

4

d

8/20

0.40

b

(late census)

Carya cordiformis

duglans nigra

0

0

0

0

0

0

4

10

0

0

2

b

14/20

0.70

a,b

Carya glabra

0

2

3

7

0

0

0

0

0

0

3

b

12/30

0.40

a,b

residua- B80

Carya ovata

0

0

0

0

0

4

3

3

0

0

1

a

10/15

0.67

a

Carya tomentosa

0

0

0

0

0

0

0

0

0

0

0/15

0.00

c

(late census)

Carya cordiformis

Juglans nigra

0

0

0

1

1

0

0

0

0

0

2

b

2/15

0.13

b,c

Carya glabra

1

4

7

4

10

0

. 0

0

0

0

2

b

26/40

0.65

a

residua- D80

Carya ovata

0

0

0

0

0

2

4

11

0

0

1

a

17/20

0,85

a

Carya tomentosa

0

0

0

0

0

0

0

0

0

0

0/20

0.00

b

(late census)

Carya cordiformis

dugians nigra

Carya glabra

0

1

11

14

1

0

0

0

0

0

1

a

27/40

0.67

a

retecta- E81

Carya ovata

0

3

3

10

0

0

0

0

0

0

2

a

16/20

0.80

a

Carya tomentosa

11

7

0

0

0

0

0

0

0

0

4

b

18/20

0.90

a

(early census)

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/20

0.00

b

Juglans nigra

0

0

9

5

0

0

0

0

0

0

3

a

14/20

0.70

a

Carya glabra

0

0

0

0

0

0

2

4

6

5

2

a

17/40

0.42

b

retecta- E81

Carya ovata

0

0

0

0

0

0

0

1

5

9

1

a

15/20

0.75

a

Carya tomentosa

0

0

0

0

0

1

8

0

0

0

- 3

b

9/20

0.45

a,b

(late census)

Carya cordiformis

0

0

0

0

0

0

0

0

0

0

0/20

0.00

c

» Juglans nigra

Carya glabra

0

0

0

1

11

3

0

0

0

0

3

a,b

15/20

0.75

a

retecta- C84

Carya ovata

0

0

0

0

3

4

1

0

0

0

1

a

8/10

0.80

a

Carya tomentosa

0

0

0

0

6

1

0

0

0

0

2

a,b

7/10

0.70

a

(late census)

Carya cordiformis

Juglans nigra

0

0

0

2

6

1

0

0

0

0

4

b

9/10

0.90

a

Carya glabra

0

1

10

,12

1

0

0

0

0

0

4

b

24/30

0.80

a

vidua- A82

Carya ovata

0

0

0

3

8

3

0

0

0

0

2

a

14/15

0.93

a

Carya tomentosa

0

1

1

2

6

0

0

0

0

0

3

a,b

10/15

0.67

a,b

(early census)

Carya cordiformis

Juglans nigra

0

0

0

1

3

1

0

0

0

0

1

a

5/15

0.33

b

Carya glabra

0

0

0

1

1

6

7

7

0

0

3

b

22/30

0.73

a

vidua- A82

Carya ovata

0

0

0

0

0

0

0

10

2

1

1

a

13/15

0.86

a

Carya tomentosa

0

0

0

0

0

1

7

1

0

0

4

b

9/15

0.60

a,b

(late census)

Carya cordiformis

Juglans nigra

0

0

0

0

0

0

4

1

0

0

2

b

5/15

0.33

b

29(3):173-194, 1990(91)

193

Table 5. No-choice laboratory rearing results with Salicaceae-feeding Catocaia
larvae. Census taken at second instar, number of larvae per rearing
container not equalized at start. Foodplants having same “Group” letter
are not significantly different at the 0.05 level (G-tests).

Species-brood

Survival

Populus

deltoides

Populus

grandidentata

Populus

italica

Salix

nigra

pa/ta- A82

live

52

7

1

dead

13

14

12

Group

a

b

b

relicta- A82

live

22

7

13

0

dead

8

11

4

10

Group

a

b

a

c

Literature Cited

Barnes, W., and J. McDunnough. 1918. Illustrations of the North American species
of the genus Catocaia. - Mem. Amer. Mus. Nat. Hist. 3: 1-47.

Forbes, W.T.M. 1954. Lepidoptera of New York and neighboring states. III.

Noctuidae. - Mem. Cornell Univ. Agric. Expt. Sta. 329: 1-433.

Gall, L.F. 1987. Leaflet position influences caterpillar feeding and development.
- Oikos 49: 172-176.

. 1991a. Evolutionary ecology of sympatric Catocaia moths (Lepidoptera:

Noctuidae). II. Sampling for wild larvae on their foodplants. - J. Res. Lepid.,
29: 195-216.

— . 1991b. Evolutionary ecology of sympatric Catocaia moths (Lepidoptera:

Noctuidae). III. Experiments on female opposition preference. - J. Res. Lepid.,
29: 217-233.

Gall, L.F., and D.C. Hawks. 1990. Systematics of Catocaia moths (Lepidoptera:
Noctuidae). I. Type material in the Strecker collection, with lectotype
designations. - Fieldiana (Zoology n.s.) 59: 1-16.

Hampson, G.F. 1913. Catalogue of the Lepidoptera Phalaenae in the British
Museum. Vol. 12 (Noctuidae). - British Museum, London. 626 pp.

Sargent, T.D. 1976. Legion of Night: the Underwing Moths. -Univ. Mass. Press,
Amherst. 222 pp.

. 1977. Studies on the Catocaia (Noctuidae) of southern New England. V. The

records of Sidney A. Hessel from Washington, Connecticut, 1961-1973. - J.
Lepid. Soc. 31: 1-16.

. 1982. Studies on the Catocaia (Noctuidae) of southern New England. VI. The

“pairing" of Catocaia neogama and Catocaia retecta. - J. Lepid. Soc. 36: 42-53.
SAS Institute Inc. 1985a. SAS User’s Guide: Basics, Version 5 Edition. - SAS
Institute Inc., Cary, North Carolina. 1290 pp.

. 1985b. SAS User’s Guide: Statistics, Version 5 Edition. - SAS Institute Inc.,

Cary, North Carolina. 956 pp.

194

J. Res. Lepid.

Schlenoff, D.H. 1985. The startle responses of blue jays to Catocala (Lepidoptera:
Noctuidae) prey models. - Anim. Behavior 33: 1057-1067.

Schweitzer, D.F. 1979. Effects of foliage age on body weight and survival in larvae
of the trip Lithophanini (Lepidoptera: Noctuidae). - Oikos 32: 403-408.

. 1982. Field observations of foodplant overlap among sympatric Catocala

feeding on Juglandaceae. - J. Lepid. Soc. 36: 256-263.

. 1987. Catocala p retiosa, the precious underwing moth: results of a global
status survey, with a recommendation for retention in category 2. - Status
Survey Report to the US Fish & Wildlife Service, (Newton Corner, MA). 24 pp.

Sokal, R.R., and F.J. Rohlf. 1982. Biometry (2nd Ed.). - Freeman, San Francisco.
859 pp.

Wojtusiak, J. 1979. Studies on locomotor activity during the post-embryonic
development of Lepidoptera. - Folia Biol. 27: 305-342.

Journal of Research on the Lepidoptera

29(3):195-216, 1990(91)

Evolutionary Ecology of Sympatric Catocala Moths
(Lepidoptera: Noctuidae)

II, Sampling for Wild Larvae on their Foodplants

Lawrence F. Gall

Entomology Division, Peabody Museum of Natural History, Yale University, New Haven,
CT 06511 USA

Abstract. Phenologies and foodplant use are documented for wild
larvae of 13 species of Nearctic Catocala moths whose larvae feed on
trees and shrubs in the plant families Juglandaceae, Fagaceae, and
Myricaceae. Individual Catocala species restricted feeding to one
foodplant family, with larvae of 4 species taken only on oaks ( Quercus ),

1 on sweet fern ( Comptonia ), and 8 on hickories ( Carya ) and walnuts
( Juglans ). Taxonomic divisions within Juglandaceae defined further
boundaries to foodplant acceptability, with larvae of 2 species found on
walnuts, and the other 6 on hickories. All Carya feeders preferred
shagbark hickory, Carya ovata , with 3 of 6 being limited to it. Mature
larvae were found on more kinds of foodplants than young larvae of the
same species. Larval resting sites shifted from leaves to branches and
bark as larvae matured. Different Catocala species had differing
overall resting sites, with 3 of the 6 Carya feeders specializing in hiding
under bark. Parasitism was lower in these 3 species compared to
species that rest exposed on branches. Periodic competition for food
between Catocala larvae and gypsy moth ( Lymantria dispar) larvae is
shown to be intense.

Introduction

Although previous workers have generally assumed larval foodplant
breadth to be limited among the Nearctic Catocala Schrank, little has
been published other than ex ovis foodplant acceptance of reared larvae.
Schweitzer (1982b) has presented some quantitative field data for
larvae, documenting that seven Juglandaceae-feeding Catocala use
shagbark hickory, Carya ovata (Mill.) K. Koch (of section §Eucarya DC.
of Carya Nutt.), as a foodplant in southern New England (see also the
larval collecting summaries by Rowley, 1909, and Rowley and Berry,
1910). Detailed knowledge of larval foodplant use by the Nearctic
Juglandaceae feeders is most desirable, as these taxa have long been
known to have synchronous larval and adult phenologies, and 20 or more
species occur sympatrically at most locations in eastern North America
(e.g., Sargent, 1970, 1977; Miller, 1977).

This paper examines the foodplant-linked biologies of wild larvae of 13
Nearctic Catocala species that feed on trees and shrubs in the plant
families Juglandaceae, Fagaceae, and Myricaceae. Treated herein are
distributions of larvae on different foodplant species, spatial distribu-
tions of larvae within individual foodplants, and temporal patterns in

196

J. Res. Lepid.

larval abundance. This is the second of three articles examining the
evolution of foodplant use in this speciose genus; the other articles cover
experimental work on larval foodplant specificity (Gall, 1991a) and
female oviposition biology (Gall, 1991b).

Materials and Methods
DESCRIPTION OF STUDY SITES

The primary field study site for larval sampling was West Rock Park, an open-
canopied xeric trap rock ridge in New Haven County, Connecticut, USA (see Gall,
1987, 1991a). West Rock is near the northern limit for many elements of more
southerly biota and in addition contains the usual fauna and flora of southern
New England broadleaf forests. The canopy is dominated by oak, hickory, and
ash, with a sparse under story of rosaceous, ericaceous, and other shrubs.

The juglandaceous tree fauna of West Rock is somewhat atypical for southern
New England, in that shagbark hickory ( Carya ovata ) is not the overwhelmingly
dominant member of this tree genus. Pignut hickory (C. glabra [Mill.] Sweet, of
section §Eucarya) is nearly as abundant as Carya ovata, with mockernut hickory
(C. tomentosa Nutt., of section §Eucarya) being less common. Additionally, many
of the hickories on the ridgetop are small or medium-sized (5-25 cm dbh), whereas
large (>25 cm dbh) hickories are usually encountered in most other New England
habitats. On the top of West Rock ridge there are only scattered bitternut
hickories ( Carya cordiformis [Wang.] K. Koch, of section §Apocarya DC.) and
butternuts ( Juglans cinerea L.), and virtually no black walnuts ( Juglans nigra
L.). I surveyed other Juglans trees (mostly >20 cm dbh) at the base of West Rock,
in or near parks in the city of New Haven, and on the Yale campus at the Marsh
Botanical Gardens.

There are ten Quercus species on West Rock, five each in the white oak
(. Lepidobalanus Endl.) and red oak ( Erythrobalanus Spach) subgenera, and a
non-trivial minority of probable infra-subgeneric hybrids. Seven of these oaks
can be found commonly on the ridgetop itself. Among Quercus (. Lepidobalanus )
these include: red oak, Quercus borealis Michx. ; scarlet oak, Q. coccinea Muench. ;
scrub oak, Q. ilicifolia Wangenh.; and black oak, Q. velutina Lam. Among
Quercus (. Erythrobalanus ) are: white oak, Q. alba L.; chestnut oak, Q. prinus L.;
and post oak, Q. stellata Wangenh. An eighth red oak, dwarf chestnut oak (Q.
prinoides Willd.), occurs sporadically on the ridge. Swamp white oak (Q. bicolor
Willd.) and pin oak (Q. palustris Muench.) occur only at the base of West Rock.

TECHNIQUES FOR SAMPLING LARVAE

The procedure adopted here is known to lepidopterists as “beating” or “whip-
ping,” and variations on the basic theme have long been employed by arthropod
workers. First, a close visual inspection of a tree is made for larvae. With trees
under 5-10 cm dbh, it is often possible to gently bend the trunk and upper
branches down, and thereby inspect the entire canopy. After larvae are removed
from reachable branches, a collecting sheet is spread under the tree, and the
remaining larvae dislodged from their resting positions by hitting the trunk. F or
this I used sharp blows with a softball bat, with 1-2 full size bedsheets to catch
falling larvae.

While beating might at first seem somewhat inelegant, it is nevertheless a
robust and easily implementable quantitative technique. There are a number of

29(3):195-216, 1990(91)

197

sampling biases to be considered, the four principal ones relevant to juglandaceous
foodplants being described here, with others treated in the Results as they
pertain to particular questions.

First, by beating one cannot effectively sample the larval faunas of trees much
greater than 10-15 cm dbh. This does not greatly affect conclusions in the present
study, since the majority of juglandaceous trees on or near West Rock are small
and beatable (being young, or occasionally slightly stunted). I did sample larvae
from all the large trees which I could both climb and easily reach and beat the
principal branches.

Second, the effective sampling radius for any given tree is the width of the
collecting sheet. To the extent that this bias is operant for all trees sampled, it
would not be expected to alter qualitatively comparisons among foodplant
species, assuming tree size class distributions to be largely concordant (cf.
Section 1. under Results for Juglandaceae feeders). One expects to sample a
somewhat lower percentage of the total fauna of larger trees, since the canopy
diameter increases but the collecting surface remains constant.

Third, beating the trunk might not jar a tree enough to dislodge all larvae on
branches and leaves. I investigated this possibility often, by first noting the
resting locations of Catocala and other lepidopteran larvae on hickories (but not
removing them), and then beating the trees. The majority (ca. 50-100 percent)
of larvae so located usually landed on the collecting sheet, and reinvestigation of
the tree showed that the others were no longer at their prior resting sites. These
doubtless missed the sheet. Even when a tree was hit lightly, this usually proved
sufficient to induce Catocala larvae to shift their resting positions and/or move
rapidly along a petiole or branch, and moving larvae are quite readily dislodged.
When sampling, I would therefore hit a tree several times in rapid succession,
wait some 5-15 seconds, and then beat again to dislodge remaining larvae.

Fourth, related to point three, most Cary a ovata (even some young trees) have
long strips of exfoliating bark (“shags”) under which larvae can hide and wedge
themselves, and it is doubtful that beating dislodges a large proportion of larvae
hiding there. Sampling all the shags on a particular tree is impractical, since
shags are numerous and often occur to nearly the full height of the tree.
Moreover, to search shags thoroughly requires that they be removed or bent
substantially, a process injurious to the cambium. Thus, I made a practice of
searching under 5-10 shags on suitable trees, after checking the leaves and
branches, but before beating.

LARVAL PARAMETERS

For each Catocala larva collected in the field, I recorded the sampling date, the
species, and foodplant on which it was found. For larvae located by sight, I
recorded the resting position using the following categories: terminal leaflet,
lateral leaflet, basal leaflet, branch/trunk, or under shags. For most larvae I was
also able to record larval instar, which was indexed using a combination of head
capsule width, body size, and the instar- specific morphological patterns of larvae
of different Catocala species. Many wild-collected larvae were subsequently
brought back to the lab, where they were reared to maturity to obtain parasitoids,
and/or to verify species identity in a few problematic cases.

Wilson (1975) has documented the effects of parasitoids on larval growth in
Catocala antinympha Hubner, a Myricaceae feeder. Parasitized larvae of all
instars had smaller head capsules and smaller body sizes than their unparasitized

198

J. Res. Lepid.

counterparts. This can potentially confound determination of instar using these
two parameters, and is why my prior rearing experience with Catocala —
especially knowledge of the instar-specific morphologies of the larvae — was
essential to proper analysis. For statistical tests I was conservative, using only
two larval age classes to minimize the parasitoid problem: these were mature
(ultimate, penultimate, or antepenultimate instar) and young (all earlier in-
stars). I note here that excluding larvae that proved to be parasitized from the
total larval sample alters none of the statistics presented in the Results.

TREE PARAMETERS

For each tree, I recorded its species identity, and for most I also determined size

and location. I estimated size by dbh, with size classes as follows: small, less than
10 cm; medium, 10-20 cm; or large, >20 cm. The top of West Rock ridge is a narrow
plateau, with steep flanking southwestern and northeastern slopes. Two hiking
trails run the length of much of the ridge in the vicinity of my study sites, one each
on either side of the central plateau. Thus, I defined tree location as either west
slope, east slope, or middle slope, with the trails as boundary lines.

All statistical tests of larval patterns used the appropriate observed tree
distributions, as necessary, when calculating expected larval frequencies. The
same underlying tree distributions were used for both young and mature larval
age classes when this factor was tested, since I adjusted sampling effort weekly
during each season so that proportions of the three Carya sampled daily
remained similar (p>.25 for each year by G- tests, data pooled into four-day
intervals). Other statistical testing followed Gall (1991a).

Table 1. Distributions of juglandaeeous foodplants sampled in Connecticut for
Catocala larvae, as a function of year, location, and tree size.

Foodplant

1980

1981

1982

1983

1984

1985

1986 1987

Total

Carya glabra

17

25

20

25

20

19

38

9

173

Carya ovata

30

63

17

45

40

31

56

11

293

Carya tomentosa

17

16

8

12

7

7

6

2

75

Carya cordiformis

0

0

0

3

0

1

2

2

8

Juglans cinerea

3

3

2

5

3

4

2

3

25

Juglans nigra

15

10

7

11

6

9

8

9

75

Total

82

117

54

101

76

71

112

36

649

Site

Tree size

Foodplant

East

Middle

West

Large

Medium

Small

Carya glabra

24

58

52

20

74

66

Carya ovata

29

95

110

28

87

154

Carya tomentosa

2

20

19

6

19

39

20(3):195-216s 1990(91)

199

Results

JUGLANDAGEAE-FEEDING CATOCALA
1. Tree distributions

From 1980-1987 I sampled a total of 649 hickories and walnuts for
larvae. Of these, 100 were Juglans , 541 were Carya of section §Eucarya,
and 8 were Carya cordiformis of section §Apocarya (Table 1). Many of the
same individual trees were sampled in successive years. No individual
Carya was sampled twice in any given year, though individual Juglans
often were after 1982. 1 treat these repeat Juglans trees as independent
observations, -since larvae were removed each occasion (see Section 2.
below for statistical discussion).

I sampled more Carya from the western slope than either the middle
or, especially, the eastern slopes. There were no differences in the
frequencies of the three Carya sampled as a function of slope and date
(p>.50 by G-tests, data pooled across all years). For C. ovata and C.
glabra there were highly significant differences in sampling location as
a function of year, whereas for C. tomentosa there was not (pc.Ol, pc. 01
and p>.50, respectively, by G-tests). This reflects a real biological
difference in the microhabitat distribution of C. tomentosa (much more
common on the west slope than the middle or east slopes) compared to C.
ovata and C. glabra. The significances for C. ovata and C. glabra can be
traced to the fact that I did not collect on the east slope in 1983 and 1987.

I adjusted sampling effort to maintain comparable tree size class
distributions among Carya , and hence tree size as a function of tree
species proved to be the same both within and among years (p>. 10 in each
case by G-tests). Despite these adjustments, the overall frequencies of
the three Carya foodplants proved marginally heterogeneous as a func-
tion of year (G=23.88, df=14, pc. 05). This was due to a preponderance of
Carya tomentosa sampled during the first year of the study (G= 15.64,
df=12, p>.15, omitting 1980). Tree size also proved marginally heteroge-
neous among Carya (G= 13.496, df=6, pc. 05), with this being traceable to
an excess of medium-sized Carya glabra (G 1.711, df=4, p>.50, without
it). For each Carya , tree size also proved constant as a function of
sampling site (p>.25 by G-tests for each species).

These heterogeneities in tree sampling have minimal influence on later
conclusions about larval foodplant use, with the possible exception of
epione. In specific: first, regarding the site by year heterogeneity for C.
ovata and C. glabra , few Catocala larvae were collected anywhere during
1983 and 1987 (9 percent of the total Carya larva sample; 16 percent of
epione , and 3 percent of all other taxa combined). Second, regarding the
1980 oversampling of Carya tomentosa , only epione used this foodplant
to any significant extent, and sampling in 1980 was late in the season,
after most epione larvae had pupated.

200

J. Res. Lepid.

Table 2. Distributions of wild Connecticut Juglandaceae-feeding Catocala larvae:
above, by Catocala species by year; below, by foodplant species by year.

Species

1980

1981

1982

1983

1984

1985

1986

1987

Total

epione

8

115

3

17

1

5

19

11

179

hahilis

4

25

3

1

0

4

0

0

37

judith

1

3

1

0

0

1

0

0

6

neogama

23

8

11

0

0

0

0

7

49

palaeogama

20

42

7

3

3

11

7

3

96

piatrix

0

0

2

1

0

0

1

0

4

residua

11

6

1

1

0

0

0

0

19

retecta

28

35

17

5

6

12

23

1

127

Total

95

234

45

28

10

33

50

22

517

Foodplant

1980

1981

1982

1983

1984

1985

1986

1987

Total

Carya ovata

49

180

23

23

6

28

40

12

361

Carya glabra

17

20

6

3

0

3

7

3

59

Carya tomentosa

5

26

3

1

4

2

2

0

43

Juglans cinerea

3

4

4

0

0

0

0

1

12

Juglans nigra

21

4

9

1

0

0

1

6

42

Total

95

234

45

28

10

33

50

22

517

Table 3. Distributions of wild Connecticut Juglandaceae-feeding Catocala larvae, as
a function of foodplant species. P = significance level for preference of
favored* foodplant genus versus all other juglandaceous foods (G-tests or
exact multinomial probabilities).

Species

Carya

ovata

Carya Carya Juglans Juglans
glabra tomentosa cinerea nigra

Total

Favored
Foodplant p

epione

128

21

30

0

0

179

Carya <.01

habilis

37

0

0

0

0

37

Carya <.01

judith

6

0

0

0

0

6

Carya .67

palaeogama

68

22

6

0

0

96

Carya <.01

residua

17

2

0

0

0

19

Carya .06

retecta

105

14

7

0

1

127

Carya <.01

neogama

0

0

0

11

38

49

Juglans <.01

piatrix

0

0

0

1

3

4

Juglans <.01

Total

361

59

43

12

42

517

29(3):195-216, 1990(91)

201

2. Larval preferences among foodplant genera

I collected 517 Catocala larvae from the 649 hickories and walnuts
(Tables 2-3). A total of 464 were larvae of epione Drury, habilis Grote,
judith Strecker, palaeogama Guenee, residua Strecker, and retecta
Grote; the other 53 were neogama Smith and piatrix Grote. I did not find
larvae of either obscura Strecker or subnata Grote. All neogama and
piatrix larvae were found on Juglans , and 463 of 464 larvae of the
remaining six species were on Carya (Table 3). The single Carya- to-
Juglans crossover was an antepenultimate retecta larva, found on 31
May 1980 on Juglans nigra in a New Haven park.

Repeat sampling of Juglans trees from 1982 through 1987 biases the
foodplant-genus level statistical tests in Table 3. The preferences of
neogama and piatrix for Juglans are substantially underestimated by
this repeat sampling bias, and, accordingly, the preferences of the other
Catocala species for Carya are slightly overestimated. For epione ,
habilis , palaeogama , and retecta , the bias does not affect the biological
conclusion — that is, restriction to Carya — drawn from the Table 3
statistical tests, since the preferences of these four Catocala for Carya
remain significant if only the 1980 and 1981 larval samples are consid-
ered (the years in which no repeat sampling of Juglans was done; pc. 01
for each species by G-tests). However, for residua , whose larval samples
were small, the test for Carya preference is p=0. 1 12 for the 1980 and 198 1
seasons. Note that Schweitzer (1982a, p. 258) searched a total of 6
Juglans and 10 Carya in Connecticut in 1979-1980, using comparable
sampling methods, and captured 5 residua on C. ovata. Inclusion of his
additional trees and larvae with the 1980-1981 data from Table 3 gives
p=0.031 for preference of Carya by residua.

3. Larval preferences within foodplant genera

Table 4 shows the foodplant distributions of the six Carya -feeding
Catocala species as a function of larval age. Different species had
different preferences among the Carya foodplants (G=46.07, df=10,
pc. 01), although all preferred C. ovata, and foodplant use broadened in
each species as the larvae matured (G=7.31, df=2, pc. 05). The repeat
sampling of Juglans effectively precludes a precise analysis of preference
among Juglans for neogama and piatrix, but neither of these Catocala
seemed to discriminate sharply between J. nigra and J. cinerea (p>.50 for
each by G-tests, treating the repeats as independent). None of the Carya
or Juglans feeders showed differences in foodplant preference within
their foodplant genus as a function of year (p>.25 in each case by G-tests,
controlling for instar, foodplant, and year).

Young larvae of epione , habilis, palaeogama, and retecta were all found
disproportionately on C. ovata (pc. 01 for each, G-tests). Mature larvae
of habilis, residua, and retecta also strongly preferred C. ovata (pc. 01 for
each, G-tests). Mature larvae of epione remained primarily on C. ovata,
but not as faithfully (G=8.33, df=2, pc. 05), and mature larvae of

202

J. Res. Lepid.

Table 4. Distributions of wild Connecticut Juglandaceae-feeding Catocala larvae, as
a function of larval age and foodplant species. Mature larvae are found on
more kinds of foodplants than are young larvae of the same Catocala
species. See text for elaboration and statistical analyses.

Species

Carya

Larval Age

Carya

ovata

Carya

glabra tomentosa

epione

Young

65

3

16

Mature

63

18

14

habilis

Young

9

0

0

Mature

28

0

0

judith

Young

1

0

0

Mature

5

0

0

palaeogama

Young

23

5

0

Mature

45

17

6

residua

Young

2

0

0

Mature

15

2

0

retecta

Young

25

3

1

Mature

80

11

6

palaeogama were found nearly equally on the three Carya species
(G=4.20, df=2, p>.15). Young Judith and residua larvae were not
analyzed due to small sample sizes, but it seems clear from other field
data (Gall, 1991b) and their mature larval profiles that these two
Catocala , like habilis, are restricted to C. ovata throughout their larval
cycle.

Larvae of epione were found overwhelmingly on small C. ovata (G= 14.59,
df=l, pc. 01) and small C. tomentosa trees (G=14.01, df=l, pc. 01; test is
large trees against all others). Larvae of the other five Carya-feeding
Catocala were more often found on large C. ovata trees (pc. 05 for each
species, similarly), the trend being most pronounced for habilis Judith,
and residua.

4. Larval phenologies

Figure 1 shows larval instar plotted as a function of sampling date, for
each of the Catocala species collected. For epione , habilis, palaeogama,
residua, and retecta, larval instar increased smoothly as the season
progressed (pc. 05 by ANOVA for each species, for difference in instar as
function of date; instar means shown in Table 5). Instars of neogama did
not increase with sampling date (p>.25 by ANOVA), reflecting its greatly
staggered egg hatch in comparison to the synchronous egg hatch of the
aforementioned five Carya feeders. The Juglans- feeding piatrix has

29(3):195-216, 1990(91)

203

20

iudith

10

Figure 1 . Phenologies of wild Connecticut Juglandaceae-feeding Catocala larvae.
Black squares = ultimate instars; stippled squares = penultimate instars;
hatched squares = antepenultimate instars; white squares = earlier instars.
Collection data pooled for 1980 through 1987.

204

J. Res. Lepid.

Table 5. Mean collection dates for wild Connecticut Juglandaceae-feeding Catocaia
larvae, by larval instar. Data from 1 980 through 1 987 pooled, dates begin at 1 May
(32 = 1 June, 62 = 1 July), instar means increase smoothly for Garya-feeding
species, reflecting synchronous egg hatch; no comparable smooth increase for
Juglans- feeding species due to asynchronous egg hatch.

Species

Larval Instar

Younger Antepenultimate Penultimate

Ultimate

Total

epione

10.6

18.3

24.9

26.8

176

habilis

28.6

33.9

47.4

56.0

37

Judith

30.0

29.0

34.0

41.0

6

palaeogama

17.6

27.8

31.3

35.9

89

residua

29.0

29.4

29.0

33.7

16

retecta

21.8

30.8

37.1

39.4

126

neogama

45.3

47.0

44.8

38.3

34

piatrix

42.0

43.3

4

greatly staggered egg hatch, and the Cary a -feeding Judith has synchro-
nous egg hatch, but the phenologies of these two Catocaia were not tested
due to small sample sizes.

The 1982 and 1984 larval seasons proved to be significantly later than
all other years (pc. 05 by ANOVA, controlling for instar and Catocaia
species). The 1981 season was earliest, followed quite closely by 1985,
1987, and 1980, with 1983 and 1986 being intermediate; but neither of
these trends was statistically significant.

Among the Carya- -feeding species, larvae of epione were found consid-
erably earlier than larvae of the other tax: a, and larvae of habilis later
(pc. 01 by ANOVA; Mann- Whitney tests used for comparisons involving
Judith). Larvae of palaeogama were slightly earlier than those of retecta
(pc. 05, similarly), and larvae of the remaining species closely overlapped
each other in time (p>.25 similarly for each comparison). The timing of
egg hatch is the same for these six Carya' -feeding Catocaia ; hence, the
earlier phenology of epione can be linked to its larger egg (and more
robust first instar larva), and 5 larval instars compared to 6-7 in the other
taxa. Larvae of habilis simply develop less rapidly than all the others (cf.
Gall, 1991a).

For epione, palaeogama, and retecta , whose larvae were found occasion-
ally on foodplants other than C. ovaia , there were no significant differ-
ences in larval collection date as a function of foodplant species (p>.15 by
ANOVA for each). For epione , a tree by instar interaction (pc. 05 by
AN OVA) was due to progressively lower collection date means for mature
larvae on C. ovata compared to C. tomentosa as the season advanced.
This probably reflects an accelerated growth rate on C. ovata in older
epione larvae, a pattern also seen in this species’ no-choice rearing
experiments (Table 4 in Gall, 1991a).

29(3):195-216, 1990(91)

205

5. Larval densities on foodplants

The mean numbers of larvae collected from each tree species are
presented in Table 6. Adult Catocala were extremely scarce from 1982
through 1984 in southern Connecticut (trap records of myself and D.
Schweitzer), and Table 6 reflects this adult scarcity as a 10-fold decline
in larval abundance from 1980 through 1984, especially when epione is
removed (there was a bias against collecting epione in 1980, a year in
which sampling was disproportionately late in the season). The order of
magnitude decline in abundance is highly significant for each foodplant
species in Table 6 (pc. 01 by G-tests, trees split each year into two classes:
having no larvae, having one or more larva).

Table 7 shows plots of the number of larvae collected per individual
foodplants, by year and Catocala species. The variance to mean ratio, an
index of contagion, is greater than one for each of these 14 plots,
indicating that larvae were not distributed randomly among individual
trees. Stated another way, fewer trees than expected had only one larva,
and more than expected had no larvae. This makes good biological sense,
since the Carya-feeding Catocala lay varying sized clumps of eggs when
ovipositing. The plots also become less contagious as larvae mature (t=0,
n=6, pc. 05, Wilcoxon test; using those taxa having calculable young and

Table 6. Mean numbers of wild Connecticut Juglandaceae-feeding Catocala larvae
sampled per tree, as a function of foodplant species. Above, all Catocala
species included; below, excluding epione. Larval densities are always
highest on C. ovata. Note 10-fold decrease in densities on all foodplants
from 1980 through 1984.

Larvae Per Foodplant Tree

All Catocala :

1980

1981

1982

1983

1984

1985

1986

1987

Juglans

1.33

0.62

1.44

0.06

0.00

0.00

0.10

0.58

Carya

1.11

2.17

0.71

0.33

0.15

0.58

0.49

0.68

C. glabra

1.00

0.80

0.30

0.12

0.00

0.16

0.18

0.33

C. ovata

1.63

2.86

1.35

0.51

0.15

0.90

0.71

1.09

C. tomentosa

0.29

1.63

0.38

0.08

0.57

0.29

0.33

0.00

Without epione:

1980

1981

1982

1983

1984

1985

1986

1987

Juglans

1.33

0.62

1.44

0.06

0.00

0.00

0.10

0.58

Carya

0.98

1.07

0.64

0.12

0.13

0.49

0.30

0.18

C. glabra

0.94

0.32

0.25

0.08

0.00

0.16

0.05

0.22

C. ovata

1.50

1.57

1.24

0.18

0.13

0.81

0.50

0.18

C. tomentosa

0.12

0.25

0.38

0.00

0.57

0.00

0.00

0.00

206

J. Res. LepicL.

Table 7. Numbers of wild Connecticut Juglandaceae-feeding Catocala larvae sampled per
individual foodplant tree, by year, Catocala species, foodplant species, and larval
age. Distributions having a plus exclude data from large trees and Catocala epione.
Variance to mean ratios for all distributions are greater than 1 .0, indicating clumping
of larvae on foodplants; asterisks show significant clumping i.e., departure from
Poisson (0.05 level, G-tests or exact multinomial probabilities).

Larval Number of Larvae Totals Larvae/Tree

Species Age Year Foodplant 0 1 2 3 4 >4 Trees Larvae Mean Var/

Mean

epione

young

1981

Cary a ovata

13

2

1

0

1

3

20

40

2.00

9.05*

epione

mature

1981

Carya ovata

17

8

2

1

0

3

31

30

0.97

2.44*

epione

young

1981

Carya tomentosa

5

1

0

1

0

2

9

14

1.56

3.07*

epione

mature

1981

Carya tomentosa

2

0

0

2

0

0

4

6

1.50

2.00

epione

young

1983

Carya ovata

11

0

0

1

2

0

14

11

0.79

4.02*

epione

young

1986

Carya ovata

16

0

0

0

0

1

17

5

0.29

5.01 *

epione

mature

1986

Carya ovata

34

4

0

1

0

0

39

7

0.18

1.72

habilis

young

1981

Carya ovata

34

1

2

0

0

0

37

5

0.14

1.71

habilis

mature

1981

Carya ovata

8

3

2

1

0

0

14

10

0.71

1.39

palaeogama

mature

1980

Carya ovata

21

4

2

0

0

0

27

8

0.30

1.25

palaeogama

young

1981

Carya ovata

28

4

4

1

0

0

37

15

0.41

1.57

palaeogama

mature

1981

Carya ovata

7

4

2

1

0

0

14

11

0.79

1.37

palaeogama

mature

1985

Carya ovata

12

3

2

0

0

0

17

7

0.41

1.23

residua

mature

1980

Carya ovata

22

4

1

0

0

0

27

6

0.22

1.15

retecta

mature

1980

Carya ovata

16

8

1

0

2

0

27

18

0.67

1.85

retecta

young

1981

Carya ovata

31

4

1

0

1

0

37

10

0.27

2.19

retecta

mature

1981

Carya ovata

11

2

0

1

0

0

14

5

0.36

1.99

retecta

mature

1985

Carya ovata

14

1

1

0

0

1

17

8

0.47

3.49

re tecta

mature

1986

Carya ovata

26

7

5

1

0

0

39

20

0.51

1.32

all taxa

mixed

1980

Carya glabra

2

1

2

2

0

0

7

11

1.57

1.03

all taxa

mixed

1980

Carya ovata

9

8

6

1

0

2

26

36

1.39

2.20*

all taxa

mixed

1981

Carya glabra

7

1

3

0

1

0

12

11

0.92

1.88

all taxa

mixed

1981

Carya tomentosa

, 7

2

0

1

0

2

12

22

1.83

6.93*

all taxa

mixed

1981

Carya ovata

21

6

4

7

4

9

51

147

2.77

8.34*

all taxa +

mixed

1981

Carya ovata

29

7

0

6

5

0

47

45

0.96

2.23*

all taxa

mixed

1985

Carya ovata

16

6

4

2

0

1

29

26

0.90

2.08

all taxa +

mixed

1985

Carya ovata

16

4

3

1

0

1

25

19

0.76

2.55

all taxa

mixed

1986

Carya ovata

35 13

3

3

0

2

56

40

0.72

2.47*

all taxa +

mixed

1986

Carya ovata

23 10

5

2

0

0

40

26

0.65

1.23

mature larval samples in the same year), although this analysis is
heavily controlled by epione, whose larvae clump considerably more than
those of the other taxa.

Table 7 also gives similar plots of the number of all Catocala larvae
collected per tree, by year and by foodplant. Again, the variance to mean
ratios are all greater than one, and highly significantly so for C. ovata and
C. tomentosa. The clumping is due neither to the undue influence of
epione nor to overabundance of larvae on large trees, as shown by the
remaining high clumping indices when both these factors are removed

29(3):195-216, 1990(91)

207

(V* plots in Table 5). This remaining clumping might in part be an
artifact of pooling larval age classes. However, with plots having
mixtures of Caiocala species, the tree samples cannot be readily parti-
tioned and tested by groups corresponding to young and mature larvae.

6. Larval resting sites

Table 8 shows resting locations for the sampled larvae of the Carya
feeders, and Caiocala antinympha , as a function of larval age. The larvae
of antinympha were collected on West Rock on its myricaceous food plant,
sweet fern ( Comptonia peregrina [L.] Coult.).

Different Carya feeders showed markedly divergent preferences for
resting locations (G-109.40, df=2, pc, 01). Larvae of all species shifted
their resting sites from leaves to woody material as they matured
(G=63.30, df=4, pc. 01), the magnitude of the change varying from species
to species, with retecta and epione larvae changing their habits most
strongly. Comparable age-specific shifts in resting behavior were re-
corded for Caiocala antinympha on Comptonia (G=20.96, df=l, pc. 01).

Table 8. Resting sites for wild Connecticut Juglandaceae-feeding Catocala larvae,
and larvae of Catocala antinympha on Comptonia peregrina. Terminal
leaflets preferred over other leaflets by young larvae; woody sites preferred
over leaves by older larvae. Different Catocala species have different
overall preferences for type of resting site. See text for statistical analysis.

Species

Larval Age

Leaflet Resting Site

Basal Lateral Terminal

Woody Resting Site
Branches Shags

epione

young

5

10

23

4

0

mature

1

3

3

30

1

habilis

young

0

0

2

4

1

mature

0

0

1

0

11

Judith

young

0

0

0

0

1

mature

0

0

0

0

2

paiaeogama

young

0

0

11

1

0

mature

0

5

14

0

0

residua

young

0

0

0

0

0

mature

0

0

0

1

5

retecta

young

2

2

4

0

0

mature

1

2

2

11

4

antinympha

young

10

1

mature

3

13

208

J. Res. Lepid.

Table 9. Parasitism rates among wild Connecticut Juglandaceae-feeding Catocala.

Collection data pooled across all years. Parasitism is higher for species that
rest exposed on branches compared to species that rest under bark.

Species

Preferred Resting Site

Parasitized

Healthy

Judith

hidden under shags

0

3

habilis

hidden under shags

0

12

residua

hidden under shags

0

12

epione

exposed on bark/leaf

15

40

palaeogama

exposed on bark/leaf

7

37

retecta

exposed on bark/leaf

11

73

Table 1 0. Distributions of wild Connecticut larvae of Catocala arnica and Catocala
lineella, as a function of their Quercus foodplant species. Larvae of arnica

prefer Lepidobalanus, larvae of lineella prefer Erythrobalanus (see text for
statistical analysis).

Foodplant

Trees

Sampled

Catocala

arnica

Catocala

lineella

Quercus ( Erythrobalanus ) borealis

27

2

8

Quercus ( Erythrobalanus ) cocci nea

8

0

0

Quercus ( Erythrobalanus ) ilicifolia

22

2

3

Quercus ( Erythrobalanus ) velutina

13

0

0

Quercus ( Lepidobalanus ) alba

20

4

0

Quercus ( Lepidobalanus ) prinoides

1

0

0

Quercus ( Lepidobalanus ) prinus

20

2

2

Quercus ( Lepidobalanus ) stellata

31

3

2

Among Carya-feeding larvae collected on woody material, there were
differences in utilization of shags versus branches/trunks (G=46.43,
df-5, pc. 01), with habilis , Judith , and residua specializing in hiding
under shags. The low numbers of larvae of these three species collected
by beating in comparison to epione, palaeogama , and retecta mostly
reflects this difference in resting behavior, as well as my tendency not to
search exhaustively under shags on C. ouata trees.

Larvae resting on leaves showed a strong preference for terminal
leaflets (G=41.95, df=2, pc. 01). Table 8 underestimates this preference
for terminal leaflets, since the availability is 1:2:2 for terminals, laterals,
and basals, respectively, on C. ovata and C. glabra (or greater, as for
example in C. tomentosa , which regularly has 7-11 leaflets). The
differing leaflets of Carya present qualitatively different food resources
to Catocala larvae, and the influence of leaflet position on larval survival
and development, and its more general implications for plant-phytophage
interactions, have been treated elsewhere (Gall, 1987).

29(3): 195-216, 1990(91)

209

7. Larval parasitoids

Table 9 gives numbers of parasitized and healthy larvae, for those wild
larvae of the Carya feeders that were subsequently brought back and
reared in the lab. These data are likely to underestimate the actual
parasitism rates in the field, since many young larvae were collected
before heavy parasitism might be expected (but note, for example, the
several tachinid fly species whose minute eggs are laid on leaf surfaces
and ingested by phytophagous larvae). The Catocala whose larvae rest
primarily under shags (habilis Judith, residua) had significantly lower
parasitism rates (G=9.55, df l, pc. 01) than those whose larvae rest
exposed on branches and trunks (epione, palaeogama, retecta).

RESULTS: QUERCUS-FEEDWG CATOCALA

A sample of 142 oaks was also beaten during 1980-1987 (Table 10). All
of these oak trees were classified as small or medium in size. A total of
31 Catocala larvae were taken from oaks.

All but 3 of the sampled larvae proved to be arnica Hubner or its sibling
lineella Grote. Table 10 gives the distributions of arnica and lineella
among their Quercus foodplants. Considering only those five oak species
on which larvae were taken, these two Catocala species had the same
foodplantuse (but only marginally; G=8.61, df=4, .10>p>.05). Pooling by
oak subgenus, these two species differed (G=4.96, df=l, pc. 05), arnica
being principally on white oaks, Quercus ( Lepidobalanus ), and lineella
on red oaks, Quercus ( Erythrobalanus ).

The other 3 larvae beaten from oaks included 2 micronympha Guenee
and 1 coccinata Grote, all taken on Quercus stellata. Curiously, no ilia
Cramerlarvae were captured, as the adults are ubiquitous each year in
the study sites (adult micronympha and coccinata are uncommon there).
The absence of ilia larvae could in part be due to my sampling rather late
in its larval season — the mature larvae tend to rest off the foliage and
trunks during the day, and hence would be less collectable during
daytime beating. Larvae of ilia might also prefer larger trees. Elsewhere
in North America, larvae of each of ilia, micronympha, and coccinata
have been recorded on trees from both Quercus subgenera (my unpub-
lished field records).

Discussion

GENERAL TRENDS IN FOODPLANT SPECIFICITY

The larval records demonstrate that foodplant specificity in these 13
Nearctic Catocala species is absolute at the foodplant family level — no
larvae of J uglandaceae-feeding taxa were taken on Quercus , and no
larvae of Quercus -feeding taxa were taken on juglandaceaous plants.
Field data on larval foodplant use for Catocala that eat other foodplant
genera have also been accumulating recently (Schweitzer, 1987; unpub-
lished field data of myself and others). The conclusion emerging from
these studies is that the taxonomic boundaries defined by all nine known

210

J. Res. Lepid.

Catocala foodplant families are respected in the field by the larvae (these
other collection data will be summarized in the forthcoming taxonomic
monograph of the North American Catocala).

Within single plant families, foodplant specificity is similarly sharp. At
the generic level within Juglandacae, 2 Catocala species in southern
Connecticut appear restricted to Juglans ( neogama , piatrix ), and 6 to
Carya ( epione , habilis, Judith, palaeogama, residua , retecta ). Adult
subnata occur near the larval field sites, and I consider the absence of
subnata larvae on Juglans and Carya section §Eucarya in the present
work, and its overriding preference for Carya section §Apocarya in
studies of larval foodplant use (Gall, 1991a) and oviposition (Gall, 1991b)
to indicate restriction to Carya section §Apocarya.

Among the few oak-feeding Catocala collected, arnica appears to prefer
white oaks, Quercus ( Lepidobalanus ), while lineella prefers read oaks,
Quercus ( Erythrobalanus ). However, the foodplant specificity is not
nearly so sharp within these two subgenera of Quercus as it is between
the Juglandacae genera.

THE INFLUENCE OF LARVAL AGE

Larval age as indexed by instar proved to be a pivotal variable, and
analysis of other factors in the absence of age would have led to erroneous
conclusions about the foodplant-linked biologies of all the Juglandaceae-
feeding Catocala species. This is particularly so for the age specific shifts
in resting behavior observed for these Catocala larvae (Table 8).

Larval foodplant use broadens as larval age increases (Table 4):
mature larvae of epione , palaeogama , and retecta all were more common
on C. glabra than were young larvae of the same species, and for

Table 11. Species identities of wild Connecticut Juglandaceae-feeding Catocala
larvae taken on C. ovata, for trees yielding seven or more larvae. Asterisks
indicate tree samples from Schweitzer ( 1 982b) , pluses indicate tree samples
from Godwin (unpublished data from tree defaunation studies).

Number

Year

Tree Size

Larvae

epione

habilis

judith palaeogama residua retecta

1981

small

26

22

2

0

2

0

0

1981

large

20*

0

8

1

1

0

10

1981

small

19

15

0

0

3

0

1

1981

large

17*

0

0

0

5

1

11

1980

medium

15 +

5

0

0

9

0

1

1981

small

11

8

0

0

3

0

0

1982

large

10

2

0

0

4

1

3

1981

medium

9

6

0

0

2

0

1

1981

small

8

8

0

0

0

0

0

1981

large

8

0

3

2

0

3

0

1980

large

7

0

1

0

1

1

4

1979

large

7 +

0

6

0

0

1

0

29(3):195-216, 1990(91)

211

palaeogama and retecta , C. tomentosa was used essentially only by
mature larvae. This broadening in foodplant use with age is primarily
a function of larval movement (differential mortality may contribute to
a small extent: in principle, inferior field survival of larvae on C. ovata
compared to C. glabra and C. tomentosa could broaden foodplant distri-
butions for mature larvae, but this is known not to be the case for the
Cary a -preferring Catocala treated here; Gall, 1991a). In this regard,
note that my efficiency at sampling trees was poorest when larvae were
mature, for two reasons. First, young larvae rest essentially only on the
undersides of leaves, on the midribs, where they are quite visible; mature
larvae are cryptically patterned, and shift their restings sites to branches
and under shags (but note palaeogama which stays on leaves), and so are
more difficult to locate. Second, closer visual scrutiny of foodplants is
critical when searching for younger larvae, since such small larvae are
more likely to be blown by breezes far from the collecting sheet when trees
are beaten. Thus, the lack of young (and abundance of mature) larvae on
C. glabra and C. tomentosa is not likely to be explainable by collecting
bias.

INTERSPECIFIC LARVAL ASSOCIATIONS

Tables 6 and 7 demonstrate that Catocala larvae often co-occur on
individual trees (and larval densities are certainly higher in general than
indicated here). The corollary question is whether the co-occurences
represent larvae of the same or different Catocala species. Table 11 lists
the identity of Catocala larvae taken on individual C. ovata trees from
which 7 or more larvae were collected. Several trees from Schweitzer
(1982b) and unpublished pyrethrin defaunation studies conducted by B.
Godwin are included here (larvae determined by me). All but one tree in
Table 11 contains a mixture of at least two species, and most trees have
three or four. Note also the field reports by Rowley ( 1909 ) and Rowley and
Berry (1910) detailing the frequent co-occurence in Missouri of numbers
of angusi Grote, epione , habilis Judith, palaeogama, residua, and vidua
Smith larvae on C. ovata.

The percentage of all trees from which I collected more than one larva,
on which two or more Catocala species were also present, was high for
both C. ovata (34 of 57) and C. glabra (7 of 1-1). Furthermore, these
percentages remained comparable from year to year, despite the great
fluctuations observed in larval densities (for C. ovata, G=5.13, df=6,
p>.50; for C. glabra, G=3.28, df=3, p>.25).

Another way to express the association of different Catocala species as
larvae is to calculate resource overlap indices. Table 12 lists overlap
indices for larvae of the Cary a feeders, on foodplant use and resting site
(the MacArthur-Levins index as modified by Lawlor [1980] was used to
reflect consumer electivities i.e., forage ratios sensu Stanton [1982] are
used as Lawlor’s alphas). Table 12 also shows the directional change in
the overlap indices between young and mature larvae, and these graphs

Table 12. Overlap indices for wild Connecticut Juglandaceae-feedsng Catocaia larvae, as a function of larval age. Above, forfoodpiant use; below,
for resting site. “Difference” is directional change in overlap index as function of age: zero = change of less than 0.10; plus =■ increase
of more than 0.10; minus = decrease of more than 0.1 0. “Difference” table interpretable biologically as change in association of different
Catocaia species as larval season progresses.

212

J. Res. Lepid.

29(3):195-216, 1990(91)

213

are therefore interpretable biologically as the change in association of
different Catocala species on the same foodplant array over time (refer
as well to the larval phenologies given in Figure 1). These overlap indices
are offered mainly as instructive guides, since they help visualize the
principal findings of the larval collection data; no statistical testing is
done on the indices, because in most instances they are based on small
larval samples.

Note the uniformly high values for foodplant use overlap, in both young
and mature larvae. Larvae of epione became more associated with larvae
of retecta and palaeogama on foodplants as larvae matured, but the
resting sites of epione and palaeogama diverged sharply at the same
time. Larvae* of habilis, Judith, and residua strongly overlapped with
each other in both foodplant use and resting site throughout the season
(each specializes on large C. ovata trees, and hides under shags).

The larvae biology of epione differs substantially from those of the other
five Cary a' -feeding species examined here (5 vs 6-7 instars, early vs later
season phenology, small vs large tree preference, broad vs restricted
foodplant use among Cary a). It is notable that the other five Cary a
feeders comprise part of a closely knit monophyletic unit in the genus
Catocala , with epione being removed at considerable distance (on the
basis of egg, larval, and adult morphological characters). I suspect that
eating Juglandaceae has been arrived at convergently by epione.

LARVAL COMPETITION FOR RESOURCES

For Catocala , there are at least two distinct resources for which the
larvae might compete: food, and avoidance of predators and parasitoids.
Although some Catocala species appear to be reasonable candidates for
studying foodplant competition — particularly the C. ovata specialists
habilis Judith, obscura, and residua — no experimental evidence is yet
at hand (Sargent, 1982, conducted infra and interspecific rearings of
neogama and retecta larvae on Cary a ovata, but those data do not address
foodplant competition among Catocala , as neogama does not feed on
Cary a in the field). Certainly, the higher parasitism rates for epione,
palaeogama and retecta (whose larvae rest exposed on bark) compared to
habilis Judith, and residua (whose larvae rest concealed under shags)
underscores the importance of resting site to possible larval competition
for antipredator niche space (Table 9).

Roth Sargent (1977, 1982) and Schweitzer (1982a, 1982b) consider
Catocala not food-limited, with larval numbers being held low relative to
leaf availability by unknown factors, probably predators and parasites.
This is a reasonable assumption in years of low Catocala abundance, and
perhaps whenever large trees are being considered, but seems untenable
in years of peak larval abundance, and for most (if not all) times on small
trees. One factor alone, the disdain of larvae for basal leaflets (Table 8;
Gall, 1987), reduces the effective preferred leaf surface area on all
hickories by perhaps as much as one fourth or one third.

214

J. Res. Lepid.

Table 13. Right forewing lengths (in mm) of wild-collected adult Connecticut
Catocala specimens, as a function of species, sex, and severity of gypsy
moth defoliation during the year of collection. ANOVA table shows that
peak years of gypsy moth defoliation produce smaller adult Catocala. See
text for discussion.

Level of Gypsy Moth Infestation
At Peak Below Peak
mean mean

Species

Foodplant

Sex

N wingspan N

wingspan

arnica

Quercus

female

11

19.2

3

20.7

male

10

19.3

6

20.7

blandula

Rosaceae

female

8

21.9

3

23.7

male

9

20.1

3

22.0

connubialis

Quercus

female

3

18.7

1

19.0

male

4

18.3

3

19.7

crataegi

Rosaceae

female

11

20.5

1

21.0

male

6

19.2

4

19.5

ilia

Quercus

female

16

34.9

5

37.6

male

11

35.5

7

37.1

judith

Carya

female

8

23.9

2

26.5

male

8

25.2

5

26.0

paiaeogama

Carya

female

7

34.0

3

35.3

male

6

32.2

5

33.2

residua

Carya

female

6

32.8

3

34.0

male

4

31.3

8

33.0

retecta

Carya

female

5

32.8

4

33.8

male

6

31.0

14

33.8

Source

df

SS

MS

F

P

sex

1

0.025

0.025

6.90

0.009

gypsy

1

0.104

0.104

28.90

< 0.001

species

8

9.493

1.187

328.90

< 0.001

sex*gypsy

1

0.000

0.000

0.01

0.937

species*sex

8

0.041

0.005

1.43

0.185

species*gypsy

8

0.011

0.001

0.38

0.931

species*sex*gypsy

8

0.015

0.002

0.51

0.846

residual

183

0.660

0.004

29(3):195-216, 1990(91)

215

A relevant factor to intra- and interspecific Catocala larval interactions
is the influence of the gypsy moth ( Lymantria dispar L.), an introduced
pest that periodically defoliates deciduous forests in North America.
Gypsy moth larvae defoliated many canopy trees locally at my West Rock
study sites throughout 1980 and especially 1981. In 1981, defoliation
was essentially complete for all tree species by the first few days of June.
Defoliation of C. tomentosa occured well before other Cary a species,
although by early June most of the C. ovata had been stripped of 90-100
percent of their leaves (most oaks had been stripped by late May).

In 1981, only epione among the Juglandacae feeders would have been
able to pupate in numbers prior to the brunt of the gypsy moth defoliation
(but note the earlier defoliation of C. tomentosa , which epione uses
heavily). On 2 June 1981, 1 collected 31 Catocala larvae, representing all
six of the West Rock Cary a -feeders, from 10 C. ovata trees from which the
foliage was nearly or entirely stripped by gypsy moth larvae. Only 5 of
the 31 larvae were in the final instar; with little question, the remaining
26 would either have died from starvation within a few days, or have been
forced to pupate at subnormal larval size.

Indeed, wild adult Catocala collected at West Rock and North Stam-
ford, Connecticut, during the years of peak gypsy moth abundance were
significantly smaller than adults taken in the same areas in years when
gypsy moths were scarce (Table 13). This adult size difference is almost
certainly the result of forced early larval pupation in response to lack of
foliage. Note that the peak years for gypsy moth abundance at both
Connecticut sites were also peak years for Catocala abundance, and so
the observed dwarfing could perhaps in part be due to increased intraspe-
cific competition between Catocala larvae. However, dwarfed adults are
not associated with peak Catocala years in other localities in the United
States where the gypsy moth is absent (personal observations in Arkan-
sas and Tennessee). Additionally, collecting of Catocala adults at
artificial bait sources was extraordinarily successful in the peak gypsy
moth years compared to other years, especially at the beginning of the
flight period for each Catocala species. Again, this likely indicates that
adult moths had to bolster their reduced somatic reserves carried over
from larval feeding, prior to mating and oviposition (my unpublished
field notes and those of D. Schweitzer).

Acknowledgements. I thank Richard Harrison, David Hawks, Charles Remington,
James Rodman, Dale Schweitzer, and Bruce Tiffney for many helpful discussions
about Catocala , and John Hartigan for statistical advice. Dale Schweitzer and
David Furth helped to collect and rear larvae. Financial assistance was provided
by the George D. Harris Foundation, the E. Tappan Stannard Fund, a Sigma Xi
RESA grant, and a Yale University Prize Teaching Fellowship. This paper was
drawn from a dissertation submitted to Yale University in partial fulfillment for
the degree Ph.D. Philosophy.

216

J. Res. Lepid.

Literature Cited

Gall, L.F. 1987. Leaflet position influences caterpillar feeding and development.
- Oikos 49: 172-176.

-. 1991a. Evolutionary ecology of sympatric Catocala moths (Lepidoptera:

Noctuidae). I. Experiments on larval foodplant specificity. - J. Res. Lepid., 29:
173-194.

. 1991b. Evolutionary ecology of sympatric Catocala moths (Lepidoptera:

Noctuidae). III. Experiments on female oviposition preference. - J. Res. Lepid.,
29: 217-233.

Lawlor, L.R. 1980. Overlap, similarity, and competition coefficients. - Ecology 61:
245-251.

Miller, W.A. 1977. Catocala (Noctuidae) species taken in Clay County, Tennessee.
-J. Lepid. Soc. 31: 197-202.

Rowley, R.R. 1909. Another season with Catocalae. - Entom. News 20: 127-135.
Rowley, R.R., and L. Berry. 1910. Further study of the Catocalae. - Entom. News
21: 104-116.

Sargent, T.D. 1976. Legion of Night: the Underwing Moths. - Univ. Mass. Press,
Amherst. 222 pp.

. 1977. Studies on the Catocala (Noctuidae) of southern New England. V. The

records of Sidney A. Hessel from Washington, Connecticut, 1961-1973. - J.
Lepid. Soc. 31: 1-16.-

. 1982. Studies on the Catocala (Noctuidae) of southern New England. VI. The

“pairing” of C. neogama and C. retecta. - J. Lepid. Soc. 36: 42-53.

SAS Institute Inc. 1985a. SAS User’s Guide: Basics, Version 5 Edition. - SAS
Institute Inc., Cary, North Carolina. 1290 pp.

. 1985b. SAS User’s Guide: Statistics, Version 5 Edition. - SAS Institute Inc.,

Cary, North Carolina. 956 pp.

Schweitzer, D.F. 1982a. The larva and status of Catocala pretiosa (Noctuidae),
with designation of a lectotype. - J. Lepid. Soc. 36: 18-30.

% 1982b. Field observations of foodplant overlap among sympatric Catocala

feeding on Juglandaceae. - J. Lepid. Soc. 36: 256-263.

. 1987. Catocala pretiosa, the precious underwing moth: results of a global

status survey, with a recommendation for retention in category 2. - Status
Survey Report to the US Fish & Wildlife Service, (Newton Corner, MA). 24 pp.
Sokal, R.R., and F.J. Rohlf. 1982. Biometry (2nd Ed.). -Freeman, San Francisco.
859 pp.

Stanton, M.L. 1982. Searching in a patchy environment: foodplant selection by
Colias p. eriphyle butterflies. - Ecology 63: 839-853.

Wilson, L.F. 1975. Notes on the biology and parasitoids of the sweet fern
underwing (Lepidoptera: Noctuidae) in Michigan. - Great Lakes Entom. 8: 145-
153.

Journal of Research on the Lepidoptera

29(3):217-233, 1990(91)

Evolutionary Ecology of Sympatric Catocala Moths
(Lepidoptera: Noctuidae)

III. Experiments on Female Oviposition Preference

Lawrence F, Gall

Entomology Division, Peabody Museum of Natural History, Yale University, New Haven,
CT 00511 USA

Abstract. Fecundities and lifespans are documented for 56 species of
Nearctic Catocala moths, and detailed analysis of oviposition prefer-
ence is presented for 9 species whose larvae feed on trees in the plant
family Juglandaceae. Among these 9 species, 1 restricted oviposition in
the field to walnuts ( Juglans ), 1 to pecans (section §Apocarya of Carya ),
and 7 to hickories sensu strictu (section §Eucarya of Carya). Shagbark
hickory, Carya ovata , was greatly preferred by all 7 Carya section
§Eucarya feeders. Large trees were favored over small trees, and there
were no temporal differences in timing of oviposition. Oviposition
specificity became less crisp as the season progressed and females aged.
Patterns in oviposition preference were similar in two geographically
distant sites in North America. Laboratory arena tests of preference
concorded only modestly well with the field studies, due to the con-
founding effects of female age, “shagginess" of the oviposition sub-
strate, and arena size.

Introduction

This paper examines the foodplantdinked oviposition biologies of
North American Catocala Schrank moths, and is the third in a series
addressing the evolution of foodplant use in this speciose genus. Fecun-
dity and longevity data are tabulated herein for 56 of the 108 Nearctic
Catocala species, from all foodplant groups, and more detailed field and
laboratory studies of oviposition specificity are presented for 10 of the 25
Juglandaceae-feeding species. Prior papers in this series treat choice and
no-choice laboratory foodplant preference tests with Catocala larvae
(Gall, 1991a), and the distributions of wild larvae on their foodplants
(Gall, 1991b).

Materials and Methods
OVIPOSITION AEENA PROTOCOL
An arena design similar to that for larvae (see Gall, 1987, 1991a) was used to
assess oviposition preferences of female Catocala. Since females of the
Juglandaceae-feeding Catocala deposit eggs under exfoliating bark and cracks,
pieces of bark (ca. 5x15 centimeters) from different trees were used as the food
choices. White oak ( Quercus alba L.) bark was offered as a Type I control choice
item ( sensu Gall, 1991a) in each arena. Females confined to such arenas were fed
daily with a honey/syrup solution, diluted by at least one-half to two-thirds with
water. Arenas were also frequently sprayed with water to prevent desiccation.
Females were generally allowed to live out their lives in arenas, but were

218

J. Res. Lepid.

removed when they clearly could no longer make choices among the food items
(i.e., legs lost, and/or wings battered).

I used circular plastic containers and small paper bags (ca. 15 cm diam, 15 cm
deep) as arenas initially in this work, and subsequently switched to larger paper
bags and wire mesh arenas (ca. 30x20x15 cm). Bark samples were taped into the
plastic arenas, and sewn into the others. Developing suitable arena designs
proved to be a most recalcitrant aspect of these laboratory studies (see below).
This stands in contrast to the larval work, in which differing arena geometries
and rearing containers had little effect on assays of foodplant preference.

FIELD METHODS FOR STUDYING OVIPOSITION BY WILD
FEMALES

Schweitzer (1982) has described the techniques used here to locate ovipositing
female Catocala. One first carefully scans a trunk of a foodplant tree using a
powerful focused-beam search lantern and head lamp. Most ovipositing Catocala
are readily visible, and sit facing head up on the trunk, with their ovipositors
inserted deep into cracks or under shags (some species e.g., subnata Grote, tend
to oviposit on bark near the base of their foodplants, but most species oviposit
from about waist level to 3-5 m up the trunk). After scanning the trunk and major
branches, the trunk is then hit several times with a net or stick to flush remaining
moths. Many Catocala so disturbed from oviposition activity spiral downward,
alight, and commence oviposition again on the same tree trunk within a short
time. Hitting the trunk thus proved to be »a reliable procedure for enticing
unreachable females to within net range. Ovipositing female Catocala are
clearly intent on the task at hand, being largely oblivious to the bright collecting
lights and attendant noise when one is searching trees (twig snapping, etc.). This
contrasts sharply with their wary behavior at artificial bait sources at night, and
flightiness during the daytime.

I conducted several pilot trials using an infra-red “sniper scope” to observe
female Catocala , but abandoned this method early on in favor of lanterns and
head lamps, for simplicity. Doubtless such infra-red viewing devices would prove
useful in examining the oviposition behavior of other night-flying Lepidoptera
that are disturbed by light in the visible spectral range.

DESCRIPTIONS OF STUDY SITES

Searches for ovipositing Catocala were carried out from 1980-1987 at four
localities in Connecticut and Tennessee (USA). Three of the Connecticut sites
were in New Haven County: West Rock, Brooksvale Park (Hamden), and
Southbury; the other, West Redding, was in Fairfield County. Data from the
Connecticut localities were recorded separately, but are pooled here for analysis
as the biological patterns proved consonant among sites. The Tennessee searches
were conducted during 1985 and 1986 at Celina (Clay County), a locality with a
rich Juglandaceae-feeding Catocala fauna (Miller, 1977), and where many of the
more southerly species replace or are numerically commoner than those present
in Connecticut.

The collecting routes were all selected to include a representative array of
available juglandaceous foodplants. At West Rock, West Redding, and Celina the
routes were largely woodland paths; at Brooksvale Park and Southbury, the
routes were woodland margins. However, the routes at each locality included
both types of habitat. Small to medium (ca. 5-15 cm dbh) hickories and walnuts

29(3):217-233, 1990(91)

219

predominated at West Rock, while large (>25 cm dbh) trees occured at the other
habitats. Shagbark hickory, Carya ovata Mill. K. Koch of section §Eucarya DC.,
was the most prevalent hickory at all sites, although West Rock had a substan-
tially larger number of pignut hickory (C. glabra Mill. Sweet of section §Eucarya)
than did the other localities. Shellbark hickory (C. laciniosa Michx. Loud, of
section §Eucarya ) occured only at Celina, and this locality was the only one
lacking butternut ( Juglans cinerea L.).

TREE AND MOTH PARAMETERS

For each tree surveyed, I recorded the collecting locality and species identity.
For most I also recorded size as: small (5-10 cm dbh), medium (10-20 cm dbh) or
large (>20 cm dbh). The majority of trees sampled greatly exceeded 20 cm dbh,
except at West Rock. I recorded the species identity of each moth, but did not
index age because of lack of mark-recapture data on Catocala with which one
would calibrate age. Only those moths which were positively identifiable to
species, and whose ovipositors were clearly inserted into bark crevices were
scored as ovipositing (these were the vast majority of all observations). At each
locality, I worked a predefined collecting route. I usually repeated the route twice
(occasionally three or four times) per night. Since moths were removed from trees
each time, and at least 30-45 minutes separated repeat checks, I treated all tree
samples as independent observations. Most sampling was conducted between
0830 and 1130 hours. Statistical testing follows Gall (1991a).

Results

LABORATORY STUDIES

1. Longevity and Fecundity of Female Catocala

Table 1 presents longevity (in days) and fecundity data for 56 Catocala
species that I have confined over the years for arena tests or for securing
eggs for subsequent rearing (information is presented only for those
moths that did in fact lay eggs, which was the great majority of all
females). Few of the moths whose data are shown were freshly eclosed
when captured. Thus, the Table 1 data must be considered minimum
estimates of longevities and fecundities of wild Catocala females. In
addition, Table 1 is somewhat biased toward more mature female moths,
since newly eclosed females often will refuse to lay and/or will produce
infertile eggs, and my primary reason for confining female Catocala
always was to obtain breeding/experimental stock for the subsequent
larval field season.

Considering all 56 species, there were significant positive correlations
between longevity and moth size (Spearman rank: r=+0.22, n=211,
p<.01), and fecundity and moth size (r~+0.45, n=253, pc. 01; size indexed
by forewing length taken from species accounts in Sargent, 1976).
Within each of the six foodplant groups represented here by more than
one taxon, the correlations between longevity-size and fecundity-size
were again positive, but none of the correlations was significant. There
were no significant differences in longevity as a function of foodplant
group, but fecundities of females in the Salicaceae, Juglandaceae, and
Myricaceae groups tended to be greater than fecundities of females in the

220

J. Res. Lepid.

Table 1. Longevities and fecundities of female Catocala in oviposition containers
(arenas and otherwise). Means, maxima, and sample sizes indicated for
both parameters. Means by foodplant family given below main Table;
families connected by the same vertical bar are not significantly different at
the 0.05 level (Tukey HSD tests). Egg size gives mean maximal diameter
and mean maximal vertical height of sample of 10 eggs, rounded to the
nearest 5 micrometer units (scale: 50 micrometer units per mm).

Species

Egg Size
diameter height

Number Eggs Laid
mean n max

Lifespan
mean n max

Foodplant

Family

andromedae

50

30

53.2

6

159

13.4

5

19

Ericaceae

gracilis

50

30

32.3

3

68

16.0

2

26

Ericaceae

louisae

45

30

20.5

2

21

9.0

1

9

Ericaceae

sordida

50

30

30.5

6

57

10.3

3

13

Ericaceae

arnica

45

20

46.0

3

72

16.5

2

18

Fagaceae

coccinata

70

30

77.2

6

116

11.5

4

14

Fagaceae

connuhialis

50

20

54.8

4

120

11.0

3

13

Fagaceae

herodias

65

40

118.4

5

198

16.0

2

20

Fagaceae

ilia

70

40

122.3

7

312

14.6

5

19

Fagaceae

lineella

45

20

89.2

5

150

18.7

3

21

Fagaceae

micronympha

55

20

53.8

4

104

8.8

4

17

Fagaceae

similis

50

20

52.1

7

109

12.3

3

18

Fagaceae

angusi

50

20

463.5

4

1493

31.8

4

35

Juglandaceae

censors

40

30

123.0

2

151

Juglandaceae

dejecta

65

20

219.5

2

258

23.5

2

28

Juglandaceae

epione

55

45

85.6

5

136

11.4

5

18

Juglandaceae

flebilis

50

20

44.0

5

109

21.0

5

42

Juglandaceae

habilis

45

20

100.3

4

248

28.3

3

38

Juglandaceae

insolabilis

60

20

61.8

4

190

10.0

5

15

Juglandaceae

judith

50

20

59.3

4

169

13.3

3

25

Juglandaceae

lacrymosa

55

20

169.3

7

381

25.4

7

40

Juglandaceae

luctuosa

50

20

304.0

1

304

26.0

1

26

Juglandaceae

maestosa

50

20

224.5

2

388

18.0

2

20

Juglandaceae

nebuiosa

55

25

58.2

5

153

22.8

4

29

Juglandaceae

neogama

55

20

377.8

4

1102

21.5

6

32

Juglandaceae

obscura

55

20

291.5

11

621

21.4

9

39

Juglandaceae

palaeogama

55

20

317.9

14

694

22.9

12

52

Juglandaceae

piatrix

50

25

164.0

5

400

14.0

2

20

Juglandaceae

residua

55

20

352.7

3

480

26.3

6

37

Juglandaceae

retecta

55

20

277.0

14

625

18.7

13

41

Juglandaceae

robinsoni

45

20

46.0

1

46

41.0

1

41

Juglandaceae

serena

50

20

72.0

1

72

38.0

1

38

Juglandaceae

subnata

60

25

237.3

4

506

12.3

4

15

Juglandaceae

ulalume

50

20

44.0

1

44

9.0

1

9

Juglandaceae

vidua

50

20

104.5

2

165

Juglandaceae

innubens

50

35

59.5

4

150

11.0

3

19

Leguminoseae

antinympha

45

30

98.5

4

226

13.5

2

15

Myricaceae

badia

45

30

96.0

4

150

10.5

2

15

Myricaceae

coelebs

50

30

171.7

3

405

Myricaceae

muliercula

40

30

248.0

2

270

18.0

2

22

Myricaceae

29(3):217-233, 1990(91)

221

Egg Size Number Eggs Laid Lifespan Foodplant

Species

diameter

height

mean

n

max

mean

n

max

Family

hlandula

55

20

29.2

5

49

10.0

4

14

Rosaceae

crataegi

50

20

80.0

3

92

13.5

2

16

Rosaceae

grynea

50

20

62.6

5

99

15.3

7

27

Rosaceae

mira

55

20

49.6

7

79

9.2

6

17

Rosaceae

praeclara

50

20

57.5

13

278

14.8

6

34

Rosaceae

pretiosa

55

20

31.1

7

72

13.5

10

19

Rosaceae

uitronia

55

20

84.4

7

193

17.3

7

24

Rosaceae

amatrix

45

35

452.0

2

600

20.0

3

32

Salicaceae

briseis

60

35

74.5

2

99

Salicaceae

cara

50

35

286.7

3

538

15.9

9

27

Salicaceae

carissima

45

30

42.0

1

42

63.0

1

63

Salicaceae

concumbens

55

40

157.8

5

241

15.5

4

21

Salicaceae

parta

55

35

112.0

2

174

25.0

1

25

Salicaceae

relicta

60

45

153.3

4

202

11.8

4

25

Salicaceae

unijuga

65

40

75.0

2

100

35.5

2

37

Salicaceae

cerogama

45

35

73.5

2

107

10.0

3

14

Tiliaceae

Means:

214.7

20.9

Juglandaceae

177.7

19.6

Salicaceae

137.6

14.0

Myricaceae

80.4

13.3

Fagaceae

55.4

13.7

Rosaceae

37.7

12.6

Ericaceae

Fagaceae, Rosaceae, and Ericaceae groups (see Table 1). Note that
nearly all of the Salicaceae, Juglandaceae, and Myricacae feeding species
in Table 1 are larger than the corresponding Fagaceae, Rosaceae, and

Ericaceae feeders.

2. Arena Tests

Table 2 gives the results of 20 arena experiments with Juglandaceae
feeders. In 11 of the arenas, oviposition on either the Type I control
Quercus alba and/or the arena itself (noted in Table 2 as “Fraction
Haphazard”) accounted for nearly 20 percent or more of all eggs laid.
Using wire mesh cages greatly reduced oviposition on the arenas them-
selves, but nevertheless did not curb oviposition on the Type I control.

Haphazard laboratory oviposition on plants not in the preferred
foodplant family is squarely at odds with field oviposition results with
these same moth species (see below under Results and Discussion). I
therefore consider only 9 of the 20 arenas to be “successful,” these being
starred in Table 2 and characterized by less than five percent of all eggs
being laid on arenas and Type I controls (the piatrix Grote and subnata
arenas are considered successful despite ca. 10 percent hapazard ovipo-
sition, since the numbers of misplaced eggs were very low).

Of the seven successful arenas with Cary a feeders, five are for
palaeogama Guenee, and single females represent each of residua

T able 2. Oviposition arena results with Juglandaceae-feeding Catocala. Values in T able are percents of eggs laid on choice items. “Successful'
arenas starred. See text for elaboration.

222

J. Res. Lepid.

%

O

£

s

£

c

o

o

CD

©

C

o

CD

c

JO

CD

0)

=3

2

CD

o

I

o

03

D)

03

03

CD

O)

03

O)

o>

03

O)

03

O)

03

CL

O)

03

03

O)

03

CD

3

©

3

3

3

3

3

3

3

3

3

3

3

3

©

3

3

U

©

3

3

3

©

< < <

<

<

< < < < <

<

< < < CD

<

<

©

©

< < <

Q

00

r-

00

CO

h*.

CO

fv

00

03

h.

co

00

00

'f

CD

Q

00

00

rv

00

CM

T—

CM

CM

CM

CM

CM

CM

CM

CM

CM

CM

CM

1”

CM

CM

CM

.9

.9

.9

.9

©

c

.9

©

to

O)

©

©

©

to

to

to

03

©

©

©

03

03

©

©

©

©

to

©

©

©

rrt

V-

©

©

©

m

rrt

©

*—

©

©

L.

CO

CL JD

<:

£

>

CL

CL

cl n

>

'3:

3: j5

JO

>

>

>

<

-4— »

0)

CL

i

'3S

■a

q

CO

r-

03

T=~

O

CO

O

CM

o

U*3

03

o

CM

|v.

O

|v

q

c

0

©

q

q

m

N

cO

sz

C3

CO

d

00

cd

CM

cd

d

d

q

d

d

i —

d

00

CM

q

w

N

©

jC

CM

CM

d

CO

CM

CM

T“

03

CO

IV.

CO

CM

CM

0

©

LO

T=

T”

Q,

CL

CO

©

LL

©

X

X

©

o

CO

00

00

03

i —

’M"

T —

co

r--

o

CO

CM

O

Lf3

|v.

CD

03

o

__

©

CM

O

03

00

CM

00

CO

03

LO

00

o

00

00

CM

LD

T—

CD

©

0

’M'

>

o

■M-

CO

■M"

CM

CO

CM

CD

CM

CM

CD

CM

IT”

O

>

0

©

o

o

CL

CO

f=~

©0

©

f“

LO

LO

o

O

o

o

O

o

CM

o

O

q

CM

O

O

O

c

CO

O

O

©

03

q

d

d

d

CM

d

d

00

d

d

d

T“

03

d

d

d

o

©

06

d

d

L.

CO

■J—

"M"

CM

1.

CM

<

o

©

<

in

o

©

o

CM

T-

o

CO

O

CM

o

CM

LO

03

o

o

q

o

IV

00

c

©

2

0

w

LD

LD

q

c

d

CM

d

00

cd

d

cd

d

CM

q

d

d

d

q

00

CM

q

©

©

C

q

CM

d

o

i —

CM

f— *

03

T=”

rv

CM

CM

c

O

CM

T”

1™

o

O

O

D)

©

CO

“3

.©

00

£

r-

|v

q

o

q

■M"

r--.

q

q

o

o

q

cq

LD

q

LD

C

O

&

1

C3

O

O

03

<5

cd

T— -

d

T-’

d

d

CM

03

cd

T-"

d

d

CM

CM

q

q

03

O)

©

q

d

d

|v

£

CM

O

JO

©

©

0

?

8

LD

©

5

in

CM

id

CO

d

CO

03

o

d

03

CM

00

in

q

d

q

CM

CM

q

LO

q

d

LD

q

CM

fv

CM

d

CO

03

s

£

0

Jugians

2

.03

IV.

cd

0

d

CD

©3

CO

CO

CO

fv.

CO

CO

LO

CM

CD

■M'

CO

CO

CD

CM

O

CO

|V

CM

iS

■M-

o

-M;

■m-

CO

CD

00

CM

O

03

^r

o

T“

q

q

2

ID

ID

CO

CO

T-l

cd

03

T 1

cd

q

CM

LO

q

CM

hi

q

cd

CM

cd

cd

03

0

©

d

q

LO

o

CM

CO

■J—

CM

"M"

co

LO

CM

co

O

O

O

O

o

O

00

co

00

00

5

5

5

~o

o

o

CD

O

00

<

o

CO

CD

5

<

<5

CD

5

o

<

cb

5

©

CO

cb

5

©

o

©

E

©

9

©

6

©

CD

©

E

©

o

©

E

©

9

©

E

©

o

CO

DQ

O

03

o

oo

<r

s

CD

GO

Q

JO

0

0

9

0

00

X

©

00

5

DO

CO

©

©

©

©

cb

O) 03 O) CD 03 CD 03

©

©

i

03

O

O

cb

E

<

A

iS

©

'o

w,

O

©

o

8

5-..

©

o

©

o

o

©

o

©

o

©

o

©

o

©

o

©

o

©

s

o

dj

o

©

0

©

CD

©

c

©

C/)

CO

CO

C/3

CO

iS

iS

J©

iS

iS

iS

,©

©

©

.©

©

O

.©

CL

-Q

-Q

Xj

-Q

-O

©

©

©

©

©

©

©

£

2

2

2

CL

©

©

CO

O

O

O

O

c

CL a CL CL a a Q.

CD

©

CL ©

*

*

■K

*

*

■K

*

*

*

29(3):217-233, 1990(91)

223

Strecker and retecta Grote. In all seven arenas, Carya ovata and C.
glabra were prefered over C. tomentosa Nutt. (pc. 05 by AN OVA for
palaeogama ; G=296.18, df=2, pc. 01 for residua', G=96.75, df=2, pc. 01 for
retecta ). In the remaining 2 arenas, the J uglans -feeding piatrix pre-
ferred Juglans (G=31.71, df=2, pc. 01), and the Carya -feeding subnata
likewise preferred C. cordiformis (Wang.) K. Koch of section §Apocarya
DC. (G=9.50, df=2, pc. 05).

FIELD STUDIES

1. Tree Distributions

I surveyed a total of 1552 hickories and walnuts for Catocala from 1980
through 1987. Table 3 gives relevant summaries of the trees at each
locality, and breaks down the Connecticut trees as a function of size.
Since the same routes were repeated each year, there was by definition
no year to year variation in the relative frequencies of juglandaceous
foodplants sampled at each locality. Although an unbiased analysis of
tree size is precluded by repeat sampling of the same trees each year, tree
size clearly differed as a function of tree species (G=38.02, df=10, pc.01,
treating repeats as independent).

2. Opposition as a Function of Foodplant Genus

I recorded 259 ovipositing female Catocala from these 1552 trees (Table
4). A total of 10 females were noted ovipositing on Juglans, 13 on Carya
cordiformis of section §Apocarya, and 236 on Carya of section §Eucarya.

Females of subnata oviposited in Connecticut on C. cordiformis (one
observation also on Juglans cinerea\ subnata is rare at the Celina,

Table 3. Numbers of juglandaceous trees searched for ovipositing female Catocala.
Above, by state; below, tree size distributions for Connecticut localities.

Carya

Carya

Carya

Carya

Carya

Juglans

Juglans

State:

ovata

glabra

tomentosa laciniosa

cordiformis

cinerea

nigra

Total

Connecticut

640

361

68

184

66

39

1358

Tennessee

97

36

4

26

6

25

194

Total

737

397

72

26

190

66

64

1552

Carya

Carya

Carya

Carya

Carya

Juglans

Juglans

Tree Size:

ovata

glabra ■

tomentosa laciniosa

cordiformis

cinerea

nigra

Total

Large

383

181

27

120

41

18

770

Medium

137

115

25

32

10

4

323

Small

71

35

13

26

10

2

157

Total

591

331

65

178

61

24

1250

Table 4. Field oviposition records for female Juglandaceae-feeding Catocala, by state and tree species. P = significance level for preference
of favored foodplant genus versus all other juglandaceous foods (G-tests or exact multinomial probabilities).

224

J. Res. Lepid.

c

J2

Q.

TD

O

o

CO

TO eg

It

<2 to
to eg
eg to

■P CD
§>.TO
P O
.to

§> 1

to ^
O ^
o
o

eg

2. .8

<5 .5

eg

co

5 Q
o to
eg 0)
O 5

p

2 5

OP
eg to
O D)

2,13

P eg
TO £
O O

3

o

o

o

c

c

o

o

(H r- t- T" CD

eg o o o o

V v v

as.

s,s,§,

TO TO TO

o o o

, _ TO TO C3

uj m uj uj lu

too coj €03 coo

TO TO
CJ CJ
TO TO

TO TO TO .

o o o o o

oo T- cd to in

oo o CO t-

o o o o o

o o o o o

o o t- o o

0 0 1^-00

a o h'-

00 r- ^ ^ CO
00 00 00 1-

TO

5

TO

5 g5 TO

TO
CJ

75 CO -ffi

§ S5 5

TO -g
TO -2

co

P CO

§ P S o5 to
•2, o Cl. to c!

a q
v v

8

O

a

5,

O

©

CO CO

2j

TO §>
O p

oo r-~

o eg

uo

o o

o o

o o

TO

TO TO

TO

P

TO

C0

CM

CD

00

5

o eg

§f

eg

co

TO eg
TO TO

TT

J2 TO

| 5

§> TO
P O
.CO

1 1

o ^

8

TO

TO -S

o p

TO

co

TO p
TO 0

o 5

Q

2 5

OP
TO TO
O O)

TO TO

O TO
TO ^
O O

O) T- LO

t W r-

TO TO TO TO TO
to1 to4 ^ ib cb
TO TO TO TO TO

O CJ CJ O CJ

P P P ,=3 P

Uj Uj Uj Uj Uj
cco too coa COj COj

o o a o o

0©000
0 0 0 0 0

TO TO TO TO TO

to4 to4 St4 to4 ^

TO TO TO TO TO

o o o o o

eg oo oo to

o o o o o

o o o o o

o o o o o

o o

eg oo

o o

m

■2 TO

TO

^ 95 eg

TO

to

TO P TO $ 0
5; Q Cl k (/)

TO

TO

TO

§>

eg

eg

eg

eg

29(3):217-233, 1990(91)

225

Tennessee locality). Females of neogama Smith oviposited on Juglans in
both Connecticut and Tennessee. Females of flebilis Grote, Judith
Strecker, obscura Strecker, paiaeogama, residua , serena Edwards, and
retecta all oviposited on Carya in both Connecticut and Tennessee. No
ovipositing female nebulosa Edwards were observed in Tennessee, though
I collected the species commonly there at lights and on tree trunks during
the daytime.

In Connecticut, C. ouata was significantly preferred over other Carya,
by Judith , obscura , paiaeogama , and residua (pc. 05 for each Catocala ,
exact binomial probabilities; p=0.053 for retecta). Females of obscura ,
paiaeogama , and residua preferred to oviposit on large trees (pc. 05 for
each as above, test is large trees versus all others; for Judith , p=.Q91). In
Tennessee, C. ovata was similarly preferred by obscura and residua
(pc. 05 for each as above; for paiaeogama , p=.103). There were too few
observations to draw firm conclusions about oviposition specificity in
flebilis and serena .

No Catocala species showed differences in oviposition preference as a
function of year (p>.25 in each case by G- tests, controlling for state). Two
Connecticut ovipositions represent what I consider to be “mistakes” — a
paiaeogama ovipositing on C. cordiformis, and a subnata on Juglans
cinerea. These two observations are discussed later in the context of
female moth age.

3. Temporal Patterns

Table 5 presents the mean time of night for oviposition by the seven
Connecticut Juglandaceae-feeding Catocala species. There were no dif-
ferences in timing of oviposition among any of these species (p>.15 by
ANOVA considering all, p>.25 for just the Carya feeders). Curiously, I
found no ovipositing habilis Grote in this study, although adults were
generally common at bait and light on the same nights in both Connecti-
cut and Tennessee. Since my tree searches ended prior to midnight, it
may be that this species oviposits

Table 5. Mean time of night (EST) for ovi-
position by wild female Jugland-
aceae-feeding Catocala in Con-
necticut. All species oviposit at
similar times.

Time of Oviposition

Species

Mean

SD

N

judith

20:56

00:31

4

subnata

21:01

00:28

11

paiaeogama

21:09

00:36

88

residua

21:10

00:22

22

obscura

21:13

00:43

23

retecta

21:15

00:35

12

neogama

21:50

00:34

6

principally late at night. Late
night to pre-dawn flight character-
izes a number of other Catocala ,
such as the oak-feeding herodias
Strecker (personal observations
of myself, D. Hawks, and D.
Schweitzer). Sampling was con-
ducted rather late in each season
to expect epione Drury (adults fly
in late June and early July in
Southern Connecticut), and I sus-
pect from other work with epione
females, and the distribution of
its young larvae in the field (Gall,

226

J. Res. Lepid.

Table 6. Mean numbers of ovipositing female Juglandaeeae-feeding Catocala per
tree in Connecticut, as a function of year and foodplant species. Note
decline in numbers from 1980 through 1984 on all tree species.

Species

1980

Number of Females Per Foodplant Tree
1981 1982 1983 1984

1987

Carya glabra

0.12

0.03

0.00

0.00

0.00

0.00

Carya ovata

0.45

0.20

0.00

0.02

0.36

0.25

Carya tomentosa

0.22

0.04

0.00

0.00

0.00

0.00

Carya cordiformis

0.15

0.04

0.00

0.00

0.06

0.00

Jug Ians sp.

0.31

0.08

0.00

0.00

0.00

0.00

Total Trees Surveyed

371

674

125

55

87

46

1991b) that this species oviposits on small trees, near the base of the

trunk.

4. Moth Densities on Foodplants

The mean number of ovipositing females per tree is presented in Table
6, by foodplant and year. As with the larval field data for the same
Catocala species (Gall, 199 lb), these oviposition records reflect the sharp
drop in adult Catocala numbers at light trap samples in Connecticut
during 1980-1984. The decline in female abundance from 1980-1984 in
Table 6 is highly significant for each foodplant genus (pc. 01 by G- tests,
trees split each year into two classes: having no females, having one or
more female).

Table 7 shows plots of the number of ovipositin gpalaeogama, residua ,
and obscura per C. ovata in Connecticut and Tennessee, for years when
moth abundance was high. The variance to mean ratios are not substan-
tially greater than one, no distribution being significantly different from
poisson (i.e., random; p>.25 for each by exact binomial tests). Stated
another way, ovipositing females of the same species did not congregate
on individual C. ovata trees (compare these distributions to those for
larvae of the same Catocala species in Gall, 1991b).

Table 7 also shows similar plots of the number of all ovipositing
Catocala per tree. Again, the variance to mean ratios showed no clear
trends, and the distributions were similarly random (p>.25 in each).
Thus, there was no tendency for ovipositing Catocala females on the
whole to congregate on individual C. ovata or C. glabra trees. The pattern
was the same for the other Cary a foodplants, for which far fewer
observations were available (virtually all records being of single ovipos-
iting Catocala , or none). Therefore, it appears that individual Catocala
females select suitable oviposition sites without regard to the presence
or absence of other females. Since there is a superabundance of such sites

29(3):217-233, 1990(91)

227

Table 7. Numbers of ovipositing female Juglandaceae-feeding Catocala per individual tree:
by year, Catocala species, collection locality, and foodplant species. Variance to
mean ratios for all distributions are near 1 .0, indicating no clumping of females on
trees (no distribution is significantly different from random i.e., Poisson at the 0.05
level; G-tests or exact multinomial probabilities).

Species

State

Year

Foodplant

0

Number of Moths

1 2 3 >3

Totals

Moths Trees

Moths/Tree
Mean Var/
Mean

obscura

CT

1980

Carya ovata

80

11

2

1

0

18

94

0.19

1.38

obscura

CT

1981

Carya ovata

247

8

1

0

0

10

256

0.04

1.16

palaeogama

CT

1980

Carya glabra

96

9

1

0

0

11

106

0.10

1.09

palaeogama

CT

1980

Carya ovata

70

19

3

2

0

31

94

0.33

1.20

palaeogama

CT

1981

Carya ovata

235

19

1

0

0

21

256

0.08

1.01

palaeogama

CT

1984

Carya ovata

44

7

2

0

0

11

53

0.21

1.18

palaeogama

TN

1980

Carya ovata

48

5

1

0

0

7

54

0.13

1.18

residua

CT

1980

Carya ovata

85

8

1

0

0

10

94

0.11

1.11

residua

CT

1981

Carya ovata

240

15

1

0

0

17

256

0.07

1.06

residua

TN

1986

Carya ovata

45

8

1

0

0

10

54

0.19

1.03

all taxa

CT

1980

Carya glabra

93

12

1

0

0

13

106

0.13

1.02

all taxa

CT

1980

Carya ovata

50

30

8

3

3

67

94

0.71

1.35

all taxa

CT

1981

Carya ovata

207

40

7

2

0

60

256

0.23

1.20

all taxa

CT

1984

Carya ovata

39

9

5

0

0

19

53

0.36

1.19

all taxa

TN

1986

Carya ovata

35

15

4

0

0

23

54

0.43

0.94

on individual C. ouata trees, and eggs are placed far into crevices, it is
doubtful that females assess egg load on trees before ovipositing (I
regularly find eggs and eggshells of several Catocala species under the
same piece of exfoliating bark).

Implicit in the above analyses is that the spatial distribution of these
hickories was not strongly contagious, a pattern which by itself could
confound interpretation of moth densities on trees, as well as influence
conclusions about female preferences among different foodplants spe-
cies. Both intraspecific and interspecific distributions of the foodplant
trees are relevant (Courtney and Courtney, 1982, for the former; Stanton,
1982, for the latter). At the Connecticut localities of Brooksvale Park,
Southbury, and West Redding, farming and landscaping practices have
generated an approximately regular distribution of all hickories. Large
trees (often hickory and maple) are planted, or culled, at intervals along
the edges of fields. This reduces the possibility of “edge-effects” within
monotypic clumps of trees, but probably does not greatly affect interspe-
cific tree clumping. At West Rock and Celina, where neither farming nor
landscaping is now manifest, there may be some tendency for small trees
to be clumped (both intra- and interspecific stands), but medium sized
and large trees are encountered only sporadically.

228

J. Res. Lepid.

Discussion

GENERAL TRENDS IN OVIPOSITION SPECIFICITY

The field studies presented above show that ovipositing female
Juglandaceae-feeding Catocala discriminate among the three major
taxonomic divisions within Juglandaceae: neogama oviposits on Juglans;
subnata on Cary a section §Apocarya; and Judith , obscura, palaeogama,
residua , and retecta on Carya section §Eucarya. Within Carya section
§Eucarya, shagbark hickory (C. ovata ) is always preferred over other
hickories. In both Connecticut and Tennessee, two geographically
distant sites with differing juglandaceous floras, oviposition preferences
by wild female Catocala are quite close even down to the foodplant
species level.

Oviposition by both flebilis and serena, represented here by small
sample sizes, will likely also prove to be restricted to Carya (the few other
recorded ovipositions for serena are all C. ovata ; Nielsen, 1978, and in
lift.). Field data on nebulosa are presently lacking, but I have determined
a wild egg collected in Michigan on Carya cordiformis by M. Nielsen to
be nebulosa (confirmation by scanning electron microscopy; the egg SEM
profiles of nearly all Nearctic Catocala are now known), and its larvae
only ate Carya section §Apocarya in foodplant preference tests (Gall,
1991a). Thus, it seems probable that nebulosa will prove to limit
oviposition to Carya section §Apocarya, as does subnata .

In contrast to my field studies with Juglandaceae-feeding Catocala
larvae (Gall, 1991a, 1991b), I did not systematically sample non-
juglandaceous trees at the same time for ovipositing females, to allow
explicit tests of foodplant family level oviposition specificity. However,
I have certainly examined well in excess of 300 trees (most of which were
oaks) in other Catocala foodplant groups during the course of these
nightly walks, and never noted an ovipositing Juglandaceae-feeding
Catocala on these other foodplant trees. Additionally, during studies of
oviposition by Rosaceae-feeding Catocala , Schweitzer (1987, and in litt.)
has similarly failed to record other than Rosaceae feeders on those plant
arrays. Thus, there is little question that the Juglandaceae-feeding
Catocala discussed here oviposit solely on juglandaceous foodplants.

Note that Rowley & Berry (1912) did report finding eggs of the oak-
feeding ilia Cramer very rarely on C. ovata in Missouri. I have taken only
one ilia egg under a C. ovata shag (in Connecticut in 1984, among
numerous shags examined for eggs and hatched egg shells over the
years), and Schweitzer (1982) obtained exclusively Juglandaceae-feed-
ing Catocala (over 100 individuals representing 4 species) from 18 C.
ovata whose egg faunas he surveyed by placing shags into rearing sleeves
in early spring.

OVIPOSITION SPECIFICITY VS. LARVAL FOODPLANT USE

Whether female oviposition choice and larval feeding biologies agree is
of interest to studies of the evolutionary history of foodplant ecology in

29(3):217-233, 1990(91)

229

the genus Catocala. For the species treated in this paper, such concor-
dances appear to be absolute at the foodplant family level, with all 10
Juglandaceae-feeding Catocala ovipositing solely on Juglandaceae in the
field, their first instars preferring only these same plants in arena tests
(Figures 1-3 in Gall, 1991a), and the wild larvae being found only on these
same plants (Table 3 in Gall, 1991b). The concordances appear nearly as
absolute within Juglandaceae, with one taxon ( neogama ) ovipositing and
feeding as larvae in both lab and field on Juglans , one on Cary a section
§Apocarya ( subnata ), and five on Carya section §Euearya (Judith , obscura,
palaeogama, residua , retecta). Furthermore, for each of these seven
Catocala species, larval growth and survival are also uniformly best on
the foodplants preferred by wild females for oviposition (Table 4 in Gall,
1991a).

Only the laboratory oviposition specificity tests do not agree closely
with the larval feeding biologies. Most of the disagreement seems
traceable to the confounding influences of female age and bark sample
“shagginess” (see below), as the few successful oviposition arenas with
palaeogama, piatrix , residua , retecta , and subnata do match larval
feeding patterns moderately well at the foodplant genus level. Among
the southern Connecticut Carya -feeding Catocala , there is some vari-
ability in the otherwise precise correspondences between wild female
oviposition and larval foodplant specificities, but only for palaeogama
and retecta , the two most oligophagous of the taxa ( hahilis , Judith ,
obscura , and residua are all specialists linked tightly to the biology of C.
ovata ). For 1 retecta and 3 palaeogama broods, data are available on the
foodplant preferences of the females and their larval progeny (cf. Table
2 above; and Table 4 in Gall, 1991a).

The mother of retecta brood E81 was collected as she oviposited on C.
ovata. She was not tested in an oviposition arena, and her progeny ate
much more C. glabra in first instar arenas than other food choices.
Survival and growth rate of her larvae were but slightly higher on C.
ovata than on C. glabra. The mother of palaeogama brood A80 was
collected while she oviposited on C. ovata. In oviposition arenas, she laid
more eggs on C. glabra than on C. ovata. Her progeny overwhelmingly
preferred C . ovata to other Carya in first instar arena tests, and similarly
developed fastest on C. ovata in no-choice sleeve rearings. The mother of
palaeogama D80 was taken while she oviposited on C. tomentosa. She
laid most of her eggs on the arena (a paper bag) rather than on bark
samples. Her progeny ate equal amounts of C. ovata and C. tomentosa ,
but they were not reared in sleeve experiments. The mother of palaeogama
brood E80 was taken while she oviposited on C. glabra. She was not
tested in an arena, and her first instars preferred C. ovata two to one over
C. tomentosa , and ate virtually no C. glabra. Her larvae developed most
rapidly on C. ovata in sleeve rearings.

The only consistent pattern to emerge from the foregoing comparisons
is that C. ovata seems to best support larval growth, irrespective of

230

J. Res. Lepid.

female oviposition history. There is no indication that pronounced “host
races” are developed at present in the hickory-feeding Catocala who
regularly use Cary a other than the universally preferred C. ouata.
However, I suspect that local or larger-scale host races may prove to be
common among the Rosaceae-feeding Catocala when their field biologies
are examined closely (my unpublished rearing data).

THE OVIPOSITION ARENA FAILURES: THE EFFECT OF
FEMALE AGE

Limitations in arena methodology are suggested by the failure of most
captive females to discriminate against the arena walls and, especially,
the Type I control (white oak, Quercus alba ) — on which oviposition by
the same Catocala species does not occur in the field. First, the flight
space offered in these arenas is relatively small compared to the wing-
span of most Catocala. Second, volatile olfactory cues that might be used
by Catocala to identify foodplants will mix in plastic and paper arenas
closed to strong ventilation (the wire mesh arenas were well ventilated),
and when coupled with small arena size, this could easily occlude cues
which guide female choice. The exclusive use of wire mesh arenas in 1981
reduced the problem of oviposition on arena walls, but the females
continued to oviposit heavily on the Type I control, Quercus alba.

Two uncontrolled intervening variables seem relevant to this persis-
tent tendency of females to oviposit on Quercus alba in arenas: female
age, and the “shagginess” of bark samples. Both variables need to be
explored further, but low adult Catocala numbers in New England since
1981 (and the more pressing need to secure adequate rearing stocks each
summer from the few captured females) have limited efforts to design
follow-up arena tests.

For the Carya-feeding species listed in Table 2, there is a positive
correlation between the date a female was confined to her arena and the
fraction of eggs she placed “haphazardly” on arena walls and Type I
controls. The trend is significant for the 1980 arenas (Spearman rank
r=+0.78, n=8, pc. 05), but not so for 1981, although it is positive (r=+0.31,
n=9, p>.25). Table 8 shows 1981 Connecticut ovipositions for palaeogama
(the most abundant species in the best year), on C. ouata versus all other
foodplants, as a function of sampling date. The percentage of ovipositions
on trees other than Carya ouata increases as the season progresses
(Spearman rank r=+0.98, n=6, pc. 05; more conservative pooling strate-
gies for sampling date do not eliminate the significance). Schwarz (1923),
working with caged Catocala car a females, noted a comparable increase
in ovipositions on other than the preferred foodplant (willow) as the
females aged (his data are summarized in Table 9).

In addition, the only Connecticut residua (n=l) and retecta (n=2)
ovipositions on Carya other than C. ouata were similarly late in the 1980
season (27 August, 30 August, and 1 September, respectively). Moreover,
the two “mistakes” noted earlier in the Results were also late in the

29(3):217~233, 1990(91)

231

Table 8. Oviposition by wild Catocaia pa laeogama females in Connecticut during
1981 , by date and foodplant species. Preference for oviposition on Carya
ovata breaks down as the season progresses and females age.

July August

Week 4 Week 1

Number of Ovipositions

August August August
Week 2 Week 3 Week 4

September
Week 1

Carya ovata

5 4

10

4 3

0

another tree

0 0

1

2 2

1

T able 9. Oviposition by caged Catocaia cara females, by date and foodplant species.

Preference for oviposition on willow breaks down as females age. Data

from Schwarz (1923).

Number of Oppositions

August

August

September

September

19

20

15

10

Female 1

willow

10

21

0

0

poplar

0

0

2

1

hickory

0

0

0

2

Female 2

willow

11

30

0

0

poplar

0

0

0

0

hickory

0

0

0

1

season: a palaeogama on Carya cordiformis , 29 August 1981; a subnaia
on Juglans cinerea , 28 August 1980.

These several lines of evidence indicate that oviposition specificity
broadens as females grow older (although adult Catocaia typically live
for several weeks or more, and pupal eclosion in the field is no doubt
spread over at least a 1-2 week period, it is certainly safe to state that the
average age of females in late August and September is greater than in
late July and early to mid August). The broadening in oviposition
specificity is likely in part a simple correlate of deteriorating sensory
capacity. However, it may also in part be an adaptive response to aging,
as somatic reserves are depleted and fecundity and fertility decline
(sharpest discrimination might be favored selectively in younger fe-
males: larvae hatching from these earlier eggs would likely be better
nutritionally provisioned than those from late eggs, and perhaps be at a
competitive advantage, both intra- and inter specifically). Thus, older
females might be viewed as having more freedom to experiment with
secondarily suitable foodplants (e.g., trees other than C. ovata for the
Carya- feeders), with less risk of compromising already successful repro-
ductive effort.

232

J. Res. Lepid.

THE OVIPOSITION ARENA FAILURES: THE EFFECT OF BARK
SHAGGINESS

The roughness of bark samples also seems responsible in part for the
unusual haphazard oviposition by captive Juglandaceae-feedingCafocaZa
in arenas. Bark of the Type I arena controls, white oak ( Quercus alba),
is rough, like that of Cary a ovata and C. glabra, and exfoliates slightly (as
in only C. ovata). Only the bark of C. tomentosa is smooth (in contrast to

C. ovata, C. glabra, and Quercus alba), and even massive C. tomentosa
have roughened bark largely only near their bases.

In all arenas with Carya feeders, fewer eggs were laid on the non-
shaggy C. tomentosa than on the other Carya species (pc. 05 by ANOVA.
successful and unsuccessful arenas combined). Likewise, fewer eggs
were laid on C. tomentosa than on Quercus alba plus arena crevices (t=20,
n=16, pc. 01 by Wilcoxon test; t=3, n=8, pc. 05 for same test using 1981
arenas only, in which the use of wire mesh cages eliminated oviposition
on arenas). Lastly, 3 of the 4 arenas in which more eggs were laid on C.
tomentosa than on the oak/arena were with palaeogama confined in
plastic containers, in which no crevices or seams were present.

Thus, the shagginess of an oviposition substrate seems also to be
important to ovipositing Catocala, along with foodplant species. In this
context, consider too that the genital morphology of female Catocala is
clearly geared toward concealing eggs in inacessible places such as shags,
crevices and cracks: the ovipositor is strongly dorso-laterally flattened,
secondarily sclerotized, and highly extensible; and the eggs of most of the
Juglandaceae feeders, the Nearctic Rosaceae feeders, and some of the
Nearctic oak feeders are also dorso-laterally flattened, in contrast to the
usual hemispherical Catocala egg .

Acknowledgements. I thank Richard Harrison, David Hawks, Charles Remington,
James Rodman, Dale Schweitzer, and Bruce Tiffney for many helpful discussions
about Catocala, and John Hartigan for statistical advice. Dale Schweitzer, David
Furth, and Victor DeMasi helped search for ovipositing females in Connecticut;
Wayne Miller did so in Tennessee. Financial assistance was provided by the George

D. Harris Foundation, the E. Tappan Stannard Fund, a Sigma Xi RESA grant, and
a Yale University Prize Teaching Fellowship. This paper was drawn from a
dissertation submitted to Yale University in partial fulfillment for the degree of
Ph.D.

Literature Cited

Courtney, S.P., and S. Courtney. 1982. The ‘edge-effect’ in butterfly oviposition:
causality in Anthocaris cardamines and related species. -Ecol. Entom. 7: 131-
137.

Gall, L.F. 1987. Leaflet position influences caterpillar feeding and development.
- Oikos 49: 172-176.

. 1991a. Evolutionary ecology of sympatric Catocala moths (Lepidoptera:

Noctuidae). I. Experiments on larval foodplant specificity. - J. Res. Lepid., 29:
173-194.

29(3):217-233, 1990(91)

233

1991b, Evolutionary ecology of sympatric Catocala moths (Lepidoptera:

Noctuidae). II. Sampling for wild larvae on their foodplants. - J. Res. Lepid.,
29: 195-216.

Miller, W. A. 1977. Catocala (Noctuidae) species taken in Clay County, Tennessee.
- J, Lepid. Soc. 31: 197-202.

Nielson, M.C. 1978. Field Summary for 1977, Michigan report. - Lepid. Soc. News
(May/June), p. 6.

Rowley, R.R., and L. Berry. 1912. A dry year’s yield of Catocalae (Lepid.). - Entom.
News 23: 207-214.

Sargent, T.D. 1976. Legion of Night: the Underwing Moths. -Uni v. Mass. Press,
Amherst. 222 pp.

SAS Institute Inc. 1985a. SAS User’s Guide: Basics, Version 5 Edition. - SAS
Institute Inc., Cary, North Carolina. 1290 pp.

. 1985b. SAS User’s Guide: Statistics, Version 5 Edition. - SAS Institute Inc.,

Cary, North Carolina. 956 pp.

Schwarz, E. 1923. The reason why Catocala eggs are occasionally deposited on
plants upon which the larva cannot survive; and a new variation (Lepid.,
Noctuidae). - Entom. News 34: 272-273.

Schweitzer, D.F. 1982. Field observations of foodplant overlap among sympatric
Catocala feeding on Juglandaceae. - J. Lepid. Soc. 36: 256-263.

. 1987. Catocala pretiosa , the precious underwing moth: results of a global

status survey, with a recommendation for retention in category 2. - Status
Survey Report to the US Fish & Wildlife Service, (Newton Corner, MA). 24 pp.
Sokal, R.R., and F.J. Rohlf. 1982. Biometry (2nd Ed.). - Freeman, San Francisco.
859 pp.

Stanton, M.L. 1982. Searching in a patchy environment: foodplant selection by
Colias p. eriphyle butterflies. - Ecology 63: 839-853.

Journal of Research on the Lepidoptera

29(3):234-236, 1990(91)

Book Reviews

A REVISION OF THE INDO -AUSTRALIAN GENUS ATTACUS. Richard S.
Peigler. 1989. 167 pp., 9 maps, 24 figures (line drawings), 36 color figures.
Published by the Lepidoptera Research Foundation, 9620 Heather Road, Beverly
Hills, California 90210, USA. ISBN 961 1464-2-7

Format: 20.2 by 27.8 cm. Price US $30.00 postpaid. Text in English.
Summaries in German, Indonesian and Japanese. Thesis for the PhD degree in
Entomology, Texas A&M University, College Station, Texas, December 1983.

The genus Attacus is particularly known for comprising the largest Lepidopt-
era in the world. Its distribution, categorized as Indo- Australian with a slight
incursion into the Palearctic zone, extends from the southern Himalayas to Sri
Lanka at Taiwan to the Philippines and to northern Australia, with a gap in New
Guinea where the genus is replaced by its sister-group Coscinocera. Of the 14
species cited by Peigler, 11 are insular and endemic to their respective islands or
archipelagoes, and only Attacus atlas has a range that is both insular and
continental, including all of southeastern Asia, Taiwan, Sumatra, Java, and
Borneo. The absence of endemics on these four islands is noteworthy. The
economic role of Attacus is quite minimal, the culture of their fagara silk having
been practically abandoned. The larvae, of which the known ones are very
polyphagous, have only been well known as pests of cacao and avocado.

In spite of their gigantism, except for three species which are largely commer-
cialized on Taiwan (A. atlas) and the Philippines (A. lorquinii and A. caesar),
these insects have remained relatively little known; in fact, several taxa are still
represented in collections by one to ten specimens. The existing literature, itself
very unequally extensive for the various species, gave numerous errors and gaps,
and a revision of the genus appeared necessary.

The plan to the book is that which is employed in every work of this kind. After
the preface, the introduction, and an interesting historical review of the litera-
ture (the first Attacus were cited and figured around the beginning of the 18th
century), the author offers his methods of work which made particular use of
considerable material that was examined, and notably, the collections containing
the majority of the type material, in the principal American and European
museums ( 17 of which were visited). The pertinent data from all these specimens
plus the correspondence with other scientific institutions and private collectors
is faithfully reproduced in the treatment of each species concerned, constituting,
notably for the biogeographer, documentation of the first order. The following
two chapters treat the study of the morphology of the imago and early stages, the
latter of which only four species are known, an examination of the systematic
position of the genus and a review of nomenclatural problems.

The genus Attacus, in its current concept, considered without iheArchaeoattacus
and Coscinocera which are nonetheless figured by the author, is a group that is
monophyletic and absolutely homogeneous. The study of the 14 species cited by
Peigler, following an alphabetical order which I would have preferred to have
been phylogenetic, is preceded by a key that is based entirely on external

29(3):234-236, 1990(91)

235

characters of the habitus. The treatment of each species consists of the following
sections: bibliographic references, type specimens, diagnosis, early stages if
these are known, geographical distribution, material examined, discussion
(designation of lectotypes, synonymies, specific characters, etc.). Of the 50 taxa
of various ranks proposed in the literature, 14 are judged to be valid; I note 29 new
synonymies of which 23 refer to Attacus atlas alone. Two new species have been
described in a preliminary note (1975, Nachr. Ent. Ver. Apollo , Frankfurt (N.F.)
6: 53-60) and four taxa are elevated to specific rank.

The 14 taxa treated are all considered to be full species, Peigler considering
that the concept of subspecies is not applicable here. In spite of his earnest
attempt to demonstrate it, this appears to be difficult to prove. As a matter of
generality, declaration of specific or subspecific rank of certain taxa, in the
absence of definitive characters, becomes somewhat subjective. Thus doubts
remain in assigning specific rank to taxa such as A. mcrnulleni from the Andaman
Islands and particularly to A taprobanis from Sri Lanka. The position that the
reviser has cornered himself in is hardly comfortable!

The work is finally composed of four chapters in which the following topics are
covered: biology and ecology (see also Appendix I, hostplants), intergeneric
relationships based on a cladistic study founded on 20 characters, a phylogenetic
analysis of the genus and speciation. The solutions proposed in phylogeny are
pertinent and generally convincing, but unfortunately the final result is to group
the American genera on the one hand and the Indo-Australian genera on the
other, with the African Epiphora by itself in the tribe. Finally, the bibliography
is plethoric with more than 300 cited works.

The only criticism to be expressed concerns the illustrations. The drawings are
mediocre, particularly those figuring the male genitalia and even more so the
female, which appear schematic and often incomplete. The color plates are
technically excellent but the scale adopted is regrettable (not indicating the
sizes), reducing the moths to two-thirds their actual size, thus depriving them of
their most attractive character, their great sizes. Nevertheless every species is
absolutely identifiable and, where doubt exists, it is sufficient to refer to the
biogeographical data where necessary. Finally, I have some regret that the
author did not go ahead and revise all of the Palearctic and Indo-Australian
genera of Attacini, three more genera containing perhaps 10 additional species,
as this would have doubled the value of this basic work.

As it is, this work is a wealth of information of which I cannot give a fair
appraisal in the space allotted in this review. The work of Richard Peigler is a
contribution far superior to that of a simple revision and every entomologist, even
if not particularly interested in Attacus , will find a documentation and assem-
blage of valuable data beyond the limits of the subject treated, for all of the fauna
of the geographical zone concerned. It is a book particularly well done and of
beautiful quality, including the printing and the presentation, and appears to be
available at a very reasonable price.

Claude Lemaire, La Croix des Baux, F -84220 Gordes , France.

230

J. Res . Lepid .

THE BUTTERFLIES OF THE CAUCASUS. 1. Y. P. Nekrutenko. 1990.
Naukova Dumka. Kiev (SU). In Russian. 215 pp., 32 col. pis., 106 figs. ISBN 5 -
12- 001352-X. Price 2.10 Rbl. (hardback, plastic, approx. 17 x 24 cm).

This important work is the result of many years of field work and study of
literature and original material in museums in the USSR and abroad by the
author, a leading Ukrainian lepidopterist. The book contains comprehensive
treatment of Papilionidae, Pieridae, Satyridae and Danaidae of the Caucasus
(remaining Rhopalocera families are to form the second volume of this two
volume handbook; I understand that the manuscript is near completion). The
work opens with general chapters on the geography of the Caucasus and on the
history of the study of butterflies of the Caucasus. The latter is a detailed 25 pp.
account illustrated by photographs of leading lepidopterists who published
important contributions on the butterflies of the Caucasus. The systematic part
contains a short introduction, outline of anatomy and keys to families, as well as
monographs on all species. There are also keys to genera and species. The
individual species monographs provide comprehensive and painstakingly com-
piled bibliographies, descriptions of external features and anatomy including
drawings of male and female genitalia, variation, type-locality and etymology of
the valid name. Comprehensive information on distribution, biology, larval food
plants, and often on the conservation or status are also provided.

Throughout the text references are made to information sources; the bibliog-
raphy at the end of the volume comprises well over 400 references commencing
with the “Systema Naturae” of 1758. Particularly valuable are consistent
references to original descriptions. The color plates feature all species treated in
this volume, near to life sizes. The quality of printing is probably very good by
Russian standards, but not surprisingly leaves much to be desired by ours. As
this book is above all a taxonomic work, the author calls it a key which seems to
me an understatement, a discussion of the classification of some species should
await the publication of the second volume. Suffice to say that the author
deserves both our congratulations and our thanks. It is a great pity that the book
is written in Russian only, without an English summary, and thus is inaccessible
to the vast majority of European lepidopterists. The author should certainly
consider publishing, if not an English translation, at least an extensive English
summary of and a guide to the Russian version. I impatiently look forward to the
publication of the second volume of this important handbook.

Otakar Kudrna, Karl- Straub -Str. 21, D 8740 Bad Neustadt - Salz
( Germany ).

INSTRUCTIONS TO AUTHORS

Manuscript format: Two copies must be submitted, double-spaced, typed, with wide
margins. Number all pages consecutively. If possible italicize rather than underline scientific
names and emphasized words. Footnotes are discouraged. Do not hyphenate words at the right
margin. All measurements must be metric. Time must be cited on a 24-hour basis, standard
time. Abbreviations must follow common usage. Dates should be cited as: day- Arabic numeral;
month-Roman numeral; year- Arabic numeral (ex. 8. IV. 1984). Numerals must be used for ten
and greater e.g. nine butterflies, 12 moths.

Electronic submission: The Journal is now being produced via desktop publishing, allowing
much shorter publication times. Although typewritten manuscripts are acceptable, those
submitted on computer disc are highly preferred. After being notified of your paper's
acceptance, submit either a Macintosh or IBM disc version. Include on your disc both the fully
formatted copy from your word processing program and a text-only (ASCII) copy. The two most
preferred formats are Microsoft Word for the Macintosh and either Microsoft Word or Word
Perfect for the IBM, although translation utilities will allow conversion from most formats. Put
returns only at the ends of paragraphs, not at the end of each line. Use one tab to indent each
paragraph. Even if your printer is incapable of outputting italics, please specify italics rather
than underline in your disc copy. Please note any special characters that are used in either the
body of the text or the tables {e.g. e, u, '% §, p, 6, 9 ).

Title material: All papers must include the title, author's name, author's address, and any
titular reference and institutional approval reference. A family citation must be given in
parenthesis (Lepidoptera: Hesperiidae) for referencing.

Abstracts and Short Papers: All papers exceeding three typed pages must be accompanied
by an abstract of no more than 300 words. Neither an additional summary nor key words are
required.

Name citations and Systematic Works: The first mention of any organism should include
the full scientific name with unabbreviated author and year of description. There must be
conformity to the current International Code of Zoological Nomenclature. We strongly urge
depositing of types in major museums, all type depositories must be cited.

References: All citations in the text must be alphabetically listed under Literature Cited in
the format given in recent issues. Abbreviations must conform to the World List of Scientific
Periodicals. Do not underline or italicize periodicals. If four or less references are cited, please
cite in body of text not in Literature Cited. For multiple citations by the same author(s), use
six hyphens rather than repeating the author’s name.

Tables: When formulating tables, keep in mind that the final table will fill a maximum space
of 11.5 by 19 cm either horizontally or vertically oriented. Number tables with Arabic
numerals. When submitting tables on disc, use tabs between columns rather than multiple
spaces.

Illustrations: Color can be submitted as either a transparency or print, the quality of which
is critical. Black and white photographs should be submitted on glossy paper, and, as with line
drawings, must be mounted on stiff white cardboard. Authors must plan on illustrations for
reduction to page size. Allowance should be made for legends beneath, unless many consecutive
pages are used. Drawings should be in India ink. Include a metric scale. Each figure should
be cited and explained as such. Each illustration must be identified by author and title on the
back. Indicate whether you want the illustration returned at your expense. Retain original
illustrations until paper is accepted. Legends should be separately typed on pages entitled
"Explanation of Figures." Number legends consecutively with separate paragraph for each
page of illustration.

Review: All papers will be read by the editor(s) & submitted for formal review to two referees.

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

Volume 29 Number 3 Fall 1990(1991)

IN THIS ISSUE

Date of Publication: December 31, 1991

Evolutionary Ecology of Sympatric Catocala Moths 173

(Lepidoptera: Noctuidae)

I. Experiments on Larval Foodplant Specificity

Lawrence F. Gall

Evolutionary Ecology of Sympatric Catocala Moths 195

(Lepidoptera: Noctuidae)

II. Sampling for Wild Larvae on their Foodplants

Lawrence F. Gall

Evolutionary Ecology of Sympatric Catocala Moths 217

(Lepidoptera: Noctuidae)

III. Experiments on Female Oviposition Preference

Lawrence F. Gall

Book Reviews 234

>-v

Cover Illustration:

Photograph by Lawrence F. Gall.

SKT THE JOURNAL OF

ON THE LEPIDOPTERA

Volume 29

Conservation Issue

Number 4 Winter 1990(1992)

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

ISSN 0022 4324
Published By:

Founder:
Editorial Staff:

Associate Editors:

The Lepidoptera Research Foundation, Inc.
9620 Heather Road
Beverly Hills, California 90210
(310) 274-1052

William Hovanitz

Rudolf H. T. Mattoni, Editor
Scott E. Miller, Assistant Editor

Emilio Balletto, Italy
Henri Descimon, France
Philip DeVries, U.S.A.

Thomas Emmel, U.S.A.

Lawrence Gall, U.S.A.

Hansjuerg Geiger, Switzerland
Otakar Kudrna, Germany
Arthur Shapiro, U.S.A.

Atuhiro Sibatani, Japan
Karel Spitzer, Czechoslovakia

Manuscripts and Notices Material may be sent to the Editor at:
9620 Heather Road, Beverly Hills, CA 90210 (213) 274-1052

The JOURNAL is sent to all members of the FOUNDATION.

Classes of Membership:

Regular (Individual)
Contributing

Student/Retired-Worldwide
Subscription Rate/Institutions
Life

$ 20.00 year (vol.)

$ 30.00 or more, year (vol.)
$ 18.00 year (vol.)

$ 28.00 year (vol.)

$ 250.00

STATEMENT OF OWNERSHIP AND MANAGEMENT

THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA is published four times a year by THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The office of publication and the general business
office are located at 9620 Heather Road, Beverly Hills, California 90210. The publisher is THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The Editor is R. H. T. Mattoni at the above address.
The Secretary-Treasurer is Leona Mattoni at the general business office. All matters pertaining to
membership, dues, and subscriptions should be addressed to her, including inquiry concerning mailing,
missing issues, and change of address. The owner is THE LEPIDOPTERA RESEARCH FOUNDA-
TION, INC., a non-profit organization incorporated under the laws of the State of California in 1965. The
President is R. H. T. Mattoni, the Vice President is John Emmel, the Secretary-Treasurer is Leona
Mattoni. The Board of Directors is comprised of Barbara Jean Hovanitz, Leona Mattoni, and R. H. T.
Mattoni. There are no bond holders, mortgages, or other security holders.

Journal of Research on the Lepidoptera

29(4):237-253, 1990(92)

Conservation of butterflies in Australia

T.R. New

Department of Zoology, La Trobe University, Bundoora, Victoria, Australia 3083.

Abstract. The development of understanding of butterfly conserva-
tion in Australia is reviewed. A summary is given of the dramatic
changes to Australian terrestrial environments which have occurred
since Caucasian settlement, and some factors leading to butterfly
decline in recent years are itemised. The study of Australian butterflies
is summarised, and the limitations of current knowledge in relation to
conservation concern are stressed. Recent legislative measures involv-
ing butterflies are itemised and discussed. Species-orientated conser-
vation cases are still rare, but the roles of several taxa in increasing
political and public conservation awareness are outlined.

Introduction

“Butterfly Conservation” is a relatively new topic in Australia, and has
not yet developed to encompass the broad public and scientific concern it
engenders in much of the northern hemisphere. However, during the last
few years it, together with other aspects of invertebrate conservation,
has started to appear on scientific and political agendas as part of a
growing more general concern over the future of the Australian environ-
ment and biota. A preliminary report on conservation status of Austra-
lian insects by Key (1978) aroused considerable interest. A broader
appraisal (New 1984) and a number of other accounts and surveys have
been published during the last decade, and many of these are noted
below.

Thus, in contrast to the more detailed historical and factual treatments
which are feasible for butterfly conservation in Europe or North America,
where this theme has long been accepted readily, this account traces the
emerging awareness of the topic and how it is gradually becoming
acknowledged as important in Australia. Specific case histories are
sparse, and there is thus not a “bank” of experience of conservation of
particular taxa equivalent to that available for parts of the northern
hemisphere. This account must, in contrast, indicate some of the
principles, restrictions, and increasing public and legislative sympathy
for butterfly conservation in the country.

The Problem

The Australian environment has been changed dramatically during
only 200 years of European settlement, and much of the indigenous biota
of the island continent has suffered accordingly. The major change has
undoubtedly been the destruction of natural vegetation, and related
effects. Vast tracts of land have been cleared for pasture and arable
agriculture, so that most categories of forest and woodland have declined.

238

J. Res. Lepid.

Figure 1 . Australia, indicating major political boundaries, biogeographical regions,
approximate numbers of butterfly species from each State and Territory,
and some major centres of butterfly diversity. States and Territories
denoted by initial letters (Western Australia, Queensland, New South
Wales, Victoria, South Australia, Tasmania, Northern Territory, Australian
Capital Territory); numbers of butterfly species summarised from Common
& Waterhouse (1981) and more recent literature; centres of diversity shown
are (1) Cape York Peninsula, (2) southern coastal Queensland, (3) Cairns
area - after Kitching (1981), see text; the regions shown are the southern
Bassian, with distinct Western and Eastern Provinces, the Torresian
(northern and north eastern), and the more arid Eyrean, delimited by the
500mm isohyet. The northern boundary of the East Bassian is debatable,
and there are frequent northward incursions of southern fauna at higher
altitudes.

Only small proportions of the original vegetation remain in some major
grain-growing areas (such as parts of Western Australia and South
Australia), and nearly three-quarters of forests in Victoria have been
modified substantially, for examples. As much as half the continent may
have suffered degradation in land quality (Ive and Cocks 1989) by loss of
topsoil resulting from combinations of vegetation clearing, over-cultiva-
tion, overstocking, irrigation and the activities of feral exotic animals -

29(4):237-253, 1990(92)

239

ranging from rabbits and goats to horses, camels and water-buffalo. A
second category of change considered also to have affected insects
adversely has been the introduction of large numbers of exotic animals,
including the above and other mammals, birds, freshwater fish and the
cane toad (Bufo marinus) as well as an array of arthropods, including
non-specific biological control agents, and aggressive plant weeds.

Urbanisation in Australia, with its attendant effects, has tended to be
concentrated along the eastern side and the southern corners of Austra-
lia, the areas of greatest rainfall and ecological complexity (Fig. 1). Much
of the vast semiarid to arid inland region, in contrast, is very sparsely
settled, although still subject to grazing by stock, and related degrada-
tion pressures.

It is reasonable to infer that the above, and other, changes have caused
decline and/or loss of many insect taxa, but this is difficult to substantiate
because of lack of historical information on most groups. For many insect
orders, there are still large numbers of unnamed species, and many taxa
have yet to be collected and categorised; this lack of knowledge, with
perhaps only half our insect species, or even less, yet having names,
constitutes a substantial “taxonomic impediment” (Taylor 1976) to
communicating concern about loss of insect diversity. As they are
elsewhere, butterflies are one of the best-known groups of insects, and
attention is now being paid to both the wellbeing of individual taxa, and
to their use as indicators of a greater conservation need: as reflections
of invertebrate biodiversity and thus of the “health” of natural environ-
ments.

However, Australia, even neglecting various political outliers, is a vast
area, and there are many gaps in distributional and faunistic knowledge
even for such “well-known” groups. The continent spans a wide range of
environments, from lowland tropical rainforest in the north, through
arid regions and cool woodlands, to cold montane environments in the
southeast, and this mosaic of environments can be used to provide foci for
insect conservation in various ways. A general account of some of the
problems involved in planning conservation of the fauna is given by
Greenslade & New (1991).

Australian butterflies, and their study

In contrast to many groups of the 20,000 or so species of moths in
Australia the butterflies are, in general, taxonomically tractable and
most of the nearly 400 species are recognisable without difficulty. A
recent handbook (Common & Waterhouse 1972, 1981) marks the zenith
of a series of texts produced at intervals over the previous century. The
most important of these are a Catalogue by Masters (1873, noted by
Moulds, 1977, as the first book wholly on Australian butterflies), Olliff
(1889), Anderson and Spry (1893-94), Rainbow (1907) and two later
classics still of great value: Waterhouse & Lyell (1914) and Waterhouse
(1932). With the stimulation to collectors and biologists provided by

240

J. Res. Lepid.

these and more recent texts (McCubbin 1971, D’Abrera 1971 and later
editions) and a number of less comprehensive volumes, a useful frame-
work of distributional, taxonomic and biological information on Austra-
lian butterflies now exists, leading to possibility of some generalisation
over faunal composition and larval food-plant associations (Symons
1980), for examples. More research than ever before is currently being
pursued on butterfly biology in Australia and representatives of many
endemic genera have been studied in considerable detail during the last
decade or so.

It is likely that a few further species of Hesperiidae and Lycaenidae will
be found in less extensively collected regions, and that the taxonomic
status of members of some complexes of putative subspecies will also
change. Otherwise, very few species are undescribed. As with many
other insect groups in Australia, endemism is high, so that there are
some very characteristic elements in the fauna. The butterflies of
northern Australia have much in common with those of New Guinea and
many non-endemic taxa are restricted, or largely restricted, to this
warmer region. They represent an attenuation from the richer faunas of
New Guinea and the Indo-Malayan fauna. A few species are widely
distributed elsewhere: Danaus plexippus (L.) (which has been in Austra-
lia for about 100 years), D. chrysippus petilea (Stoll), Lampides boeticus
(L.) and Pieris rapae (L.) are the best-known. The most characteristic
endemic complexes occur in the Hesperiidae: Trapezitinae, Nymphalidae:
Satyrinae, and Lycaenidae: Theclinae and Polyommatinae, and these
groups contain a very high proportion of the butterflies of conservation
concern. Others occur. Several Papilionidae, for example, have been
considered in conservation legislation (see below). In general, there is a
tendency for endemism at both generic and specific levels to be highest
in the southern regions of Australia, probably reflecting a longer history
of evolutionary isolation as relatively earlier arrivals than most butter-
flies now confined to the North. Torres Strait, between New Guinea and
Queensland, may have existed for only some 5.5-6 thousand years, before
which these areas were contiguous. Many of the least tractable groups
of butterfly species or putative subspecies occur in the south, indicating
that speciation is continuing. Many of them have very localised distribu-
tions and some are known to have declined in distribution range. Thus,
for example, the island State of Tasmania has only three endemic species
and one endemic genus (all Satyrinae) but many endemic subspecies (
Couchman & Couchman 1977). Some other Tasmanian subspecies are
shared only with southern Victoria. Figure 1 shows the main political
areas of continental Australia with indications of the relative diversity
of butterflies in each, together with major biogeographic zones.

Butterfly diversity is generally highest in the east, and decreases
rapidly to the west of the Great Dividing Range. Major coastal centres
of diversity (Fig. 1) include (1) the northernmost part of Queensland,
Cape York Peninsula, with the most diverse array of Australian butter-

29(4):237-253, 1990(92)

241

flies, around 140 species (Monteith & Hancock 1977), (2) an area in
southern Queensland where there is strong overlap of northern and
southern faunas (Kitehing 1981), and (3) the area around Cairns,
between these two, which Kitehing (1981) emphasises has been subject
to especially high collecting intensity over many years. In contrast,
butterfly diversity is generally low in the arid interior of the continent,
and few endemic taxa occur there. Many of the Northern Territory taxa,
for example, are restricted to the more clement climate of the “Top End.”
Only some 40 species occur towards the south of the Territory in the
centre of the continent. Because of their degree of taxonomic isolation,
many Australian butterflies are readily designatable as worthy species
for conservation attention when their restricted habitats are perceived to
be threatened, or are known to be declining.

The gross distributional ranges of many species are known, and
distribution maps for each species and subspecies are included in
Common & Waterhouse (1981), but the “fine grain” distribution of most
is by no means clear and collectors continue to record taxa from up to
hundreds of km outside their earlier-recorded ranges.

Many species are known from only very few localities, some from only
single small colonies occupying very restricted areas. In some instances
there is little apparent reason for this as the habitats appear to be
enclaves within much larger areas of seemingly similar terrain and
vegetation. In others, restricted distributions are related more clearly to
foodplant availability or to altitude or other physical features. A number
of southeastern Satyrinae, for example, are alpine or subalpine and
restricted to rather precise altitudinal ranges in the small mountainous
regions. They often occur as discrete colonies on different mountains
widely separated from each other by lower land, and some such isolated
populations have been accorded subspecific names.

As in other parts of the world, the species of direct conservation concern
tend to be those whose abundance or distribution has either been reduced
by human activities or where this is considered likely to occur. They
encompass widespread species which are decreasing in abundance and
species which are known from few sites, so that “more general” and “more
specific” threats are both relevant. The latter could result in rapid
extinctions but there are few documented cases of this in Australia.

There has never been a large population of butterfly hobbyists in
Australia, and yet much of our knowledge and the bulk of collections in
major institutions derives from their activities. Together these collec-
tions comprise a very substantial data base which is being progressively
recorded and assessed with the aim of refining knowledge of butterfly
distributions and how these may have changed in relation to land use.
This is being augmented by increased and more selective survey and/or
collecting in little-studied areas of several States to increase knowledge
of some critical species. It will clearly be a long time before detailed
distributional knowledge is relatively complete and lack of this definitive

242

J. Res. Lepid.

information (in comparison with, particularly, some parts of Europe) is
a communication hindrance with decision-makers in promoting cases for
the conservation need of particular taxa. Formalised distribution data
are not yet sufficiently comprehensive, in general, to be persuasive to
people seeking information on the detailed status of particular butter-
flies or to reinforce claims that particular butterflies “only occur” in a
stated place and this, coupled with taxonomic uncertainty over the
precise status of particular populations, weakens the case for some
single-species conservation efforts. Most collectors reside on the eastern
coastal part of Australia or in the southwest corner, the areas of greater
butterfly diversity and greatest concerns over environmental effects on
the insects. Some other parts of the continent are still relatively
inaccessible and the total area needing detailed survey is, indeed,
enormous.

Perhaps the most intensively - collected area of mainland Australia is
Victoria, currently subject to a butterfly mapping scheme (ENTRECS)
operated by the Entomological Society of Victoria. On a relatively crude
10 x 10 minute grid system (defining areas of about 15 x 18 km), only
about half these areas in the State (one of the smallest) have any
butterflies formally recorded from them, and very few are comprehen-
sively known (ESV 1986). Victoria contains only about 120 species of
butterfly, about 10 being recorded there for the first time during the last
decade. Maps for rainforest butterflies in New South Wales (Nadolny
1987) emphasise known distributions, again often sparse, of species of
conservation interest, and more extensive survey work is currently in
progress for that State. A list for the Australian Capital Territory
(Kitching et al. 1978) also cites localities for that restricted area. A wider
scale distribution - recording scheme for butterflies in Australia is also
current (Dunn & Dunn 1990), but the prospect of providing detailed
distribution maps for most butterflies over the vast areas of the larger
Australian States is not realistic at present. There is need for decision-
makers to accept the impracticality of doing this, or of being able to
emulate various European mapping-schemes with very limited logistics
and interest, and to accept the opinions of knowledgable lepidopterists on
the status of most species rather than insist on quantitative information.

Protective Legislation

The Legislative system in Australia is two-tiered, involving Federal
and State Governments. In general, the former lacks power to over-ride
a decision made by a State Government, and States with a common
border may have very different environmental protection laws. Insects
have received little specific attention. Butterflies have generally been
included, with other invertebrates, under such general terms as “wild-
life” or “fauna” and it is only recently that such designations have
diverged in functional interpretation from “vertebrates.” In all States
and Territories, insects are protected in National Parks, and sometimes

29(4):237-253, 1990(92)

243

other categories of reserves, and permits are needed to collect or study
butterflies in such areas.

Australia is a signatory to CITES, although no Australian butterflies
are included in Appendix 1. Export permits are needed for any Austra-
lian native fauna under the Wildlife Protection (Regulation of Exports
and Imports) Act 1982. The operation of this Act is discussed by Monteith
(1987), who emphasises that there has never been any significant
amount of exploitative butterfly collecting in Australia for export. This
is thus of little conservation concern at present but, clearly, could
increase. There is little doubt that some species could in principle be
endangered by unscrupulous collecting, and the Entomological Society of
Victoria (the largest group of amateur entomologists and hobbyists in the
country) has listed several species under a Voluntary Restricted Collect-
ing Code, whereby no more than two adult specimens should be taken by
any one collector in a season; this is, in general, heeded responsibly.
Some desirable skippers could be overcollected, as their larvae produce
conspicuous leaf-shelters which can be systematically taken. Some ant-
attended Lycaenidae, including rare species of Ogyris Westwood, are
traditionally collected in the pupal stage which occurs under the loose
bark of eucalypts. Many collectors of such groups are tempted to take a
surfeit of specimens to compensate for anticipated loss of specimens from
parasitisation, and there are unconfirmed suggestions that this form of
collecting has led to some rare species becoming endangered, or even
locally extinct.

Until recently, the only butterflies formally “protected” (that is, which
collectors are prohibited from taking!) in any part of Australia were
Papilio ulysses joesa Butler and species of Ornithoptera in Queensland,
designated under the Queensland Fauna Protection Act in 1974. The
Queensland Act, incidentally, defines “fauna” as only indigenous birds
and mammals plus other species specifically declared to be “fauna” by
government decree - so that Ulysses and the birdwings had to be specifically
declared to have this status! Permits are needed for capture of any
individuals of these taxa in the State, and a compulsory royalty of A$20/
specimen was introduced. Several collectors have indeed been pros-
ecuted for transgressing this legislation. Proposals were made in 1980
to have a number of insects, including the butterflies Ornithoptera
richmondia (Gray), Euschemon rafflesia (Macleay), Argyreus hyperbius
inconstans (Butler) and Tisiphone abeona Joanna (Butler), listed for
legislative protection in New South Wales under the State’s National
Parks and Wildlife Act. After considerable debate, and objections from
sections of the entomological community (who were not consulted over
the earlier Queensland legislation but opposed it strongly in retrospect),
the New South Wales proposals were withdrawn.

The arguments mounted against this proposed legislation included (1)
that the species appeared not to have been selected as priority taxa by
any scientific process, but as examples of a much larger range of notable

244

J. Res. Lepid.

insects in New South Wales many of which were at least equally
deserving of protection; (2) that the conservation status of the nominated
species was not clear but some, at least, appeared not to need such
legislation which could merely serve to increase their value in the
perception of collectors; (3) that such legislation without any concomi-
tant provision for habitat protection/preservation or ecological study of
the species involved was little better than a token gesture, and (4) that
it could be extremely difficult to enforce. More particularly for the non-
butterflies proposed, several species were sufficiently similar to other,
non-protected, species that they could be differentiated confidently only
by detailed examination, and it would not be reasonable to expect field
wardens to be able to do this. In general, habitat protection was seen as
far more urgent, significant and worthwhile than legislation of this sort.
Many people have emphasised that the Queensland habitats for P.
ulysses and Ornithoptera priamus (L.) are well-represented and that the
species are not uncommon. Indeed, P. ulysses has recently extended its
range into various towns and suburbs where its food plants have been
planted.

No other State has had butterfly listings equivalent to the above,
although some other insects are legally protected in Western Australia
and Tasmania. Recent pioneering legislation in Victoria has specifically
provided for invertebrates (and non-vascular plants) under the Flora and
Fauna Guarantee Act, 1988, which in emphasis is more akin to the U.S.
Endangered Species Act. Nominations can be made of any taxon deemed
worthy of conservation or believed to be in need of it, and an interim
conservation order may then be issued to protect the taxon and its habitat
(for example, from any urgent threat of development or despoliation)
pending investigation of its conservation status. The nomination process
is simple and does not require massive documentation. The onus for
subsequent investigation of conservation need falls on the State Minis-
try, together with the duty to formulate a management plan if a
conservation need is indeed demonstrated. Several Lycaenidae per-
ceived as endangered from particular development projects, such as
mining exploration and urbanisation, have been nominated for listing
(together with other insects) and it remains to be determined whether or
not this legislation will be logistically swamped in its undertaking to
ensure that “no native animal or plant will be allowed to become extinct
in Victoria.” It is as yet too early to comment on the functioning of this
Act, which is being watched with keen interest by many conservationists
in Australia. Such increased legislative awareness for insects (and other
invertebrates) has the potential to draw considerable attention to the
content of the modified form of the European Charter on Invertebrates
which was adopted in Australia in July 1989. The Australian Entomo-
logical Society has long had an active Conservation Committee which
maintains a watching brief on any such developments and seeks to foster
awareness of the importance of invertebrates in conservation activity.

29(4):237-253, 1990(92)

245

A recent legislative step in Australia has been to designate a species of
butterfly in Queensland as “Permanently Protected Fauna” in that
State, a status which confers a very high degree of protection and which
has typically been accorded only to “high profile” vertebrates such as
koala and platypus. Specimens held in the State by collectors, regardless
of when captured, may be forfeit.

Acrodipsas illidgei Waterhouse & Lyell (see below) was gazetted on 21
July 1990 as the first insect to be declared Permanently Protected Fauna
(«PPF”) Owners of specimens held in private collections may be given
permission to retain these under a Permit to keep Fauna, renewable
annually at a cost of $20, but are the permanent property of the Crown.
The permit specifies the premises at which the specimen(s) is/are to be
held and separate Movement Permits ($10 each) are needed to allow
movement of specimens to other premises or in or out of Queensland.
Maximum penalties specified for illegally collecting PPF are “200 pen-
alty units or 2 years imprisonment, or both” and for “illegal possession”
are “100 penalty units or 12 months imprisonment or both.” A. illidgei
had been cited, as an example only , of a notable insect in a general
document on the State of the Environment in Queensland (1990) and
seemingly should have no greater conservation priority than very many
other Queensland insects: it is not clear why it was selected for such
special treatment.

Priority species for Conservation

No regional Red Data Book or similar full register of species of
conservation interest has been produced for Australian insects and the
prospect of doing this is remote. Nine Australian insects are included in
the IUCN Invertebrate Red Data Book (Wells et aL 1983): no butterflies
are noted. However, two documents published in 1988 do list Australian
butterflies with implications for priority in conservation assessment.
The more widely-known (IUCN 1988) includes seven taxa (Appendix 1),
six of them from Tasmania, so that there is a strong bias towards
evaluation of that island State to the relative neglect of the mainland. All
six Tasmanian forms were categorised as “indeterminate,” and the
unusual mainland skipper Euschemon rafflesia as “insufficiently known.”

A broad survey of insect taxa meriting conservation attention in
Australia by Hill & Michaelis (1988) involved requesting relevant opin-
ion and information from more than 600 entomologists in the country.
The 54 responses collectively listed 61 butterfly taxa and it is notable (in
view of the Queensland legislation noted earlier) that no Papilionidae
were nominated.

The six Tasmanian taxa of the IUCN list appear also on the Hill &
Michaelis list, although E. rafflesia does not. The latter list includes 28
other Hesperiidae, 24 Lycaenidae, 7 Nymphalidae and 2 Pieridae, some
from only parts of their range. A select list of threatened taxa included
23 of these butterflies, and enumerated the perceived threats to these.

246

J. Res. Lepid.

They include fire and agricultural activity (including trampling by
stock), roadworks, urbanisation, mosquito control by insecticides, sand
mining and bauxite mining. The most frequently cited threats are some
form of “clearing” (10 species), followed by fire (7), the latter reflecting
concern over vulnerability of remnant habitats even for a biota generally
believed to be well-adapted to withstand low intensity fires as a sporadic
natural occurrence.

The citation of “roadworks” for four species emphasises the extreme
vulnerability of several small and highly localised populations.
Hypochrysops piceatus , Kerr, Macqueen & Sands (Lycaenidae), for
example, is known to be extant in only one roadside habitat (extending
for about 200 m) in Queensland and could seemingly be eradicated
merely by injudicious road-widening or careless use of a bull-dozer in a
rather remote area. It has apparently already become rarer because of
tree removal. Urbanisation effects, including direct habitat destruction,
are considered likely to be important for some remnant populations of
formerly more widespread taxa. The reference to mosquito-spraying is
to A. illidgei, which is restricted to a few coastal mangrove areas in
Queensland. However, many other butterflies may also have declined,
and their increasing rarity and vulnerability increases their priority for
conservation in many forms of such comparative assessment. Although
the taxa listed by IUCN (1988) have not had their precise status formally
determined, suggestions from knowledgeable collectors imply that sev-
eral of these, and others, should be regarded as being in the priority
categories “Endangered” or “Vulnerable.”

Species-orientated Conservation

A few such localised taxa have recently been the targets of efforts to
determine their status more precisely, or to more clearly appraise their
need for conservation. Species-orientated butterfly conservation
programmes are still a novelty in Australia, and the few cases have
involved localised taxa coupled with an imminent perceived threat, that
is “crisis-management” conservation with its attendant emotion. Little
was known of the detailed biology of any of these species until their
profiles were raised in this way. One case, in particular, has been very
important in helping to establish the acceptance of butterfly conserva-
tion in Australia.

This is the Eltham Copper, Paralucia pyrodiscus lucida Crosby
(Lycaenidae). In 1951, a brightly coloured form of the Dull Copper, P.
pyrodiscus (Rosenstock), from outer northeastern Melbourne (Victoria)
was described as distinct (Crosby 1951). It had been captured in
reasonable numbers over a small area since about 1938, but declined
substantially during the 1950s as its habitats became overtaken by the
urban sprawl, and was presumed to have become extinct near Melbourne
by about 1960. In early 1987, a large and apparently thriving colony was
discovered at Eltham, within the historical range of the butterfly (Braby

29(4):237-253, 1990(92)

247

1987, 1990), and on land imminently threatened with development as a
housing estate. An approach to the State Minister for Conservation,
Forests and Lands led to discussion with the developers, who agreed to
a moratorium on development until it was determined whether it might
be possible to raise funds to purchase and reserve the habitat. No
precedent for this had occurred in Australia and the discovery occurred,
fortuitously, during the formulation of the State’s Flora and Fauna
Guarantee (see above). The case was widely viewed as a barometer of the
Government’s sincerity in adopting invertebrates as conservation tar-
gets.

A major public appeal was launched to raise funds. The butterfly
became known as the “Eltham Copper,” after the township where it was
discovered, one with a strong tradition of conservation and environmen-
tal awareness. It became a familiar sight on posters, T-shirts, jewellery,
bumper-stickers and as a local emblem. During Australia’s bicentenary
year, “buy the butterfly a birthday” became a popular slogan for the
appeal, and general awareness of butterfly conservation was fostered by
a small booklet (New 1987) sold for the appeal. Over A$425000 was
raised, of which the State Government contributed $250000 and the
Shire Council a further $125000.

Concurrently with the appeal, a detailed search for other colonies was
pursued (Crosby 1987), a tentative management plan formulated
(Vaughan 1987, 1988), and the integrity of the subspecies further
appraised. Isolated colonies in one locality each in central and western
Victoria, both widely separated from the Eltham population, are cur-
rently referred to the same subspecies but all other colonies, mainly
further east, are believed to be the nominate form. Ten colonies were
discovered around Eltham, eight of them very small and considered
unlikely to be viable in the long term. The largest colony, on the
subdivision land, contained an estimated 300-500 larvae but, at the other
extreme, only six butterflies were seen at the smallest colony.

Management recommendations for these remnant urban populations,
contingent on habitat reservation, included (1) protection from the
various threatening processes created by nearby development (such as
garbage dumping, sullage, trampling, slashing or burning vegetation,
weed invasion, activities of domestic animals), (2) provision for habitat
expansion by promoting natural regeneration of foodplants (a stunted
dwarf form of Bursaria spinosa ) and (3) provision of a ranger to foster
practical management and monitor its effects. A major part of the prime
colony habitat was designated for purchase in 1989, and the effect of this
was augmented substantially by the State Government transferring an
area of land adjacent to this, and which also supported another major
colony, to constitute part of the butterfly reserve.

In the wake of the “Eltham Copper issue,” attention has since been paid
to several other butterflies in eastern Australia. One of its closest
relatives, the Bathurst Copper ( P . spinifera Edwards & Common), was

248

J. Res. Lepid.

considered even more vulnerable (Nadolny, 1987: “the most vulnerable
species in N.S.W.;” Kitching & Baker, 1990: “Australia’s rarest butter-
fly”) and until very recently was known from only a single colony in New
South Wales. Its status is at present being investigated in more detail,
but its whole known distribution falls inside a circle of about 18 km
diameter (Kitching & Baker, 1990). The localised coastal populations of
Acrodipsas illidgei in Queensland occupied mangroves subject to clear-
ing for tourist resort development, and its presence was instrumental in
helping to prevent development of some 160 ha near Redland Bay
(Samson 1989).

As elsewhere, taxonomic problems can hamper assessment of conser-
vation status of given populations. The Yellowish Skipper (or Altona
Skipper), Hesperilla flavescens flavescens Waterhouse, has until recently
been believed to be restricted to a few swampy areas in western Victoria,
with largest colonies of this extreme bright phenotype predominantly on
two sites west of Melbourne, where the sole larval foodplant (the sedge
Gahnia filum) grows. These sites have been threatened with industrial
despoliation and urban development, and these factors have been impli-
cated in the decline of the butterfly and the loss of at least one colony.
This taxon appears to be one extreme of a cline involving the highly
variable H. donnysa Hewitson, which has nine named subspecies and
with which H. flavescens Waterhouse was formerly included. A recent
survey by Crosby (1990) has shown that some other populations are
indeed close to, if not identical with, H.f. flavescens , although the
Melbourne colonies are closed populations and considered to represent a
critical stage in speciation in this complex of skippers.

This example, and the Eltham Copper, together illustrate problems of
promoting species conservation in Australia. We are commonly dealing
with discrete and vulnerable populations, but ones whose taxonomic
status is controversial, will be difficult to clarify, and about which
specialists will continue to debate. Lack of “firm” taxonomic status of
such clinal variants tends to weaken the political case for conservation.
Two further examples, both taxa included in the Hill & Michaelis (1988)
“threatened” list, illustrate this further.

a) Tisiphone abeona Joanna , one of eight named “races” of the
Swordgrass Brown, T. abeona (Donovan) (Nymphalidae:Satyrinae) all of
which are restricted in distribution, occurs in one small area around Port
Macquarie in central coastal New South Wales. It has long been known
(Waterhouse 1922, 1928) that this is a hybrid between T.a. morrisi
Waterhouse (to the north) and T.a. aurelia Waterhouse (to the south),
and is apparently maintained consistently at the boundary where these
two “races” meet. Conservation of this localised form therefore depends
not only on direct conservation of its swamp habitat but also on ensuring
continuity of habitat with those of the parent forms. T. abeona is not
particularly vagile, so that habitat fragmentation caused by agricultural
or urban development here could mark the demise of a remarkable and

29(4):237-253, 1990(92)

249

biologically unusual Australian butterfly. New (1984) cited T. aheona as
an example of an endemic evolutionary phenomenon worthy of conserva-
tion for its scientific interest alone.

b) Pseudalmenus chlorinda (Blanchard) (Lycaenidae), the Australian
Hairstreak, is represented by seven named forms in southeastern Aus-
tralia, four of these occurring close together in Tasmania. Many are very
restricted in distribution. Two Tasmanian subspecies and one mainland
one have been listed as threatened. Couchman & Couchman (1977)
noted that in Tasmania the Hairstreak “has been exterminated over wide
areas,” due to various causes. Pasture improvement, involving destruc-
tion of mature eucalypts (used for pupation sites) and acacias (the larval
foodplant), and woodchipping, were cited for P.c. conara Couchman and
P.c. chlorinda (Blanchard), respectively, and the Couchmans also noted
the disappearance of two other distinctive but unnamed Tasmanian
forms due to land clearing, burning and housing development. Soberingly ,
they stated that they located P. chlorinda in more than 50 localities after
1945 but most habitats had been destroyed so that (by 1977) it was
“difficult to think of 10 areas within the island where the Hairstreak may
survive.” In common with many other Lycaenidae, the habitat require-
ments of P. chlorinda include the need for presence of particular species
of ants in an obligate association.

Habitats

Widely recognised as the most important single requisite for conserva-
tion, habitat reservation and management has not occurred widely in
Australia for butterflies alone. As noted earlier, many important vegeta-
tion types have been reduced to small remnants of their former extent,
and there is considerable pressure (on non-butterfly grounds) for many
of these to be reserved, with the likelihood that some butterflies will also
be conserved as “passengers.” Continued threats to rainforests, for
example, will almost inevitably lead to loss of butterflies. Nadolny ( 1987)
commented on significant rain forest sites in New South Wales, and the
wide importance of tropical rain forests on Cape York was emphasised by
Monteith & Hancock (1977). Such key sites, in a range of vegetation
types, are being identified progressively in several States.

In addition to vegetation associations, ranging from alpine grasslands,
and swamps to woodlands and forests, topographical features are impor-
tant. Nadolny emphasised the importance of hill-topping sites for some
species. Some Lycaenidae in Queensland, New South Wales and else-
where, for example, are rarely (if ever) taken except on particular hill
tops, which have become classic collecting localities and entered into
Australian “butterfly folklore.” One or two species are known only from
such sites but, for many, it is not clear where the species breed - although
the phenomenon of hill-topping may well imply that the insects gather
from a considerable surrounding area, it is by no means clear whether the
populations are closed or open.

250

J. Res. Lepid.

The areas of habitat involved in controversy over butterfly species
conservation in Australia have generally been small: for the Eltham
Copper, for example, only some four ha. Samples of many Australian
ecosystems are indeed included in National Parks and other reserves but
for many of these areas no definitive species list of butterflies is available,
and management to conserve particular rare taxa is not undertaken.
Most of the lists which do exist reflect sporadic visits by collectors rather
than any attempt at comprehensive survey, although a number of more
rigorous surveys are in progress.

Few butterflies in Australia are regarded as “umbrella species,” al-
though this status could possibly be accorded to some alpine species, such
as most species of Oreixenica Waterhouse & Lyell (Satyrinae), as these
are very specialised forms associated with characteristic and restricted
alpine herbfield/grassland communities. They are frequently abundant
over their limited altitudinal ranges, and any pronounced diminution in
abundance may reflect wider-reaching effects on those communities.
Here, and in some other restricted habitats, Australian butterflies could
be valuable as indicators of community wellbeing, and conservation of
the conditions suitable for them also ensure the persistence of other, less
conspicuous, sensitive alpine fauna.

The Future

There is little doubt that further degradation to the Australian envi-
ronment will occur. In addition to continuing pressures for forest
(including rainforest) exploitation for timber, woodland clearing, and
other human interventions - including development of alpine and remote
coastal sites for recreation - the ramifications of possible global warming
(the “Greenhouse Effect”) give cause for concern. These influences
together are likely to result in further decline of many butterflies, and
global warming could lead, for example, to severe contraction of alpine
habitats in their present form. Decline of some butterflies will inevitably
go unheralded, and insufficient is known of the biology of most of the
rarer species to form the basis of informed management plans. This
foundation, together with knowledge of distribution and precise taxo-
nomic status, is gradually being strengthened, but the practical likeli-
hood is that most species will remain as passengers in conservation
activity for the foreseeable future. As elsewhere in the world, presence
of rare or unusual “priority” species may strengthen the political case for
reservation of particular habitats, even though they may not then be able
to be managed effectively for those species. This is, perhaps, especially
true for the labour-intensive maintenance of early serai vegetation
stages to which many notable taxa are largely restricted, and butterfly
conservationists in Australia will continue to draw on information from
cases elsewhere in the world in planning optimal conservation or manage-
ment strategies.

Urban reserves are becoming commonplace, as people in rapidly

29(4):237-253, 1990(92)

251

expanding cities seek to conserve even small remnants of native vegeta-
tion, and restoration of these is accelerating in intensity. There is some
potential for translocation of butterflies to these, though this has so far
been rare. A large skipper, Trapezites symmomus Hubner, found to the
north of Melbourne has recently been experimentally introduced to a
reserve managed by La Trobe University after a habitat was prepared for
it by planting of larval foodplants, and this could mark the way for other
such introductions. A strong move towards preferential growing of
Australian native plants in gardens is also evident, replacing the tradi-
tional concentration on exotic species, and there is some developing
interest in butterfly gardening, fostered in part by horticultural groups.
All these facets may help to counter the range declines and local
extinctions which have caused comment from collectors in Australia
since the late nineteenth century.

Public awareness of, and education on, butterflies in Australia has
recently been increased through Butterfly House exhibits, predomi-
nantly in Melbourne (the Melbourne Zoological Gardens) and several
tourist - based exhibits in Queensland. The latter enhance the image of
P. ulysses, widely adopted as a tourism emblem in tropical Queensland.
There is clearly considerable opportunity for conservation awareness to
emanate from such ventures. Because of lack of opportunity to import
live material, all such operations must depend entirely on native butter-
flies, and increased knowledge of their captive breeding and mainte-
nance is a natural result of this.

A start has, thus, been made on several aspects of practical butterfly
conservation in Australia. Specific cases, such as the Eltham Copper,
still have considerable novelty value, and it is highly unlikely that
hundreds of thousands of dollars will be made available for each of the
many similar cases which may arise in the future. But more people than
ever before are aware of butterfly conservation, the topic is no longer
treated with universal disdain by politicians and the public, and the
future for butterflies in Australia can be viewed with, at least, a
reasonable level of optimism because of this increasing interest.

Literature Cited

Anderson, e. & F.P. Spry, 1893- 1894. Victorian Butterflies and how to collect them.
Hearne & Co., Melbourne.

Braby, M.F., 1987. The Eltham Copper: a butterfly with a special appeal.
Habitat Australia 15 (6): 16-18.

, 1990. The life history and biology of Paralucia pyrodiscus lucida Crosby

(Lepidoptera: Lycaenidae). J. Aust. ent. Soc., 29: 41-50.

Common, I.F.B. & D.F. Waterhouse, 1972. Butterflies of Australia. Angus &
Robertson, Sydney.

, 1981. Butterflies of Australia. Revised Ed., Angus & Robertson, Sydney.

Crosby, D.F., 1951. A new geographical race of an Australian butterfly. Vic. Nat.
67: 225-226.

, 1987. The conservation status of the Eltham Copper Butterfly ( Paralucia

252

J. Res. Lepid.

pyrodiscus lucida Crosby (Lepidoptera:Lycaenidae). Technical Report Series
No.81 Arthur Rylah Institute for Environmental Research, Melbourne.

, 1990. A Management Plan for the Altona Skipper Butterfly Hesperilla

flavescens flavescens Waterhouse (Lepidoptera:Hesperiidae). Technical Report
Series No 98. Arthur Rylah Institute for Environmental Research, Melbourne.

Couchman, L.E. & Couchman, R., 1977. Butterflies of Tasmania. Tasmanian Year
Book, 1977: 66-96, Hobart.

d’Abrera,B., 1971. Butterflies of the Australian Region. Lansdowne, Melbourne.

Dunn, K.L. & L.E. Dunn, 1990. A computer database on Australian Butterflies.
Myrmecia 26: 37-39.

Entomological Society of Victoria (ESV), 1986. Preliminary distribution maps
of butterflies in Victoria, Melbourne.

Greenslade, P. & T.R. New, 1991. Australia: conservation of a continental insect
fauna. In Collins, N.M. & Thomas, J.A. (eds.) Conservation of Insects and
their Habitats. Academic Press, London pp. 33-70.

Hill, L. & F.B. Michaelis, 1988. Conservation of Insects and Related Wildlife.
Occasional Paper no. 13, Australian National Parks and Wildlife Service,
Canberra.

Ive, J. & D. Cocks, 1989. Rural land degradation in Australia. The role of trees
in ecological solutions. Aust. Conserv. Farmer, 1 (3): 18-22.

IUCN, 1988. The IUCN Red list of Threatened Animals. IUCN, Gland.

Key, K.H.L., 1978. Conservation status of Australia’s insect fauna. Occasional
Paper No.l, Australian National Parks and Wildlife Service, Canberra.

Kitching, R.L., 1981. The Geography of the Australian Papilionoidea. In: Keast,
A., ed., Ecological Biogeography of Australia. W. Junk, The Hague, pp. 979-
1005.

Kitching, R.L. & E.J. Baker, 1990. Hello, Goodbye? Geo 12, 92-95.

Kitching, R.L., E.D. Edwards, D. Ferguson, M.B. Fletcher & J.M. Walker, 1978.
The butterflies of the Australian Capital Territory. J. Aust. ent. Soc. 17: 125-
133.

Masters, G., 1873. Catalogue of the described diurnal Lepidoptera of Australia,
Sydney.

McCubbin, C., 1971. Australian Butterflies. Nelson, Melbourne.

Monteith, G.B., 1987. Australian Federal Import/Export controls on insect
specimens and their effect on insect conservation. In: Majer, J.D., ed., The
role of invertebrates in conservation and biological survey. Dept, of
Conservation and Land Management, Perth, W.A., pp. 21-30.

Monteith, G.B. & D.L. Hancock, 1977. Range extensions and notable records for
butterflies of Cape York peninsula, Australia. Aust. ent. Mag. 4: 21-38.

Moulds, M.S., 1977. Bibliography of the Australian Butterflies, 1773-1973.
Australian Entomological Press, Greenwich, NSW.

Nadolny, C., 1987. Rainforest butterflies in New South Wales: their ecology,
distribution and conservation. NSW National Parks and Wildlife Service,
Sydney.

New, T.R., 1984. Insect conservation: an Australian perspective. W. Junk,
Dordrecht.

, 1987. Butterfly Conservation. Entomological Society of Victoria, Melbourne.

Olliff, A.S., 1989. Australian butterflies: a brief account of the native families,
with a chapter on collecting and preserving insects. Nat. Hist. Assoc. N.S.W.,
Sydney.

29(4):237-253, 1990(92)

253

Rainbow, W.J., 1907. A guide to the study of Australian Butterflies. Lothian,
Melbourne.

Samson, P.R., 1989. Morphology and biology of Acrodipsas illidgei (Waterhouse
& Lyell), a myrmecophagous lycaenid (Lepidoptera:Lycaenidae:Theclinae).
J. Aust. ent. Soc., 28: 161-168.

State of the Environment in Queensland, 1990. Department of Environment and
Heritage, Brisbane, Queensland.

Symons, D.E., 1980. The foodplants of Australian butterfly larvae. J. Adelaide
Bot. Gard. 2: 277-292.

Taylor, R.W., 1976. (A submission to the inquiry into the impact on the
Australian environment of the current woodchip industry programme). In
Australian Senate, Hansard (Reference: Woodchip Inquiry) pp. 3724-3731.

Vaughan, P.J., 1987. The Eltham Copper Butterfly draft management plan.
Technical Report Series No. 57. Arthur Rylah Institute for Environmental
Research, Melbourne.

, 1988. Management plan for the Eltham Copper Butterfly £ Paralucia

pyrodiscus lucida Crosby) ( Lepidopter a : Lycaenidae ) . Technical Report Series
No. 79. Arthur Rylah Institute for Environmental Research, Melbourne.

Waterhouse, G.A., 1922. An account of some breeding experiments with the
satyrine genus Tisiphone. Proc. Linn. Soc. N.S.W. 47: i-xvii.

, 1928. A second monograph of the genus Tisiphone , Hubner. Aust. Zool., 5:

217-240.

, 1932. What Butterfly is that? Angus & Robertson, Sydney.

Waterhouse, G.A., & G. Lyell, 1914. The Butterflies of Australia. Angus &
Robertson, Sydney.

Wells, S.M., R.M. Pyle & N.M. Collins, 1983. The IUCN Invertebrate Red Data
Book. IUCN, Gland.

Appendix 1.

Australian Butterflies listed as “threatened animals” by IUCN (1988)

Hesperiidae

Antipodia chaostola leucophaea

Tasmania

Euschemon rafflesia

E. Australia

Hesperilla mastersi marakupa

Tasmania

Oreisplanus munionga larana

Tasmania

Lycaenidae

Pseudalmenus chlorinda chlorinda

Tasmania

P.c. conara

Tasmania

Nymphalidae

Heteronympha cordace comptena

Tasmania

Journal of Research on the Lepidoptera

29(4):254-266, 1990(92)

Changes of distribution of thermophilous
butterflies in Slovakia

Miroslav Kulfan

Univerzita Komenskeho, Prirodovedecka fakulta, Katedra zoologie, Mlynska dolina B-l, CS-
84215 Bratislava, CZECHOSLOVAKIA

and

Jan Kulfan

Ustav ekologie lesa SAV, V. I. Lenina 2, CS-96053 Zvolen, CZECHOSLOVAKIA

Abstract. Data from the literature and original studies concerning
thermophilous butterflies and skippers (Papilionoidea and
Hesperoidea) of the Pannonian region of Slovakia are presented.
Endangered, vulnerable, extinct, and species with no record of perma-
nent occurrence are discussed. Anthropogenic factors harmful to
thermophilous species are analyzed, including impacts of drainage,
artificial afforestation, use of chemical compounds, cattle grazing,
butterfly collecting, etc. Special attention is given to the state nature
reserves of south Slovakia and their lepidoptera habitats. For long
term protection of thermophilous butterflies in Slovakia it will be
necessary to preserve the integrity of their habitats, to reduce negative
anthropogenic factors and to implement further research on the
biology and population structure of these butterflies. International
cooperation for their protection will be necessary as well.

Introduction

The aim of this paper is to provide information on the distribution of
butterflies (Papilionoidea and Hesperoidea) across the Pannonian re-
gion of Slovakia. The region is the warmest in Slovakia. It includes
lowlands, plains, basins, mountains and hills, promontories, a plateau,
and highlands: Borska nizina (lowlands), Chvojnicka pahorkatina
(hills), southwestern part of Biele Karpaty (mountains), southwestern
and eastern parts of Myjavska pahorkatina (hills), Trnavska pahorkatina
(hills), southern and low-lying parts of Male Karpaty (mountains),
Podunajska rovina (plain), the southernmost part of Povazsky Inovec
(mountains), Nitrianska pahorkatina (hills), Hornonitrianska kotlina
(basin), Zitavska pahorkatina (hills), the southernmost part of Tribec
(mountains), Hronska pahorkatina (hills), Ipelska pahorkatina (hills),
southern part of Krupinska planina (plateau), Ipelska kotlina (basin),
Lucenska kotlina (basin), Cerova vrchovina (highlands), Rimavska
kotlina (basin), Bodvianska pahorkatina (hills), Slovensky kras (pla-
teau), Roznavska kotlina (basin), Kosicka kotlina (basin), southern part
of the Slanske vrchy (highland), Vychodoslovenska pahorkatina (hills),
southern part of the Ondavska vrchovina (highlands), Beskydske
predhorie (promontory), Zemplinske vrchy (highlands), and
Vychodoslovenska rovina (plain). The overall area is mapped as Fig. 1.

29(4):254-266, 1990(92)

255

Publications concerned with butterflies of the Pannonian re-
gion

Earlier faunistic data on butterflies from the Pannonian region are
summarized in Prodromus Lepidopter Slovenska (Hruby, 1964). Since
publication of this work, faunistic research has progressed, but several
territories of the Pannonian region are still not well understood faunis-
tically. The Male Karpaty and nearby hills belong to the best explored
regions of western Slovakia. These regions were studied by several
lepidopterologists including Caputa (1968a, b, 1970) who described the
butterfly fauna of the state nature reserve, Devinska Kobyla; M. Kulfan
(1982) who published the distribution of butterflies from the northern
part of the Male Karpaty and Myjavska pahorkatina; and J. Kulfan
(1990) who investigated communities of butterflies in the southern part
of the Male Karpaty.

Rysavy (1984) and J. Kulfan (1990) studied the lowland butterfly
communities of Borska nizina while the warmest Slovakian mountain
region, Kovacovske kopce (Burda) and nearby territory was explored by
Moucha (1959, 1961), Begin a et al. (1984) and Caputa (1987).. Other
authors presenting faunistic data from the Pannonian region of west
Slovakia included Kristof (1973), Kubanik (1979), Madlen (1971, 1974,
1975, 1976) and Sachl (1980).

A few publications refer to studies of the smallest part of the Pannonian
region located across central Slovakia (e.g., Caputa, 1963; Skypala,
1978; J. Kulfan, 1989). There are many references to the butterfly fauna
of the Pannonian region of east Slovakia, however (e.g., Brunnerova et
al. , 1984; Caputa , 1960; Ceplik, 1976; Panigaj, 1983; Vacula, 1975;

256

J Res . Lepid.

Vacula et al. , 197 1 ; Vacula and Vahala, 1973;Zuskinova, 1971;Lastuvka,
1988). In addition Caputa ( I960) and Lastnvka ( 1988) explored butterfly
communities in the largest Slovakian karst, Slovensky kras (Zadiel,
Plesivecka planina plateau).

The occurrence of individual species were reported by Kulfan et al ,
1986; Stiova, 1976; Svestka, 1979; Zelny, 1961 and others. Migratory
species in the Pannonian region have been reported by Moucha, Slimak,
Cerny, Jakes, Lastuvka, Pipek, Cervinka, Ivirnig, M. Kulfan and Gottwald
(in Moucha, 1965; Felix, 1971; Felix and Soldat, 1972; Felix et al , 1978;
Pipek and Soldat, 1979, 1980).

Development of the ecosystems of the Pannonian region

The Pannonian region can be characterized as mixed deciduous woods
which represent both the original and climax communities. During the
past 12,000 years, however, natural undisturbed vegetation processes
came to an end. Extensive forest clearance took place as the landscape
was colonized by Neolithic farmers. Forest areas were transformed into
derived meadow, pasture, and crop ecosystems which subsequently gave
rise locally to secondary forest ecosystems. As a result, there were
corresponding changes in the climate and in hydrology as the forest soils
were converted to agricultural land. Consequently, conditions evolved
which promoted the development of new vegetation types and land-
scapes. These new ecosystems were formed from adaptable elements of
the forest flora, by relict taxa that survived in extreme sites (as on blown
sand, flood plains, saline soil, peat bogs, and rocks), and finally by a
series of exotic species introduced by the early agricultural peoples
(Rybnicek and Rybnickova, 1988). From 7,000 to 5,000 YRP Neolithic
farmers settled into and transformed the hill areas of southwestern and
southeastern Slovakia. About 3,000 YBP there was a gradual occupa-
tion and deforestation of the submontane areas and the broad valleys of
large rivers.

During the Late Glacial period of 12,000 to 10,000 YBP, the southern
part of Slovakia was covered by forest-steppe and steppes of more or less
continental character. Pine and birch were the prevalent timber species
but some more demanding species were present also including oak and
elm. The alluvia of large rivers were covered extensively with aquatic
and swamp vegetation along with willow stands.

During the Early Holocene (Boreal) of 10,000 to 8,000 YBP, the region
was already covered by an open forest of pine, birch, elm and oak, with
an undergrowth of hazel and other shrubs and with a well developed
herbaceous layer. Earlier willow stands were gradually replaced by
floodplain, so since the end of the Boreal period, there were alder woods
with birch and possibly some oak. During the early Atlantic period
(8,000 YBP), the forested landscape gradually began a transformation to
an agricultural land. Since that time there has been a general contrac-
tion of forest stands and a concomitant increase of fields, meadows and

29(4):254-266, 1990(92)

257

pastures. Remnants of the woods that were the original open
thermophilous oak and mesophilous oak forest were replaced by horn-
beam oak woods, at first with oaks and elms dominant, and later during
early Holocene with hornbeam.

On poor sandy soil, e.g., the Zahorie region, oak-pine woods dominated.
Since the Atlantic period, but especially since the sub-Boreal, a second-
ary steppe-like vegetation evolved. Alluvial portions of the Podunajska
rovina and Vychodoslovenska rovina plains were covered by floodplain
woods since mid-Holocene. These landscapes were similar in appearance
to those of the present day, containing alder, willows, poplars, oaks, elms,
ash, and later hornbeam. Alders and willows prevailed in smaller river
valleys (Rybnicek and Rybnickova, 1986).

Before man intervened by converting the vegetation across this
country 6,000-7,000 YBP, forest covered about 90% of the area. By the
14th century about 60% of Slovakia was covered by forests. Pines and
exotic tree species were extensively planted on deteriorated and stony
soils. In vineyard areas the spread of the black locust, Robinia
pseudoacacia L., was encouraged, as this exotic plant was important for
bee culture while providing good timber for wheel making and firewood.
In recent years black locust monoculture has expanded in the southern
part of south Slovakia and now covers more than 50% of forest area in
many places.

Butterfly fauna of the Pannonian region

The Appendix lists all the species found in the Pannonian region. The
list shows that 165 species of butterflies historically occurred in the
Pannonian region, of which 148 species are probably still present. Forty
six Slovakian species of butterflies occur chiefly in the Pannonian
region. The following species are characteristic of this region, but also
occur in the warm sites of the mountains of West Carpathia (Zapadne
Karpaty mountains): Pyrgus serratulae, P. sertorius, I. podalirius,
Papilio machaon, Colias alfacariensis, Leptidea morsei, Aricia agestis,
Callophrys rubi, Glaucopsyche alexis , Maculinea arion, Nordmannia
pruni, N. spini , Polyommatus bellargus, P. coridon, P. daphnis, P.
dorylas, Scolitantides orion, Chazara briseis, Hipparchia semele,
Melitaea aurelia, and M. didyma.

Species of special significance.

ENDANGERED AND VULNERABLE SPECIES OF XEROTHER-
MIC SITES, MEADOWS AND DAMP BIOTOPES.

Carcharodus flocciferus, Pyrgus armoricanus, Colias chrysotheme,
Glaucopsyche alexis, Lycaena thersamon, Maculinea arion, Polyommatus
admetus, P. amandus, P. bellargus, P. damon, P. daphnis, P. eroides,
Pseudophilotes schiffermuelleri, Scolitantides orion, Melitaea aurelia,
M. britomartis, M. fascelis and M. phoebe are species which inhabit
xerothermic biotopes. They are all endangered, chiefly by afforestation,

258

J. Res. Lepid.

fires, natural forest and shrubs succession, agricultural conversion with
increasing areas of vineyards, gardens and orchards, waste disposal
sites, cattle and sheep grazing with consequent affect on erosion, plant
community composition, and application of insecticides and herbicides
in the vicinity of agricultural areas. These are also sites which are
considered to be soils without practical use. Such sites are used for
building and for sports, as motocross. Some of these species are moving
to secondary sites formed by man, as railway embankments and road
cuts. Generally, conservation of butterfly species cited is positively
influenced by extensive grazing because continuous shrub and tree
stands are reduced.

Nordmannia acaciae, N. pruni, N. spini and Thecla hetulae occur in
similar sites but require the presence of woody plants at the proper
growth stage which are the optimal foodplants of their larvae.

Euphydryas maturna is now present in only a few populations in the
Pannonian region. Besides the harmful anthropogenic factors dis-
cussed, the species is endangered by both overcollecting and by changes
in forestry management. The causes of the rapid decline of Lasiommata
achine are not clear over the last decades. Nordmannia w-alhum is
endangered through change in timber composition in the forest ecosys-
tem.

Neptis sappho occurred regularly at many localities a few decades ago.
The causes of its decline are not clear.

Hipparchia hermione, known in the Central European literature as H.
alcyone or H. aelia, and H. statilinus are inhabitants of sandy biotopes
as the Broska nizina lowland and in the sand dune grasslands of the
southernmost part of Slovakia. Preservation of these sandy biotopes in
their present state is necessary for their long term survival.

Colias myrmidone and Brenthis hecate are represented by some
isolated populations in the Pannonian region. The two species occur
both in xerothermic biotopes and mesophilous and wet meadows. Rela-
tively dense populations of these species occur in the southwestern part
of Biele Karpaty mountains in extensively managed meadows which are
mowed once a year. Both species are now endangered by changes of
meadow management including fertilization, cattle and sheep grazing,
and ploughing.

Arida eumedon, Lycaena aldphron, Maculinea alcon, M. nausithous,
M. teleius, Coenonympha tullia, Brenthis ino , Euphydryas aurinia and
Melitaea diamina are inhabitants of wet meadow biotopes. They are
highly endangered through changes of meadow management and wet-
land drainage.

Lycaena dispar is endangered where it occurs on xerothermic sites
near wet biotopes as in the Male Karpaty mountains and Cachtice hills.

Neptis rivularis occurs as a local riparian species along streams chiefly
in inaccessible places. It will be endangered by construction of thorough-
fares and the artificial embankment of natural watercourses.

29(4):254-266, 1990(92)

259

Parnassius mnemosyne occurs chiefly in forest-steppe biotopes in the
Pannonian region. The ecological situation is unlike that in mountain
biotopes where the species prefers forest meadows and glades. The
principles of conservation cited above for xerothermic species are valid
for P . mnemosyne. Its isolated and small populations are highly vulner-
able to overcollecting.

Zerynthia polyxena is a characteristic species of the Pannonian region,
occurring near vineyards, along railways and roads, and near streams.
There are numerous populations of Z. polyxena , especially in the region
of the Podunajska rovina plain and the Borska nizina lowland near large
streams. The many isolated populations are chiefly endangered along
the northern border of the Pannonian region by vegetation burning on
road and railway embankments, application of insecticides and pesti-
cides, repairs along the thoroughfares, and collection of both adults and
larvae.

INDETERMINATE SPECIES.

Pyrgus serratulae, Spialia orbifer, Leptidea morsel , Pieris mannii,
Aricia allous, Cupido alcetas, C. osiris, Lycaeides idas, Maculinea rebeli
and Polyommatus thersites are indeterminate species, species which are
probably endangered in the Pannonian region. Their distribution is
insufficiently known because they all can be confused easily with
commonplace species at sites where butterfly communities are dense.

There is insufficient information concerning Thymelicus acteon. This
species is easily overlooked by collectors.

Limenitis reducta often escapes recording because collecting voucher
specimens is not easy. It may also be mistaken for L. Camilla during
flight.

PROBLEMATIC SPECIES.

There are 17 species of problematic occurrence in the Pannonian
region at present. Information on their occurrence arises chiefly from
very old records: Carcharodus lavatherae, Parnassius apollo, Colias
palaeno, Cupido osiris, Iolana iolas, Lampides boeticus, Lycaena helle,
Syntarucus pirithous, Vacciniina optilete, Coenonympha hero, C.
oedippus, Pyronia tithonus, Argynnis pandora, Boloria aquilonaris, B.
eunomia, B. titania, and Nymphalis l-album.

In the recent past populations of Parnassius apollo were probably
present on promontories of the west Carpathians. At present this
species is not known even in the vicinity of Slovensky kras plateau
(Lastuvka, 1988). Many populations of this species have disappeared
across all of Europe as the result of the conversion of biotopes.

It appears that Argynnis pandora is no longer present in Slovakia. The
species occurred frequently at many localities of south Slovakia as
recently as a few decades ago. In past years Nymphalis l-album and N.
xanthomelas were observed, but only sporadically. N. xanthomelas was

260

J. Res. Lepid.

found in the Kovacovske kopce hills about 10 years ago (Caputa, 1987),
and probably still occurs in south Slovakia. Long term climatic changes
may well become a major cause of extinction of these species in the near
future.

The butterfly fauna of the Pannonian region may be sporadically
augmented by migrants from the south. Lampides boeticus and Sytarucus
pirithous appear sporadically in the Pannonian region. Colias crocea
appears very frequently and can be classified as an abundant species in
many years.

Colias erate first appeared in Slovakia in 1989 near the town Nove
Zamky (Petru and Bohm, in press). In 1990 it was found at many
localities of the Pannonian region, near Sturovo, Komarno, Piestany,
Nove Mesto n. Vah (J. Marek, M. Svestka, L. Vitaz and J. Patocka, pers.
comm.). Females were observed laying eggs and fresh adults were
observed which clearly indicate the species can permanently breed in
the Pannonian region. Adults of C. erate were observed across the same
biotopes as Colias crocea. We cannot predict whether C erate will be
permanently established in the Pannonian in the future. Its appearance
may be an indicator of long term climatic change, although the possibil-
ity of adaptive change within the species’ genetic system may explain
this new distribution.

Very recently Heteropterus morpheus and Cupido decoloratus have
expanded their ranges across the Pannonian region. Both species were
local and rare in the past. For example, Heteropterus morpheus was not
recorded from the state nature reserve Devinska Kobyla, near Bratislava,
in the 1960s (Caputa, 1970). At present it occurs regularly at this
locality. It is also known from many localities of southwestern Slovakia
(Reiprich and Okali, 1989; J. Kulfan, 1990). Cupido decoloratus occurs
at many localities of southern Slovakia along streams in the north (M.
Kulfan, 1982; Kulfan et al ., 1986; J. Kulfan, 1989, 1990; Reiprich and
Okali, 1989).

Conservation of biotopes

Protection of butterflies in the Pannonian region is not a simple
matter, mainly because the region contains mostly intensively utilized
agricultural land. Consequently, most meaningful conservation efforts
for the butterfly fauna can be carried out in the hills and the mountain
slopes. Butterflies which occur across the narrow zones of forest-steppe
vegetation between vineyards and forest stands, as in the Male Karpaty
and Tribec mountains, Krupinska planina plateau and Kovacovske
kopce hills, are now endangered by intensive viticulture with predict-
able total extirpation. The biotopes of Borska nizina lowlands represent
sand dunes grasslands which are changing not only by intensive
agriculture but also by continuous afforestation, both natural and
artificial, and by construction of tourist cabins and holiday homes.

There are damp biotopes of relatively large areas in the Borska nizina

29(4):254-266, 1990(92)

261

lowlands, many situated in the borderland near Austria. This territory
was inaccessible until a year ago. The land has now been intensively
utilized for agriculture, and wetland drainage will cause destruction of
these biotopes. Chemical application for mosquito control compounds
the problem of survival of this specialized butterfly fauna.

There are local saline soils in some places of south Slovakia, but the
remaining biotopes of these sites are isolated. No lepidopterological
research has been completed on these biotopes.

The spread of the introduced alien black locust, Robinia pseudoacacia,
has a very negative effect on the natural Pannonian vegetation. As a
legume it also modifies soil chemistry by increasing nitrogen levels in the
soil.

In light of the massive negative anthropogenic influences across a
substantial part of the region, state nature reserves and some preserved
landscape zones are now vital for the conservation of the butterfly fauna.
Too few reserves occur in the Pannonian region at the present time, and
they are often isolated without connections that can increase
metapopulation survival. In the future it will be necessary to form a
network of the reserves as well as a network of zones without intensive
agriculture. Only in such manner can reserves be interconnected, and
long term persistence of habitat values maintained.

Another management technique for the conservation of biotope diver-
sity would be necessary: short term extensive grazing and planned
mowing and burning. For instance, today there is extensive sheep
grazing on forest-steppe biotopes on the Hainburg Bergen in Austria
near the state nature reserve Devinska Kobyla in Male Karpaty moun-
tains. Here the spread of undesirable herbs, shrubs and trees is reduced
by sheep grazing. In addition, across some smaller areas weed burning
is employed. Similar management techniques are planned for the state
nature reserve Devinska Kobyla. All these methods increase floral
diversity and will positively impact butterfly survival.

Flowery meadows can be mowed, most appropriately during two
periods, each on half the land area, so that butterflies could move from
the mowed section to the unmowed one.

State nature reserves in the Pannonian region possess a rich butterfly
fauna. For example, in S. N. R. Devinska Kobyla (Caputa, 1970; M.
Kulfan, collection) S. N. R. Kovacovske kopce hills (Hruby, 1964;
Reiprich, 1977; Caputa, 1987) and S. N. R. Cachticke hradny vrch and
its near vicinity (M. Kulfan, 1982; Reiprich and Okali, 1989; M. Kulfan,
collection; Vitaz, pers. comm.), respectively 80, 116 and 101 species have
been recorded. A rich community of butterflies, including 31 species of
Lycaenidae, are concentrated on a small area in Cachticke kopce hills
(Male Karpaty mountains) (M. Kulfan, 1982; M. Kulfan, collection).
Therefore the Cachticke kopce hills can be identified as a butterfly
reservation containing both xerothermic and wet biotopes.

The greatest number of butterfly species occurs across the Plesivecka
planina plateau. This is part of a preserved landscape area, Slovensky

262

J. Res. Lepid.

kras plateau. Prior to 1988 121 species of butterflies occurred here
(Hruby, 1964; Lastuvka, 1988; Reiprich and Okali, 1989).

For protection of a thermophilus butterfly fauna in Slovakia, it will be
necessary to considerably increase our knowledge of their population
dynamics and ecology. In addition comprehensive inventories of many
present reserves do not exist. However, research must rapidly be
directed not only to the preserved areas, including the monitoring of
selected indicator species with estimates of population sizes, but also
extended to the regions which should be preserved for the future. In
other countries of central and south Europe there are biotopes that are
related to those in the Pannonian region of Slovakia and international
cooperation relating to the preservation of butterflies of these sites
should begin.

Literature cited

Brunnerova, L., P. Degma, M. Kulfan and I. Pitonakova, 1984. Splocenstva hmyzu
v oblasti Vihorlatu, pp. 46-51. /raiVII. vychodoslovensky tabor ochrancov
prirody 1984, prehlad odbornych vysledkov.

Caputa, A., 1960. Prispevok k poznaniu Lepidopter Zadielu a jeho okolia. Biologia
(Bratislava), 15: 130-132.

— ,1963. Prispevok k poznaniu Rhopalocera juznej casti stredneho Slovenska.

Biologia (Bratislava) Ser. B, 18: 604-611.

— , 1968a. Pozoruhodne nalezy motylov v rezervacii Devinska Kobyla. Zbor. Slov.
nar. Muz., Prir. vedy, 14: 99-106.

— , 1968b. Vyskyt vzacnych a chranenych druhov hmyzu v rezervacii Devinska
Kobyla. Ochr. Fauny, 2: 9-13.

— , 1970. Lepidoptera prirodnej rezervacie Devinska Kobyla. Ent. Probl., 8: 55-
153.

— , 1987. Vyskum motylov (Lepidoptera) Statnej prirodnej rezervacie Kovacovske
kopce. Ochr. Prir. Vyskumne prace z ochrany prirody 5. Priroda, Bratislava,

128 pp.

Ceplik, J., 1976. Vyskyt motylov babockovitych (Nymphalidae) na vychodnom
Slovensku. Mlady prirodovedec, 17: 18-19.

Degma, P., J. Gregor and M. Kulfan, 1984. Prispevok k poznaniu niektorych
lokalit Burdy a Ipelskej pahorkatiny. Spr. Slov. zool. Spoloc. SAV, 10: 72-76.
Felix, V., 1971. Pozorovani taznych motylu v Ceskoslovensku v letech 1965-1969.

Zpr. Ceskoslov. Spolec. ent. CSAV, Praha, 7: 7-26, 29-63.

Felix, V., P. Pipek and M. Soldat, 1978. Zprava o pozorovani taznych mozylu v
Ceskoslovensku v letech 1972-1974, 1975, 1976. Zpr. Ceskoslov. Spolec. ent.
CSAV, Praha, 14: 41-92.

Felix, V. andM. Soldat, 1972. Zprava o pozorovani taznych motylu v Ceskoslovensku
v letech 1970-1971. Zpr. Ceskoslov. Spolec ent. CSAV, Praha, 8: 125-150.
Hruby, K., 1964. Prodromus Lepidopter Slovenska. Vydavatel SAV, Bratislava,
962 pp.

Kristof, J., 1973. Vyskyt vzacnych a chranenych druhov hmyzu v juznej casti
levickeho okresu. Ochr. Prir. (Praha), 28, Priloha Ochranarsky pruzkum, 32.
Kubanik, S., 1979. Motyle (Lepidoptera) chraneneho naleziska Sedlisko (Soros)
a jeho okolia. Zapadne Slovensko, 6: 184-208.

29(4):254-266, 1990(92)

263

Kulfan, J., 1989. Heliofilne motyle (Lepidoptera) extenzivne obhospodarovanej
krajiny pri Krupine, pp. 237-248. /mBitusik, P. and J. Galvanek, (eds.),
Stredne Slovensko 8 - Prirodne vedy, Osveta, Martin.

— ,1990. Die Struktur der Taxozonosen von heliophilen Faltern (Lepidoptera) an
manchen Biotopen der Westslowakei. Biologia (Bratislava), 45: 117-126.
Kulfan, M., 1982. Lepidoptera severnej casti Malych Karpat a Myjavskej
pahorkatiny. Biol. Prace. 28, Vydavatel. SAV, Bratislava, 112 pp.

Kulfan, M., F. Slamka and L. Misko, 1986. Revizia druhov Everes alcetas
(Hoffmansegg, 1804) a Everes decoloratus (Staudinger, 1886) (Lepidoptera,
Lycaenidae) na Slovensku. Biologia (Bratislava), 41: 993-998.

Lastuvka, Z., 1988. Motyli (Lepidoptera) Plesivecke planiny. Ochr. Prir.

Vyskumne prace z ochrany prirody 6B: 303-322.

Madlen, J., 1971. Niekolko poznamok k letnej aktivite Lepidopter z hornej hranice
sahanskej stepnej oblasti. Vlastivedny Sprav. Levickeho okresu, 7: 18-22.
— , 1974. Rhopalocera z vrabelskeho okolia v studiach Stefana Necseyho. Zbor.
Pedag. Fak. Univ. Komenskeho v Bratislave so sidlom v Trnave, Prir. Vedy,
7: 127-135.

— , 1975. Motyle v studiach Stefana Necseyho. Levice, Tekovske muzeum, 186

pp.

— , 1976. Babockovite, Nymphalidae, Lepidoptera z okolia Siah. Zbor. prac Pedag.

Fak. Banska Bystrica, Ser. Prir. Biol, et Agrolog., 7: 75-102.

Moucha, J., 1959. K poznani motylu (Lepidoptera) rezervace “Kovacovske kopce”
na jiznim Slovensku. Ochr. Prir. (Praha), 14: 173-175.

— •, 1961. Soumracnik Pyrgus armoricanus Obth. ve statni prirodni rezervaci
“Kovacovske kopce” na jiznim Slovensku. Cas. Nar. Muz. v Prahe, odd.
prirodovedny, 130: 172-175.

— , 1965. Zprava o vysledcich pozorovani taznych motylu v Ceskoslovensku za rok
1964. Zpr. Ceskoslov. Spolec. ent. CSAV, Praha, 1: 10-15.

Panigaj, L., 1983. Z vysledkov prace entomologickej sekcie na I. Tabore ochrancov
prirody v okrese Kosice - vidiek. Pamiat.-Prir., Bratislava, 14: 24-26.
Petru, M., and S. Bohm, in press. Faunistic record from Czechoslovakia -
Lepidoptera. Acta Entomol. Bohemoslov.

Pipek, P. and M. Soldat, 1979. Zprava o pozorovani taznych motylu v
Ceskoslovensku v roce 1977. Zpr. Ceskoslov. Spolec. ent. CSAV, Praha, 15: 65-
73.

— , 1980. Zprava o pozorovani taznych motylu v Ceskoslovensku v roce 1978. Zpr.

Ceskoslov. Spolec. ent. CSAV, Praha, 16: 35-41.

Reiprich, A., 1977. Doplnky k Prodromu Lepidopter Slovenska. Ent. Probl.
(Bratislava), 14: 13-69.

Reiprich, A. and I. Okali, 1989. Dodatky k prodromu Lepidopter Slovenska. 3.
Vydavatel. SAV, Bratislava, 144 pp.

Rybnicek, K. and E. Rybnickova, 1986. Vyvoj po dobe ladovej, pp. 31-36. In:
Michalko, J., (ed.), Geobotanicka mapa - textova cast. Vydavatel. SAV,
Bratislava.

Rysavy, J., 1984. Prispevek k faunistice Lepidopter slovenskeho Zahori. Zpr.

Ceskoslov. Spolec. ent. CSAV, Praha, 20: 75-78.

Skypala, J., 1978. K poznani motylu ceskoslovenske casti Cerove vrchoviny. Zpr.

Ceskoslov. Spolec. ent. CSAV, Praha, 14: 115-118.

Stiova, L., 1976. Kvyskytu Libythea celtis na jiznim Slovensku. Ent. Zprav., 6:
107-109.

264

J. Res. Lepid.

Sachl, J., 1980. Prispevek k motyli faune okresu Komarno, pp. 118-128. In
Odborne vysledky zo XVI. tabora ochrancov prirody.

Svestka, M., 1979. K soucasnemu vyskytu ohnivacka Lycaena dispar rutilus
Werneburg 1864. Ent. Zprav., 9: 2-5.

Vacula, D., T. Marcik and Z. Kuchar, 1971. Vihorlat. Ent. Zprav., 1: 11-17.
Vacula, D. and J. Vahala, 1973. Vysledky entomologicke exkurze do okoli Plesivce.
Ent. Zprav. 3: 8-14.

Zelny, J., 1961. Zur Kenntnis von Pieris bryoniae O. aus den Karpaten. Z. Wien,
ent. Gesell., 46: 19-22, 58-64, 100-110.

Zuskinova, T., 1971. Denne motyle Vychodoslovenskeho muzea (Zaverecna
sprava). Kosice, Prir. fak. Univ. P. J. Safarika, 37 pp.

Appendix. Species found in the Pannonian region

Hesperiidae

C. erate

Carcharodus alceae

mo9

C. hyale

09

C.flocciferus

■o#V

C. myrmidone

■o«E

C. lavatherae

■o?

C. palaeno

o4E?

C. palaemon

09

Gonepteryx rhamni

09

Erynnis tages

09

Leptidea morsei

091

Hesperia comma

09

L. sinapis

09

Heteropterus morpheus

U09

Pieris brassicae

09

Ochlodes venatus

09

P. bryoniae

09D

Pyrgus alveus

090

P. daplidice

09

P. armoricanus

■o«V

P. mannii

■o«I

P. carthami

U09

P. napi

09

P. malvae

09

P. rapae

09

P. serratulae

091

Spialia orbifer

U09l

Lycaenidae

S. sertorius

09

Aricia agestis

09

Thymelicus acteon

■0*1

A. allous

■•I

T. flavus

09

A. eumedon

o«4E

T. lineolus

09

Callophrys rubi

09

Celastrina argiolus

09

Papilionidae

Cupido alcetas

■o#I

Iphiclides podalirius

*09

C. argiades

o#

Papilio machaon

*09

C. decoloratus

U09

Parnassius apollo

★o ♦ E?n

C. minimus

09

P mnemosyne

★o«4V

C . osiris

•I?D

Zerynthia polyxena

*mo9 ♦ V

Cyaniris semiargus

09

Glaucopsyche alexis

o»4E

Pieridae

Iolana iolas

09 ?

Anthocharis cardamines

09

Lampides boeticus

o?>

Aporia crataegi

09

Lycaeides argyrognomon

U09

Colias alfacariensis

09

L. idas

■o*I

C. chrysotheme

■o#E

Lycaena alciphron

o»E

C. crocea

L. dispar

■o«4E

29(4):305-315, 1990(92)

315

gratitude to Rick Davis, Rudi and Leona Mattoni for linguistic corrections, and to
O. Kudma and P. S. Wagener for encouraging me to participate at the Wageningen
Congress.

Literature Cited

Hama, E., M. Ishii, and A. Sibatani (Eds.), 1989. Nipponsan tyoorui no suiboo to
hogo (Decline and conservation of butterflies in Japan), Part 1. Nippon Rinsi
Gakkai (Lepidopterological Society of Japan), Osaka. In Japanese.

Moriyama, H., 1988. Sizen o mamoru towa dooiu kotoka (What is the meaning of
nature conservation?). Noobunkyoo, Tokyo. In Japanese.

Sei, K., 1988. Huzisan ni sumenakatta tyoo tati (Butterflies which failed to live on
Mt. Fuji). Tukizi Syokan, Tokyo. In Japanese.

Editor’s note: This paper is a modified version of the presentation at the Interna-
tional Congress “Future of Butterflies in Europe: Strategies for Survival” organized
by the Agricultural University Wageningen and held at the International Agricultural
Centre, Wageningen, Netherlands during 12-1 4. IV. 1989. A portion of the paper
comprised the English summary of Hama, Ishii, and Sibatani (1989) and is printed
with permission of the Lepidopterological Society of Japan.

Journal of Research on the Lepidoptera

29(4):267-276, 1990(92)

Effects of a Microbial Insecticide, Bacillus
thuringiensis kurstaki , on nontarget Lepidoptera
in a Spruce Budworm-infested Forest

Jeffrey C. Miller

Department of Entomology, Oregon State University, Corvallis, Oregon 97331-2907 USA

Abstract. Species in a guild of nontarget leaf-feeding Lepidoptera on
tobacco brush, Ceanothus velutinus Dough were monitored in the field
to assess ecological effects of one application of the microbial pest
control agent, Bacillus thuringiensis Berliner var. kurstaki [BTK]. The
Lepidoptera were sampled to compare species richness, species even-
ness, species diversity, larval abundance, and a dominance index
between an untreated and BTK treated site over a period of two years.

The guild of leaf-feeding Lepidoptera on C. velutinus consisted of 32
species. No statistically significant differences were observed in overall
species richness, although the number of species in the untreated site
was 30% higher two weeks after treatment. However, species richness
among uncommon species was significantly reduced in the treated site.

Also, no statistically significant differences were observed in species
evenness or species diversity but the indices were lower in the untreated
site in three of the four post-treatment samples. A dominace index was
consistently higher in the untreated site. The total number of caterpil-
lars per 100 sec sampling was significantly higher (5.4-fold) in the
untreated site in the early summer sample, two weeks after treatment.

Also, larval abundance in the early summer sample was significantly
higher (3.5-fold) one year later. No differences were noted in larval
abundance in the late summer sample in either year.

Introduction

Microbial pest control agents [MPCAs] are a primary means of biologi-
cal control for insect pests. In general, the use of MPCAs is targeted for
a particular pest species. The insect pathogen, Bacillus thuringiensis
[BT] Berliner var. kurstaki [BTK], is a bacterial MPCA used for suppres-
sion of pest Lepidoptera. For instance, large-scale use of BTK against the
gypsy moth and spruce budworm has been commonly employed over
forested habitats (Brookes et al. 1987, Doane and McManus 1981,
Dreistadt and Dahlston 1989). However, nontarget species that are
taxonomically related to the target pest may also be adversely effected
(Laird 1973, Lighthart et al. 1988, Pimentel et al. 1984, Podgewaite
1986). Miller ( 1990) noted that BTK treatments (three in a single season)
for the gypsy moth in western Oregon reduced species richness and larval
abundance for up to two years within a guild of native, nontarget
Lepidoptera feeding on oak. Franz and Krieg (1967) observed that other
Lepidoptera decreased in number when BT was applied to control Tortrix
viridana L. in Europe.

268

J. Res. Lepid.

Many concerns need to be addressed regarding the use of MPCAs,
particularly in large-scale programs and in the advent of genetically
altered organisms. Among these are: (1) the impact of an MPCA on
nontarget populations that have important functions in food webs; (2) the
fate of species of special interest like the monarch butterfly (Brower
1986); (3) population trends in endangered species; and (4) conflict with
other biological control agents, such as Lepidoptera on weeds (Miller
1990). Also, studies on ecological effects of MPCAs that are not geneti-
cally altered are needed to compare to those ecological effects involving
genetically engineered MPCAs (Kirschbaum 1985, Flanagan 1989, Tiedje
et al. 1989).

The objective of the current study was to determine if the use of BTK
resulted in significant differences in the abundance of individuals and
species composition of immature Lepidoptera between treated and
untreated sites.

Materials and Methods

The data presented here come from an investigation into the effects of a single
BTK treatment targeted for the spruce budworm ( Choristoneura occidentalis
Freeman) on a guild of native, nontarget leaf-feeding Lepidoptera. The study was
conducted from June 1989 to August 1990. The field samples focused on the guild
of immature Lepidoptera (caterpillars) that feed on the foliage of tobacco brush
( Ceanothus velutinus Dougl.). This plant was selected because of its general
abundance within the plant community where the spray zone was located. Also,
earlier studies (author, unpubl. data) indicated that the species richness and
abundance of Lepidoptera on tobacco brush was relatively high in the Oregon
Cascade Mountains.

The study site was 50km (31mi) south of Estacada, Clackamas Co., Oregon.
The area is on the western slope of the Cascade mountain range at 1000- 1200m
(ca. 3000-4000ft) elevation where the mean maximum temperature in July is 24-
28°C (75-8 1°F) and precipitation ranges from 160-200cm (63-79in) per year,
mostly between December and March (Franklin and Dyrness 1988). The plant
community is dominated by Douglas-fir ( Pseudotsuga menziesii (Mirb.) Franco).
Also present were alders (. Alnus spp.), willows ( Salix spp.), maples ( Acer spp.),
and many species of shrubs, such as, blueberry ( Vaccinium spp.), rhododendron
(. Rhododendron macrophyllum G. Don), ocean spray ( Holodiscus discolor (Pursh)
Maxim.), and C. velutinus.

Two transects for sampling caterpillars were established within each treat-
ment. Thus, a total of four transects were established. Each transect was 100m
(330ft) long and 2m (6ft) wide. The transects located in the treated site were
0.5km (0.3mi) within the boundary of a 4000ha (ca. 10,000ac) region that was
sprayed in late June 1989. The area where the transects occurred was treated on
June 26. The BTK was applied from a helicopter at the rate of 8 billion
international units (BIU) per 2.8 liters (3qt) of water per 0.4ha (l.Oac). The
transects in the untreated site were located outside the spray zone by at least 2km
(1.3mi). Because of topographical and floral heterogeneity the transects were
matched between treated and untreated sites using: (1) physical aspects of the
habitat (south facing slope, open canopy, elevation of 1000- 1200m (3000-4000ft);

29(4):267-276, 1990(92)

269

(2) similarly sized plants of C. velutinus that were l-2m (3-6ft) tall; and 3) the
presence of 30-40 plants of C. velutinus within an area of 100x100m (330x330ft).
In 1989, each transect was sampled on June 21 (pre-spray), July 11 (early
summer), and August 25 (late summer). In 1990, each transect was sampled on
July 11 and August 20.

The caterpillar fauna was sampled by shaking foliage over a 7 5x7 5cm (30x30in)
sheet for a timed interval of 30-45 sec. Sampling was continued until a total of
180-220 sec of sampling effort was achieved along each transect. Typically, 4-5
plants along each transect would be sampled at each date to attain the 180-220
sec sampling effort. A particular plant was only sampled once within each year.
Caterpillars were collected alive and reared to adults in the laboratory on clean
tobacco brush foliage to determine viability and verify identifications. The
number of larvae collected per 100 sec of sampling was used to compare
population density among transects and between treatments.

The samples were analyzed to determine: (1) the individual abundance of all
immmature leaf-feeding Lepidoptera; (2) species richness (s); (3) species even-
ness (JO; (4) species diversity {H ’XPielou 1974); and (5) a dominance index (J)
(Berger and Parker 1970, Southwood 1978)(Table 1). Also, a novel method for
classifying and comparing common and uncommon species was developed for this
study. This index was used to measure proportional abundance of individuals to
species richness. A rating of a species as common was given if the proportional
abundance of individuals of that species was equal to or larger than its proportion
of species richness. A species was considered uncommon if the proportional

Table 1. Indices involving species composition among leaf-feeding caterpillars.

Variable Equation

Species evenness
Variance of (J’)
Species diversity (H’)
Variance of (H’)
Dominance index (d)

(J’) = (H’)/lns
var (H’)/(lns)2
(-Ep. lnp.)-[(s-l)/2N]

[(Ep. ln2p.)-(Ep. lnp.)2/N]+[(s-l)/2N2]

N /Nt

max T

Proportional abundance-richness index

Nys

s is species richness; p. is the proportion of the ith species; Nmax is the
number of individuals of the most abundant species; and NT is the total
number of individuals of all species (see Berger and Parker 1970, Pielou
1974, Southwood 1978).

270

J. Res. Lepid.

abundance of individuals of that species was less than its proportion of species
richness. The data on species richness and larval abundance were analysed by
using a G-test for independence (Sokal and Rohlf 1981).

Results and Discussion

A total of 32 species belonging to nine families of Lepidoptera were
collected during the study. The most commonly encountered taxa were in
the Geometridae, Noctuidae, Lycaenidae, and Gelechiidae (Table 2). The
most abundant species were; two geometrids, Drepanulatrix falcataria
(Pack.) and Eupithecia sp.; a lycaenid, Satyrium saepium saepium
(Boisduval); and a gelechiid, Chionides sp.

Larval abundance. A total of 1389 immature Lepidoptera was
collected from all sites during the two year study. Prespray equality in
larval abundance between sites was demonstrated by the recovery of 75.2
and 75.5 larvae per 100 sec of sampling in the untreated and treated sites,
respectively (Table 3).

Table 2. Some of the more common species of Lepidoptera collected on
Ceanothus velutinus in and near an area sprayed with BTK for control of the
spruce budworm, Estacada, Oregon, USA. 1989-1990.

FAMILY

Genus species

Peak abundance*

GELECHIIDAE

Chionides sp.

23.7

GEOMETRIDAE

Drepanulatrix sp.

38.5

(mostly D. falcataria)

Eudrepanulatrix sp.

14.1

Eupithecia sp.

54.0

Hesperumia sulpharia

14.4

Nematocampa sp.

10.8

LYCAENIDAE

Satyrium saepium saepium

18.5

NOCTUIDAE

Orthosia hihisci

0.1

TORTRICIDAE

Choristoneura sp.

12.3

* Number of larvae collected per 100 sec sampling effort with a 75x75cm
beating sheet.

29(4):267-276, 1990(92)

271

Table 3. Effects of BTK on nontarget Lepidoptera on Ceanothus velutinus: Larval
abundance. Estacada, Oregon, USA. 1989-1990.

Date

Site

untreated treated

G

df

P

1989

June 21

75.2

75.5

0.00

1

0.999

July 11

80.1

14.9

49.16

1

0.001

August 25

29.2

29.0

0.00

1

0.999

1990

July 11

29.8

8.6

12.39

1

0.001

August 20

20.9

19.4

0.04

1

0.800

values are number of larvae per 100 sec sampling effort with a 75x75cm
beating sheet.

Differences in the number of larvae collected between the prespray
samples (June 21) and the first postspray samples (July 11) demon-
strated the impact of BTK on nontarget Lepidoptera (Table 3). The
number of larvae per 100 sec sample effort in the untreated site increased
6%, from 75.2 in the prespray sample to 80.1 in the first postspray
sample. This increase in larval abundance was due to natural seasonal
cycles of species in the late spring and early summer. In contrast the
number of larvae in the treated plots decreased 80%, from 75.5 to 14.9
indivduals per 100 sec sampling time. There was an 81% difference in
larval abundance between treated and untreated sites on July 11. The
second postspray sample, on August 25, showed no differences in larval
abundance between sites.

One year later, the early summer sample (July 11) again showed a
significant difference in larval abundance; 71% fewer larvae were found
in the treated site (Table 3). Although the BTK application occurred more
than one year prior to this sample, the effects were still present because
nearly all species involved are univoltine. The second year of the study
was in effect monitoring the first generation from individuals present at
the time of the treatment. The late summer sample one year after
treatment showed no differences in larval abundance between sites.

Species richness. The results on species richness in the prespray
samples indicated the treated and untreated sites were similar. A
comparison of treated and untreated sites indicated that statistically
BTK did not significantly effect the number of species of Lepidoptera on
C. velutinus (Table 4). Although species richness was 30% higher (6 more

272

J. Res. Lepid.

Table 4. Effects of BTK on nontarget Lepidoptera on Ceanothus velutinus:
Species richness (s) and species evenness (J’). Estacada, Oregon, USA. 1989-

1990.

Date/

treatment

s

J ’

var J’

95% Cl J ’

1989 June 21

untreated

15

0.76

0.01

0.64-0.89

treated

15

0.82

0.01

0.70-0.94

1989 July 11

untreated

19

0.67

0.01

0.58-0.76

treated

10

0.92

0.02

0.67-1.00

1989 August 25

untreated

15

0.48

0.01

0.35-0.60

treated

15

0.44

0.01

0.33-0.55

1990 July 11

untreated

10

0.61

0.01

0.52-0.70

treated

12

0.85

0.03

0.52-1.00

1990 August 20

untreated

6

0.52

0.01

0.32-0.72

treated

8

0.59

0.01

0.42-0.75

species) in the untreated site in the first early summer sample, the
difference was not significant (G=1.064, df=l, P>0.25).

Species evenness. The results on species evenness in the prespray
samples indicated the treated and untreated sites were similar (Table 4).
In both years a lower value for species evenness occurred in the late
summer samples. Thus, a seasonal trend in species evenness was
evident. No significant differences in species evenness were observed
between treatments within respective sample dates. However, in both
early summer samples species evenness was consistently higher in the
treated sites. Higher values for species evennness can be interpreted to
indicate a decreased degree of numerical dominance by any given
species.

Dominance index. An index indicating the degree of numerical
dominance by the most abundant species was higher in untreated sites
in all samples (Table 5). A difference of 13% was detected in the prespray
samples compared to a range of 45 and 54% between treatments in the
early summer samples. The late summer samples exhibited a difference
of 8 and 10%. These data suggest that the samples from untreated plots
tended to be dominated by certain species. The application of BTK had

29(4):267-276, 1990(92)

273

Table 5. Effects of BTK on nontarget Lepidoptera on Ceanothus velutinus:
Dominance index (d) and species diversity ( H ). Estacada, Oregon, USA. 1989-

1990.

Date/

treatment

d

H’

var H’

95% Cl H’

1989 June 21

untreated

0.30

2.07

0.03

1.72-2.41

treated

0.26

2.22

0.03

1.88-2.57

1989 July 11

untreated

0.35

1.96

0.02

1.68-2.25

treated

0.16

2.12

0.10

1.49-2.75

1989 August 25

untreated

0.63

1.29

0.03

0.98-1.61

treated

0.58

1.19

0.02

0.91-1.47

1990 July 11

untreated

0.51

1.41

0.01

1.21-1.61

treated

0.28

1.96

0.17

1.16-2.76

1990 August 20

untreated

0.71

0.93

0.03

0.58-1.28

treated

0.64

1.22

0.05

1.00-1.44

the effect of evenning the numbers of caterpillars among species, an
observation consistent with the indices on species evenness.

Species diversity. The results on species diversity in the prespray
samples indicated the treated and untreated sites were similar (Table 5).
In both years, species diversity was lower in the late summer samples
regardless of treatment. Thus, as with species evenness, a seasonal trend
in species diversity was evident. No significant differences in species
diversity were observed between treatments in respective sample dates.
However, in both early summer samples the index for species diversity
was lower in the untreated sites. A lower value for species diversity can
be interpreted to indicate a reduction in either species richness, species
evenness, or both.

Proportional abundance-richness index. The abundance of indi-
viduals in respective species expressed as a proportion relative to the
number of species can be used as an index to determine whether a species
may be considered uncommon or common (Table 6). This is useful
because indices of species richness and species diversity do not provide
a means for objectively classifying the relative abundance of individuals
among species as common or uncommon. For example, a species repre-

274

J. Res . Lepid.

Table 6. Effects of BTK on nontarget Lepidoptera on Ceanothus velutinus:
number and proportion of common and uncommon species. Estacada, Oregon,

USA. 1989-1990.

Date

No.

Common

species

Uncommon

Proportion of species
Common Uncommon

1989 June 21

untreated

5

10

33

07

treated

0

9

40

00

1989 July 11

untreated

4

15

21

79

treated

3

7

30

70

1989 August 25

untreated

2

13

13

87

treated

2

13

13

87

1990 July 11

untreated

2

8

20

80

treated

4

8

33

07

1990 August 25

untreated

1

5

17

83

treated

1

7

13

87

sented by one individual in a sample may carry the same weight as a
species represented by 50 individuals.

The number of common and uncommon species in the prespray samples
was similar (Table 6). The samples were dominated (73% overall) by
species considered to be uncommon. Overall, only 27% of the species were
categorized as common. In the early summer samples the number of
common species was equivalent but the number of uncommon species
was reduced by 53% in the treated site, a significant difference (G=10.66,
df=l, PcO.OOl) . This effect was not observed in the early summer sample
in the second year. These data suggest that certain uncommon species
are more likely to be removed from the system than common species and
that gross accounts of species richness may mask significant effects on
less abundant species. The importance of this observation underscores
the need to direct special management needs for rare and endangered
species, such as: timing of application, doses, skipping sensitive areas,
and mitigation.

29(4):267-276, 1990(92)

275

Conclusions

The data on larval abundance and richness of uncommon species
indicated that one application of BTK reduced the abundance of nontar-
get Lepidoptera in the guild of caterpillars feeding on leaves of C.
velutinus. The effects were most notable during the early summer period
immediately following treatment. The data regarding species richness,
species evenness, and species diversity did not demonstrate a significant
effect. However, a statistical analysis of the indices used to describe
community composition can be misleading if a biological interpretation
is not assessed as well. For instance, species richness declined in the first
year of the study following the application of BTK. The decline was not
significant but nonetheless six species occurred in the untreated site that
were not represented in the treated site. An extrapolation of these data
suggests that if any of the species had been limited in its distribution, or
a unique genotype of the species was locally endemic, then the popula-
tion/species would be at high risk of becoming extinct.

Additional studies on the impact of MPCAs on nontarget organisms
under field conditions are needed. Many variables contribute to patterns
in the data from samples of treated and untreated sites. A comparison of
this study to a similar study conducted previously (Miller 1990) shows
some of the variables that may influence the impact of a MPCA on
nontarget organisms. The dose, frequency of applications, timing of
applications, and geometry of the treated site will contribute a strong
influence on the immediate impact and overall recovery of a community
subjected to a MPCA treatment.

In the present study and that of Miller (1990), BTK was shown to
negatively impact certain nontarget organisms. Although no direct
comparisons were made with the use of synthetic pesticides, the negative
impacts on nontarget organisms would likely be more severe with
synthetic pesticides and widespread across a multitude of taxa. Many
synthetic pesticides have been shown to exhibit dramatic negative
impacts on nontarget organisms (Croft 1990; Ehler and Endicott 1984;
Martinat et al. 1988). While the use of microbial pathogens may exhibit
certain negative impacts on nontarget organisms, methods for the use of
MPCAs can be developed to minimize such impacts and reduce the use
of synthetic pesticides in present pest management programs. The
development of pathogens as biologically rational pest control agents
must involve assessments of nontarget effects for the benefit of prolong-
ing the use of microbes as ecologically and sociologically acceptable pest
management options.

Acknowledgments. Field sites were located with the assistance of Art Weber and
Beth Wilhite, USFS, Portland, Oregon. Funds for this research were provided in
part by U.S. Environmental Protection Agency, Willamette Institute for Biological
Control, and Oregon State University Agricultural Experiment Station. Field
assistance was provided by B. Scaccia and J.E. Miller. Taxonomic assistance was

276

J. Res. Lepid.

provided by P. Hammond, B. Scaccia, J. Powell, and D. Ferguson. This is Oregon

State University Agricultural Experiment Station Technical Paper No. 9762.

Literature Cited

Berger, W. H. & F.L. Parker. 1970. Diversity of planktonic Foraminifera in deep
sea sediments. Science 168:1345-1347.

Brookes, M.H., Stark, R.W. & R.W. Campbell (eds.) 1978. The Douglas-fir tussock
moth: A synthesis. U.S.D.A. For. Serv., Science and Education Agency Tech.
Bull. 1585.

Brower, L.P. 1986. Commentary: The potential impact of Dipel spraying on the
monarch butterfly overwintering phenomenon. Atala 14:17-19.

Croft, B.A. 1990. Arthropod Biological Control Agents and Pesticides. Wiley and
Sons, New York, New York, USA.

Doane, C.C. & M.L. McManus (eds.) 1981. The gypsy moth: Research Toward
Integrated Pest Management. U.S.D.A. Tech. Bull. 1584, Wash. D.C.

Dreistadt, S.H. & D.L. Dahlston. 1989. Gypsy moth eradication in Pacific coast
states: History and evaluation. Bull. Entomol. Soc. Amer. 35:13-19.

Ehler, L.E. & P. Eendicott. 1984. Effect of malathion-bait sprays on biological
control of insect pests of olive, citrus, and walnut. Hilgardia 52(5) 1-47.

Flanagan, P.W. 1989. The need for basic research on genetically engineered
microorganisms. Bull. Ecol. Soc. Amer. 70:14-19.

Franklin, J.F., & C.T. Dyrness. 1988. Natural Vegetation ofOregonand Washington.
U.S.D.A. For. Serv. Gen. Tech. Rep. PNW-8.

Kirschbaum, J.B. 1985. Potential implications of genetically engineered and other
biotechnologies to insect control. Ann. Rev. Entomol. 30:51-70.

Laird, M. 1973. Environmental impact of insect control by microorganisms. Ann.
N.Y. Acad. Sci. 217:218-226.

Lighthart, B., Sewell, D. & D.R. Thomas. 1988. Effect of several stress factors on
the susceptibility of the predatory mite, Metaseiulus occidentalism to the weak
bacterial pathogen Serratia marcescens. J. Invert. Path. 52:3-42.

Martinat, P.J., Coffmann, C.C., Dodge, K, Cooper, R.J. & R.C. Whitmore. 1988.
Effect of diflubenzuron on the canopy arthropod community in a central
Appalachian forest. J. Econ. Entomol. 81:261-267.

Miller, J.C. 1990. Field assessment of the effects of a microbial pest control agent
on nontarget Lepidoptera. Amer. Entomol. 36:135-139.

Pielou, E.C. 1974. Population and Community Ecology: Principles and Methods.
Gordon and Breach, New York, New York, USA.

Pimentel, D., Glenister, C., Fast, S., & D. Gallahan. 1984. Environmental risks of
biological pest control. Oikos 42:283-290.

Podgwaite, J.D. 1986. Effects of insect pathogens on the environment. Fortschr. der
Zool. 32:279-287.

Sokal, R.R. & F.J. Rohlf. 1981. Biometry. W. H. Freeman and Co., San Francisco,
California, USA.

Southwood, T.R.E. 1978. Ecological Methods. John Wiley and Sons, Inc., New York,
New York, USA.

Tiedje, J.M., Colwell, R.K., Grossman, Y.L, Hodson, R.E., Lenski, R.E., Mack, R.N.
& P.J. Regal. 1989. The planned introduction of genetically engineered
organisms: Ecological considerations and recommendations. Ecology 70:298-
315.

Journal of Research on the Lepidoptera

29(4):277-304, 1990(92)

The Endangered El Segundo Blue Butterfly

Rudolf H. T. Mattoni

9620 Heather Road, Beverly Hills, CA 90210

Abstract. Conflict concerning land use of the 302 acre sand dunes
parcel at the western boundary of the Los Angeles International
Airport (LAX) centers on the small butterfly, The El Segundo blue
butterfly (ESB) . Since the ESB was granted protected status in 1976
under the Endangered Species Act of 1973, the real issues involved in
its conservation have been obscured by the polemics of special interests
groups that have been arguing without proper data, or worse, with
flawed data. This paper reviews all known aspects of the history,
biology, and conservation issues, concluding with up-to-date political
actions that will affect the survival of the species.

Historical Perspective
SYSTEMATICS

The El Segundo Blue butterfly (. Euphilotes bernardino allyni) is one of
four subspecies of a polytypic species which belongs to the E. battoides
species complex of many population aggregates not yet clearly defined
systematically (Shields and Reveal, 1987; Mattoni, 1989; Pratt, unpub.).
Although the battoides complex occurs over all of North America west of
the Great Plains, from British Columbia to Baja California, E. bernardino
is distributed in southern California, southern Nevada, Arizona, and
northern Mexico, including Cedros Island.

The ESB was formally described by Shields (1975) from specimens
collected in El Segundo. These specimens were taken at the Chevron
Refinery site. Several experts recognized the ESB as distinct prior to its
formal description, including Emmel and Emmel (1973), who illustrated
it and called attention to its potential extinction. The ESB is distin-
guished from all other subspecies by a combination of underside black
spot size, amount of orange on the wings, wingspread, foodplant, and
other characters (Table 1 in Mattoni, 1989).

NATURAL HISTORY

As with all species in the genus Euphilotes , the ESB spends virtually
its entire life cycle in intimate association with the flowerheads of some
species of buckwheat; in this case the coastal buckwheat, Eriogonum
parvifolium.. The almost total involvement of all stages with a single
plant part is unique among North American butterflies. Adults find one
another to mate, usually nectar, lay eggs, perch, and in most cases
probably die, on flowerheads. Thus a significant array of population
regulating mechanisms operate within the flower he ad environment, for
example predation, parasitism, competition, nutrition and disease.

278

J. Res. Lepid.

When the time arrives to pupate, however, larvae either drop or crawl to
the ground and burrow into the soil. Typically they travel at least two
inches below grade, but stay within the root and debris zone where they
are protected from desiccation and insulated from temperature ex-
tremes. Factors affecting pupae are almost totally unknown even though
the species spends 90% of its life span in the pupal stage (see White, 1988).

The ESB has one generation per year, as is obligatory for all members
of the E. battoides complex under natural conditions. Adults fly from
mid-June through the end of August, the exact timing depending on
weather. Usually the flight lasts from mid- June to mid-August. The
onset of flight is closely synchronized to the beginning of the flowering
cycle of the foodplant. Fresh females fly to flowerheads upon emerging
from their pupae. There they are found and mated within hours by one
of the male population that is constantly moving from flowerhead to
flowerhead. The females then immediately begin laying eggs. Labora-
tory data indicate females produce 15-20 eggs per day, but must continu-
ously nectar to maintain egg production (Mattoni, unpub.). Although
field data indicate females at Chevron live an average of four days in
nature (Arnold, 1983), in captivity females live two weeks and produce up
to 120 eggs (Mattoni, unpub.). Eggs hatch within five to seven days.
Larvae undergo four instars to complete growth, a process taking from 18-
25 days. The larvae develop honey glands by the third instar, and are
thereafter usually tended by the ants Iridiomyrmex humilis or
Conomyrmex sp., which may protect them from parasitoids and small
predators. By late August the flowerheads have generally senesced and
the larvae have all pupated underground. The natural history data are
all reported from the population at LAX.

Mature larvae are highly polymorphic, varying from almost pure white
or pure dull yellow to strikingly marked individuals with a dull red-to-
maroon background broken by a series of yellow or white dashes or
chevrons. They feed in such a manner as to remain concealed by the
flowerhead, their patterning adding to this crypsis. The preferred part
of the flowerhead is young seeds, which are consumed preferentially to
other flowerparts. The latter are loosely webbed together producing the
illusion of an intact flowerhead. One larva requires two-to-three
flowerheads (which equals 10-15 involucres or 400-500 flowers or their
seeds) to complete development. The discrepancy between longevity of
adults in the field (2.3 to 7.3 days, Arnold, 1983) and lab (average 16 days,
Mattoni, unpub.) adults is most likely due to predation by lynx and crab
spiders. These spiders were found at a frequency of about one per 200
flowerheads in 1987 (Mattoni, unpub.). When 15 man hours of direct
observation of flowerheads was made, only one capture of a male ESB by
lynx spider was seen. The event was rapid, by seizure, with the prey
rapidly imbibed and discarded.

The egg population is chiefly regulated by Trichogramma sp. nr.
minutum , which also attacks the eggs of the common hairstreak and at

29(4):277-304, 1990(92)

279

least two species of microlepidoptera on the flowerheads. Pratt (1987)
found 9% of 147 eggs of the common hairstreak collected at the LAX dunes
in 1985 parasitized by the wasp. ESB eggs probably have a similar
frequency, but appear better placed for concealment in comparison to the
hairstreak, so few can be found in nature for testing.

In a sample of 30 mature ESB larvae recovered from flowerheads in
1987, six, or 20%, were parasitized by a braconid wasp, Apanteles
thurberiae. The same wasp also attacked the common hairstreak, the
moth Lorita scarified (=a6ornana)(Cochylidae), and Aroga sp.
(Gelechiidae), the latter two common microlepidoptera on the flowerheads.
The ichneumonid wasp, Diadegma sp., was found in the hairstreak and
both moth populations in 1987, but not in the ESB (Mattoni, unpub.).
Pratt (1987) reported the same pattern in his 1985 survey. The most
significant feature of both wasp parasitoids is absence of diapause. Thus
they only persist by living on a continuum of alternate larval hosts over
an annual cycle. The same is true for the Trichogramma egg parasitoid.

Arnold (1983) reported finding pupae parasitized by two unidentified
species of tachinid fly at Chevron. No quantitative data were given. The
tachinid life cycle coincides with that of its ESB host so alternate hosts
are not necessary for its persistence. A set of 28 pupae screened from sand
at LAX, just prior to the 1988 flight, produced no parasites.

Pratt (1987) found larvae of Aroga sp. and Lorita scarified predominant
in E. parvi folium flowerheads, up to 50 each, in 1985. He hypothesized
that these severely reduced the food available to the ESB larvae, but also
had an impact on the ESB by direct predation and indirect harboring of
shared parasitoids. Mattoni (1988) found that a sample of flowerheads
collected in 1987 produced 30-50% viable seed sets in spite of herbivory
from all sources.

HISTORICAL RANGE

Distribution of the ESB is dependent on the occurrance of its foodplant,
the coast buckwheat. The butterfly further appears limited to habitats
with high sand content. These sites historically consisted of the El
Segundo sand dunes, including interrupted extensions to the north into
what is present-day Ocean Park, and southerly to Malaga Cove in Palos
Verdes (Figure 1). However, after a gap at Santa Monica, the foodplant
extends further north along the well-drained, low altitude, steep sand-
stone slopes of the coastal Santa Monica Mountains and to the south on
similar formations along the Palos Verdes peninsula. The plant forms a
small colony on the sand dune at Point Dume. Variant ecotypes of
Euphilotes bernardino bernardino are associated with mixed Eriogonum
parvifolium and E. cinereum plant populations along the sea bluffs of
Palos Verdes south to San Pedro. The plant populations become pure
stands of E. cinereum above the bluff face in Palos Verdes. Both E.
cinereum and E. fasciculatum, are toxic to El Segundo dunes ESB larvae,
although ESB females will lay eggs on them in free choice experiments
(Mattoni, unpub.).

280

J. Res. Lepid.

Figure 1 . Distribution of both the El
Segundo sand dunes and
El Segundo blue butterfly,
historically and at present
(1991). Extent of the his-
torical dunes and butterfly
populations distribution
are shown by the stippled
areas. Extant butterfly and
undisturbed dunes frag-
ments indicated by the
black dots or black areas:
LAX dunes, Chevron Pre-
serve, and Malaga Cove.
Potential restorable habi-
tat is indicated: Hyperion-
DWP, and Playa del Rey
dunes. The Ballona la-
goon fragment is included
in the latter site.

The active El Segundo sand dunes historically covered about 4.5 square
miles (1295 hectares, 3200 acres), based on data from the 1894 geologic
survey (Figure 1) and Cooper (1967). An inaccurate figure of 36 square
miles, sometimes miscalculated as 18,000 hectares, has been widely
quoted for the dunes area from the summary of California sand dunes by
Cooper (1967). The misquoted value included pre-Flandrian sand depos-
its that formed yet older dunes now having more or less consolidated to
form sandstone. Cooper describes the situation clearly in his figure 2 and
text, but somehow the detail was overlooked and the original misinter-
pretation repeated by subsequent authors. This sandstone forms the
underlay or base to the present dunes. The edaphic properties of these
sandstones do not provide proper adaptive conditions for the indicator
sand obligate plants of the dunes community, particularly the coast
buckwheat. At El Segundo, coast buckwheat is a key indicator of
primary, undisturbed coastal sand dunes sites. A diagrammatic cross
section of the El Segundo dune system is given in Figure 2.

The biological community of sand dunes proper is adapted to continu-
ously moving sand and extreme aridity. Once sand is permanently
stabilized, community composition changes. Sand obligate plant species
decrease in frequency to the benefit of more widespread species and
weeds with overall cover increasing. Animal community composition is
likely affected as well.

29(4):277"304, 1990(92)

281

Prevailing Westerly Winds — »

Figure 2. Diagrammatic cross section of typical El Segundo dune in the high dune
area. Major physical features denoted.

The dunes and lee deflation plain were undisturbed until rancho
development in the 1840’s. Farming was then established on the coastal
prairie to the east of the dunes proper, but generally started at least a half
mile further inland, probably because of agricultural unsuitability of the
poorly drained sandstone soil near the backdune.The dunes themselves
were undisturbed until the late 1880’s when the cities from Redondo
Reach to Venice were established, but development was limited. Prior to
that time virtually the entire dunes area was pristine, without evidence
of disturbance. Redondo Beach separated the main dunes from south
Redondo Beach and the Malaga Cove extensions (Figure 1), with devel-
opment of Venice eliminating the dunes north of the mouth of Ballona
Creek. Conversion of the central part of the dunes was slower. Construc-
tion of the Chevron refinery in 1911 separated the dunes into two frag-
ments. The southern fragment was gradually converted to residences
starting at the turn of the century and rapidly accelerating in the late
1940s. Habitat values were totally destroyed by the 1970s. In the 1950s
Henne (pers. comm.) noted a dense ESB population and other rare
Lepidoptera in Hermosa Valley, an area soon after destroyed. In 1928 the
grid of streets on the LAX dunes were constructed, but development was
minimal following the 1929 crash. It was not until after World War II that
explosive development occurred, with virtually the entire dunes built
upon between 1946 and 1965, where almost all the land was privately
owned. Construction of the Hyperion wastewater and Scattergood
generating plants in the 1940s, along with dense housing on the present
LAX dunes, reduced the northern fragment to about 80 acres of dunes
habitat by 1960. The 1.6 acre Chevron butterfly sanctuary site was
isolated by residential development in the 1950s.

282

J. Res. Lepid.

The most important events affecting the very recent biological history
of the LAX dune segment was the purchase and clearing of residences
from nearly 200 acres between 1966 and 1972, construction of the VOR,
and the excavation and re-contouring of about 70% of the backdune in
order to re-align Pershing Drive in 1975. The newly recontoured dunes
were stabilized by hydromulch with a presumptive "natural” seed mix
and irrigated with a sprinkler system. Unfortunately, the seeds were
representative of coastal sage and not dune scrub plant community. The
common buckwheat, Eriogonum fasciculatum , was introduced by this
activity. At the same time the foredune to the south and west of the VOR
was graded along with the last coastal prairie fragment between the
backdune and Pershing Drive. From 1966 through the late 1970s the
natural biota of the LAX dune suffered its major contraction, with about
40 variably undisturbed acres as refugia for subsequent colonization.
The 1990 landform classification resulting from these events is mapped
in Figure 3.

The most drastic change was the complete destruction of the Los
Angeles coastal prairie to the lee of the dunes. This community was a
Stipa grassland, with a rich occurrence of herbaceous meadow plants and
innumerable intermittent vernal pools. The community is now com-
pletely extinct.

Extrapolating from an estimated carrying capacity of 1,000 ESB per
acre on the backdune and 100 ESB per acre on the foredune (Mattoni,
unpub.) the historic El Segundo dunes system, with about 400 acres of
backdune and a 2,800 acre foredune, should have had an average total
population of about 750,000 adults per year.

PRESENT DISTRIBUTION

The ESB is now (1991) restricted to three locations: the LAX dunes
(LAX), the Chevron Refinery dunes (CHEV), and Malaga Cove (MC)
(Figure 1, Table 1). In 1988 LAX had the largest population both in terms
of area (ca. 4 acres of moderate density and 20 acres of low density
populations), number of adult ESB (ca. 2000), and foodplants (1,114
native plants with 206,045 flowerheads) (Figure 4, Table 2). After the
initial habitat restoration program, the foodplant population was in-
creased to 3358 in 1991 (Figure 4) with an estimated 5000 ESB flying in
1990 (Table 2). CHEV in 1986 had an area of 1.6 acres, about 400 adult
ESB, and about 240 natural plants plus about another 1,000 surviving
introduced cultivated seedlings (Arnold 1986). The MC location, discov-
ered in 1983 by J. Morton and T. Leigh, covers about one acre, had a one-
day population count of about 60, and less than 50 plants with 30,000
flowerheads in 1984. It is heavily overgrown with ice plant and seriously
eroded. Since 1986 the site has been fenced. A 1990 survey (R. Rogers,
pers. comm.) indicates the 1984 status remains stable. Fifteen plants
(3,000 flowerheads) survived to 1986 on a small dune fragment at Playa
del Rey (proposed Ballona Wetlands Reserve). By 1989 half of these

29(4):277-304, 1990(92)

283

O

o

to

o

II

//

Figure 3. Recent history of land use at LAX El Segundo sand dunes. The land use pattern correlates with the degree of disturbance to the natural
biota and hence determines habitat value. The boundary of the preserve area proposed in 1991 is indicated by the heavy line.

284

J. Res. Lepid.

Table 1 . Land area of all El Segundo sand dune fragments remaining as open
space. Both actual habitat with some natural values and parcels that have the
potential for being restored are shown. Malaga Cove and Playa del Rey were not
connected to the original sand dune mass, but have many shared species. The
25 acre non-sand dune portion of the LAX site is not included. Values are in

hectares (acres).

Location Backdune Foredune Meadow Total

ACTUAL HABITATS

LAX

Pristine

Relatively

0.8

(2.0)

0.0

0.0

0.8

(2.0)

Undisturbed

2.6

(7.0)

13.6

(34.0)

0.0

16.2

(41.0)

Total

3.4

(9.0)

13.6

(34.0)

0.0

17.0

(43.0)

Chevron Preserve

0.5

(1.3)

0.1

(0.2)

0.0

0.6

(1.5)

Malaga Cove

0.0

0.5

(1.2)

0.0

0.5

(1.2)

POTENTIAL HABITATS

LAX

Disturbed Sites

7.3

(18.0)

78.0

(192.0)

9.7 (24.0)

95.0

(234.0)

DWP

2.0

(5.0)

15.0

(37.5)

5.0 (12.5)

22.0

(55.0)

Hyperion

0.0

16.4

(41.0)

0.0

16.4

(41.0)

Playa del Rey

2.5

(2.5)

0.0

0.0

2.5

(6.2)

Ballona Lagoon

0.0

2.8

(7.0)

0.0

2.8

(7.0)

TOTAL

11.8

(39.5)

126.4

(313.0)

14.7 (40.2)

156.8

(389.0)

buckwheats died. A male ESB was reported at the locality in 1985, but the
specimen was not taken and no stages have been collected since, in spite
of several attempts. A transient ESB was reported on the DWP right-of-
way between CHEV and LAX (R. Rogers, pers. comm.).

The ESB is isolated from its closest relative, a Palos Verdes ecotype of
Euphilotes bernardino bernardino, by only 2 km at Palos Verdes. To the
north E. bernardino bernardino occurs throughout the Santa Monica
Mountains, where it uses three Eriogonum foodplants: fasciculatum,
cinereum, and parvifolium.. E. bernardino bernardino ranges along the
coast to the south, from the bluffs of the Laguna Hills, feeding on the same

29(4):277-304, 1990(92)

285

(above) and 1991 following interim restoration (below).

286

J. Res. Lepid.

Figure 5. Photographs taken in the same position facing northwest from the corner
of Imperial and Pershing, 1938 (above) and 1988 (below). The early
photograph by W. D. Pierce shows Mrs. Pierce in the forground.

three foodplants. Some populations found on the immediate coast, as at
Point Loma, strongly resemble the ESB in appearance. This pattern is
in all likelihood a convergence and does not represent monophyly with
the ESB (Mattoni, 1989). There are no blues at Point Dume, which has a
small sand dune and a coastal buckwheat population.

There are no other sites remaining in the historical dunes habitat that
can support more than a few random plants. No further ESB populations
remain to be found, although two locations, the Ballona dunes and the
DWP backdune behind the Hyperion Plant, could be restored to support
viable ESB populations.

29(4):277-304, 1990(92)

287

Table 2. Relative density of adult El Segundo blue butterflies over six years (1984,
1986-1990) along transect counts at five major clusters of Eriogonum parvifolium
foodplant Number of Eriogonum parvifolium from 1986 census: the numerator is the
number of plants and flowerheads actually counted for butterflies in each cluster, the
denominator the total numbers in the entire cluster (See Figure 3).

Cluster Number of El Segundo Blue

E. parvifolium , 1986 Transect Counts

plants

flowerheads

1984

1986

1987

1988

1989

1990

Number of days sampled

4

5

9

10

11

10

Span of days sampled

19

35

56

61

54

63

Date of first flight

?

7-9

6-24

6-22

6-23

6-29

Date of last observation

8-8

8-13

8-19

8-22

8-16

8-24

1. Backdune

30

33.600

131

68

207

508

744

648

119

55,275

2. Backdune

43

15.300

60

109

187

344

466

427

(pristine)

154

49,700

3. Crest

35

1,500

2

2

0

3

9

8

91

3,185

4. Backdune VOR

20

4750

NC

70

46

159

130

78

86

8,280

5. Foredune

65

4940

NC

9

33

35

41

31

839

89,605

TOTAL

193

60.090

193

258

473

1049

1390

1192

1289

206,045

Preservation Status: Condition of the habitat

The probability of survival of a given species over time is a function of
habitat quality and habitat size relative to a minimum critical area
(Gilpin and Soule, 1986). The spatial loss of the El Segundo sand dunes
habitat to urbanization reached its maximum in the 1970s. The other
dune habitats, including restorable sites, cannot be further developed at
this time because of legal constraints, safety, or geological hazard.
Together these sites are sufficient to maintain El Segundo Blue popula-
tions indefinitely, given habitat quality. This viability analysis is based
on potential habitat area and topography that could support annual
adult populations in the order of 100,000 individuals. However, high
quality habitat values do not now exist, nor is there assurance that
quality can be maintained over time without some management. The El

288

J. Res. Lepid.

Segundo Blue occurs, or could occur, across several governmental juris-
dictions and on private land. Each situation presents unique problems.

In addition to the three present sites of the El Segundo Blue at LAX,
Chevron, and Malaga Cove; at least three additional sites could be
restored to a native dune ecosystem, thereby providing further assurance
of long term survival (Table 1). These are 1) The Los Angeles Department
of Water and Power (DWP) right-of-way for a power line between the
Scattergood generating plant and Imperial highway at the south end of
the LAX dunes. The site consists of 55 acres of seriously degraded dune
and coastal prairie habitat, including over 5 acres of potentially rich
backdune. Portions of the dune crest and foredune remnant at the
adjacent Hyperion wastewater treatment plant, presently landscaped in
exotic vegetation, could be included as restorable contiguous dune
habitat with an additional 41 acres. Hyperion-DWP would have high
value as a habitat corridor between the LAX and Chevron communities
in addition to functioning as an independent habitat unit. 2) The
approximate 7 acre Playa del Rey backdune, forming the west end of the
proposed Ballona wetlands preserve. 3) Approximately 5 acres of bank
lining the Ballona lagoon that are now (1991) being re-vegetated. A
proposal to re-introduce the ESB onto the first half acre fragment during
1992 is under consideration. Other potential habitats include Dune park
in the city of Manhattan Beach with a 2 acre degenerate backdune
fragment about 2 km south of Chevron. Although completely open to
public use, parts could be protected and restored. Public school open
space of about 1.5 acres each in Hermosa Beach and Manhattan Beach
are potential restoration sites.

Without an active restoration and management program, the outlook
for the long term persistence of the dunes ecosystems necessary to
support the El Segundo Blue is bleak indeed. The centerpiece of any
effort must be the LAX site, as LAX alone contains not only the largest
fragment, but the closest approximation to prehistoric dune ecosystem
composition. Of the 302 acres, about 250 acres are actual sand dunes, of
which 39 contiguous acres were at least partially undisturbed, including
an almost pristine 2 acres of backdune and 15 acres of foredune. The
remainder has either been extensively sandmined, graded, or built upon;
has heavy soil spoils; or is concrete or asphalt road (Figure 3). Further
trauma included heavy spraying of some sites with oil and introduction
of deleterious exotic plants for sand stabilization, and the invasion of
non-native animals (Mattoni, 1990 a, b).

Notwithstanding a gradual degradation between the 1938-1939 bio-
logical survey by Pierce (1939-1940) and the present, most substantive
changes have taken place during the past decade and a half (Mattoni
1990 a, b) as a result of the re-alignment of Pershing Drive, construction
of Imperial Highway, moving sand to build the VOR hill, and fragmen-
tation and scraping of the coastal prairie. The manifestation of degrada-
tion was extirpation of many native species on the one hand, and the

29(4):277-304, 1990(92)

289

invasion of the site by exotic plants and animals on the other. The
convergence of these forces predict a grim future. Of 20 native terrestrial
mammals recorded by von Bloeker (in Pierce notes)., most of which were
present in 1975 (LAX-EIR, 1975), only three are extant today. In their
place are introduced Norwegian rats, red fox and opossum. Of 31 species
of butterfly breeding on the site, 7 or 23% have disappeared. Of 18
species of reptiles and amphibians (von Bloeker, 1941), 7 or 39% no longer
occur and all 5 scrub dependent birds (Soule et al. 1988) have disap-
peared. The situation with native plants is fundamentally more serious
since plants are the base of food chains and hence ecosystems. A
specialized herbivore restricted to one plant species would be lost with
extirpation of its food source, as in the case of two extirpated butterflies.
Of the 73 native plant species recorded on the sand dunes proper by
Pierce et al. 22, or 30%, were not found by our 1989 survey, and 19 of the
51 surviving species occurred as less than 100 individuals and faced
imminent loss. More seriously, several alien plants, including two Acacia
species and Eriogonum fasciculatum , had been introduced to the site
within the past two decades with serious consequences. Other exotic
species present in 1938 have since become serious competitors to the
native plant community. Two closely related iceplant species, insignifi-
cant in 1938, are now dominant across most parts of the foredune thereby
closing options on their occupied territory. Storksbills and alien grasses
are also co-opting habitat for native plants. Recently observed Pampas
grass and Myoporum present yet new threats. The LAX dunes history is
an outstanding example of the effects of uncontrolled, often intentional,
introduction of exotic plants which become ecologically devastating.
Photographs of identical portions of the backdune show the changes in
the landscape between 1938 and 1989 (Figure 5).

The Eriogonum parvi folium foodplant of the ESB is only found on
undisturbed sand dune habitat. In the fifteen years the LAX dunes have
lain fallow since the massive clearing and excavating activities, only
three individual plants of 1114 counted in 1988 were found on disturbed
sites (Figure 4). Why the plant remains restricted is unknown, whether
the limiting factor is poor seed dispersal, seed viability and germination,
or establishment. Thousands of seedlings have been observed across
several sites following winter rains, yet few establish. It is noteworthy
that coastal buckwheat has both highest quality and highest density on
the least disturbed backdune sites (Table 2 and Figure 4: clusters of 119,
154, & 86 plants) where invasive exotics are in low frequency. Within
these sites, and the relatively undisturbed VOR sections (Figure. 3),
native dune scrub plant associations resist significant invasion by
exotics. From photographs and notes of Pierce (1938-1940), many of the
exotic plant species were present in his time, but were not abundant. The
two Acacia species appear exempt from this “exotic exclusion/native
cohesion"’ phenomenon, however, and threaten a breakdown of the
remaining native ecosystem by altering soil texture and chemistry.

290

J, Res . Lepid.

Acacia populations increased from none detectable in 1976 to 671 indi-
viduals in 1987 and was increasing at an annual rate of 46% prior to
removal between 1988-1991. The long lived seed bank of Acacia numbers
in the millions and will require attention for decades.

Ecosystem disturbance through changes in the mammalian commu-
nity have been profound as both rabbits and mice influence the differen-
tial reproductive efficiency of herbaceous plant species. The absence of
mammalian foraging probably relaxed substantial pressure on seed
banks, particularly the r-strategist European weeds. Loss of small
mammals is linked to the introduction of the European red fox, which
themselves now have such limited food resources that they are driven to
garbage scavenging, lizards, and even large insects.

Of the other habitats, Malaga Cove was last visited during the flight
period of 1990. The site was heavily overgrown with iceplant, which
threatens the some fifty buckwheat foodplants, although this site and
population have probably persisted in isolation over the past 40 years.
The site needs further evaluation particularly because the ESB popula-
tion has survived in spite of its very small population. According to Pratt
(pers. comm. ) the Malaga Cove population is genetically identical (allozyme
frequency) to that of LAX.

The Chevron site has been isolated at least since the mid 1950s,
subsequently carrying an ESB population of about 2000 adults annually
from at least 1965 until 1977, when intensive studies on adult population
demography began. While determining population sizes for 1977 and
1979, a precipitous drop from an estimated 1328 to 681 between those two
years was noted (Arnold 1983). Later Arnold (1986) presented data on
densities of both foodplant buckwheat and ESB for each year since 1977,
showing a decline between 1977 and 1986 from about 1850 to 350 adult
ESB adults and 420 to 300 Eriogonum parvifolium, Arnold claimed the
primary cause for the declines was stabilization of the dune process of
sand movement with correlated exotic plant growth, buckwheat senes-
cence, and reduced buckwheat seedling survival, A more likely cause of
crashes of both buckwheat and blue was the impact of the study itself. In
small fragile habitats, particularly steep sand dunes, regular walking
during the course of observation and data collection is a serious problem
(Brown, 1987) that leads to root damage and changing water balances.
Another study impact suggested is the citation by Arnold (1983, p. 90) of
rearing 839 El Segundo Blue larvae for sex ratio determination. Removal
of this many mature larvae from the population could alone explain the
drop in population size from 1977 to 1979. Further, as pointed out by
Murphy (1984, 1988), the use of mark-release-recapture (MRR) methods
on butterflies as delicate as the El Segundo Blue probably profoundly
effect behavior, survival and reproductive dynamics. Even with careful
netting, legs easily break off these fragile animals and lack of legs has
deleterious effects (e.g. Mattoni, 1988). The added trampling necessi-
tated by an intensive MRR program cannot aid recovery.

29(4):277-304, 1990(92)

291

Sociopolitical aspects of planning
HISTORICAL CONFLICTS — LAX DUNES

A major cause of conflict arose from events attendant to the expansion
of the Mines Field into the Los Angeles International Airport (LAX). The
major radar installation (VOR) was located on a 60 acre site purchased
in 1950. Home construction on the remainder of the dunes continued
unchecked until into the 1960s. With increasing air traffic necessitating
construction of the north runway, safety considerations and the onset of
jet age noise, residential living conditions became increasingly difficult.
In a 1965 referendum, over 66% of homeowners elected to be bought out
with the remaining property condemned. Between 1965 and 1975, 822
homes were vacated and over 2000 people relocated.

The $60 million cost was 75% reimbursed by the federal government.
Almost unnoticed during this hubbub, the ESB was listed as an endan-
gered species in 1976 under the Endangered Species Act passed by the
Congress in 1973. All of the LAX dunes south of Sandpiper St. (Figure 3)
were proposed as critical habitat in 1977, a finding abandoned by
modifications to the Act as amended in 1978. 1976 was also the year the
California Coastal Act was enacted in response to a mandate by the
voters to guarantee preservation of coastal environmental values. The
1970s was the first time signs appeared of a public awareness that all was
not right with the environment. This concern eluded many bureaucrats.

The Los Angeles City Planning Department thereafter began work on
a plan to develop the LAX dunes as a recreational facility, emphasizing
a 27 hole golf course. With recognition of the ESB, the plan was modified
to set aside 80 acres as a conservancy plus 12 acres as a preserve/
research/interpretive area. Following requirements set forth by the
coastal act, public disclosure and participation processes were initiated
and hearings began in early 1981. A public interest group, “Friends of the
Dunes”, was established and battle lines drawn. The essence of orga-
nized opposition to any development on the dunes was the position that
the dunes contained a rich biota, exemplified by the listed ESB, and
should be left as open space which would restore itself by natural plant
succession. After eight public hearings, the city submitted the develop-
ment plan to the Coastal Commission for action in late 1983.

Two independent studies of the ESB populations at LAX were per-
formed in 1984. Both studies indicated serious and deteriorating habitat
conditions. Consequent^ airport officials developed a memorandum of
understanding with both U.S. Fish and Wildlife and the California
Department of Fish and Game to support the recreational complex with
the key provision that development would generate funds to assure
restoration and permanent management of the preserve. A conservancy
committee was to oversee long term objectives. Both federal and state
agencies recognized the golf course as the least objectionable of the
funding solutions because in the long run the sand substratum would
remain untouched while the infrastructure of several acres of roads,

292

J. Res. Lepid.

foundations, rubble and old utility appurtenances would be removed.
With assurance that the unique biota were thereby conserved, future
generations would have the option to roll back the golf course with
minimum effort. An alternate hypothesis was advanced that the golf
course could be viewed as destruction of ESB habitat, thereby permitting
development of this area for commercial airport purposes at some
indeterminate future.

At its November 12, 1985 public meeting, the Coastal Commissioners
followed their staff recommendation and rejected the airport develop-
ment plan as inconsistent with the Coastal Act by not assuring preser-
vation of a unique sensitive habitat. However, the Commission did not
accept a staff recommendation that the airport set aside the entire 302
acre site as “environmentally sensitive habitat,” thereby leaving open
exploration of other options. In its wisdom, the Commission did recog-
nize that the basic issue of funding was necessary to both maintain and
restore the habitat. What remained unresolved was that denial of an
institutional funding mechanism would number the days of the “environ-
mentally sensitive habitat.”

From studies of population regulation of the ESB discussed above, it
was apparent that survival of the ESB was in immediate jeopardy. In the
public interest, the Board of Airport Commissioners generously provided
a small contract in early 1986 to relieve the situation until a permanent
solution could be found. This initial contract was successful in assuring
short term survival of the ESB. In 1987 a major biological survey and
ecological evaluation of the site was contracted to provide quantitative
information to elucidate habitat values and open alternatives to resolv-
ing the conflict of providing funding necessary to assure restoration and
management. The study recognized that funding must be developed from
some form of land use at the site. As appealing as restoration of the whole
LAX dune remnant might be, economic realities must be recognized and
a consensus established to provide a viable long term solution.

OTHER EL SEGUNDO BLUE BUTTERFLY HABITATS

The federal Endangered Species Act of 1973 states that “The purposes
of this act are to provide a means whereby the ecosystems upon which
endangered species and threatened species may be conserved (and) to
provide a program for the conservation of such endangered species and
threatened species,...” The Act goes on to define “critical habitat” as not
only the geographic area where such species occur at the time of listing,
but “specific areas outside the geographic area occupied by the species at
the time it is listed ... upon determination by the secretary that such
areas are essential for the conservation of the species.”

Recognizing the intent of the law and the role of the ESB as an indicator
of the unique El Segundo dunes ecosystem, several other habitats and
potential habitats must be dealt with. These have been mentioned above
and include: Chevron, Malaga Cove, DWP right-of-way and adjacent

29(4):277-304, 1990(92)

293

Hyperion property, the Playa del Rey-Ballona backdune, and other
miscellaneous parcels.

CHEVRON: In response to concerns of several local lepidopterists,
Chevron set aside and fenced the 1.6 acre habitat on their NE corner as
butterfly preserve (Oppewall, 1975). A pro bono corporate fund was
provided to make a partial restoration, limited to augmenting the ESB,
its foodplant, and some weeding. This laudatory, though imperfect
approach, produced no conflicts. If judiciously continued, the butterfly
should be sustained indefinitely at Chevron, or as long as Chevron is also
sustained.

MALAGA COVE: Until now there has been no general awareness of
this site, and there are no hard data concerning habitat value at this time.
Ownership is undetermined, but geology and landform of the area imply
further development is impossible. The site is fenced. There are no
obvious conflicts here.

DWP RIGHT-OF-WAY / HYPERION: Since an attempt to permit a
nursery was thwarted by public action, no other land use is imminent.
Although an effort to restore was suggested in 1981, no action has been
taken, with the DWP not yet recognizing the issue. There are no
apparent conflicts, however, except that the matter of funding would be
an obstacle to any restoration and management program. Major resto-
ration is necessary, but proximity to the LAX dunes suggests shared
management expense and greater efficiency, when a program is devel-
oped. The adjacent Hyperion wastewater treatment plant includes 30
acres of slope almost completely covered with non-native flora. Plans for
an inappropriate exotic landscaping plan were partially implemented,
against both the spirit and intent of endangered species legislation and
denial of an important heritage value. The issue was rapidly resolved by
discussion with a key Public Works Commissioner who understood a
restored natural community would be both appropriate and more effi-
cient to manage. The installed landscaping will be removed and the
whole area revegetated with native dune vegetation in 1992-1994.

PLAYA DEL REY - BALLONA BACKDUNE: The 6 acre dunes is
owned by Playa Vista Properties and was subject to intense public
conflict involving many parties. The property is part of the overall 950
acre Ballona Wetlands site. A resolution was tentatively reached (1990)
whereby $10 million will be provided by the developer to a conservancy
that will restore and otherwise utilize about 260 acres of badly deterio-
rated habitat that will be deeded as a perpetual preserve. The benefit to
the developer will be permission to construct residential and commercial
buildings on 748 acres it cannot otherwise now develop. A plan to restore
the dunes fragment was approved by the Coastal Commission and other
planning agencies, but has not been implemented by the developer (late
1991). The U. S. Fish and Wildlife Service granted permits to re-
introduce the ESB. In the meantime habitat values continue to decline
while diverse interests fail to agree. The situation parallels the LA-golf

294

J. Res. Lepid.

course matter, except LAX involves public property, whereas Playa Vista
is private.

BALLONA LAGOON: A 7 acre terrestrial upland exists surrounding
a 9 acre tidal lagoon located across the B a Ilona Greek channel to the
north of the above site. A small, 0.2 acre portion was re- vegetated in late
1990 with 41 species of sand dune plants, including 70 coastal buckwheat
plants. A second planting is proposed for 1992 and will include sufficient
foodplant to attempt an ESB re-introduction. A re-introduction ofBehrs
metalmark, Apodemia mormo virgulti , is scheduled for 1992 as a surro-
gate for predicting success with the ESB on a small fragment. Revegeta-
tion will presently be expanded to the entire site.

Assuring survival of both the ESB and the dunes ecosystem must
involve all of the localities cited above. Site multiplicity alone, assuming
responsible management, will go far in preserving all the sand obligate
plants and animals of the dunes. Although the sites cut across several
political boundaries and public interests, their biological commonality
must be recognized. The sites are summarized in Table 1.

Biological Aspects of Planning
BACKGROUND INFORMATION

Listing the ESB in 1976 was based largely on intuitive information
concerning distribution, abundance, and the nature of threat. Emmel
and Emmel (1973) mention that the then undescribed butterfly was in
danger of extinction. Their opinion was later backed by supporting
historical observations of the mass extirpation of other species on the
dunes by urbanization with only remnants left at LAX, Malaga and
Chevron. Arnold (1983, 1987) presented quantitative information on the
Chevron populations from 1977 to 1986 that showed a steady decline from
1600 to 400 individuals. Transect counts by Mattoni and Murphy (1984
unpub. rept.) provided population estimates at LAX of about 800, or
about twice Arnold’s (1986) MMR estimates from the same time. For
other biological parameters, Shields (1975) first found the butterfly
associated with Eriogonum parvifolium and Arnold (1983) later gave
additional information from Chevron on demographics, dispersal,
foodplant numbers, and parasitoids; but these provided only limited
useful information for developing a conservation plan and management
strategy. Information prior to the Mattoni (1990) report was incomplete
through failure to recognize the very significant differences in habitat
structure at different sites.

STUDIES CONDUCTED TO DEVELOP A CONSERVATION PLAN:
CHEVRON

The conservation plan Arnold (1986) implemented at Chevron essen-
tially followed conventional wisdom without benefit of a planned study
or research program. Detailed numerical estimates of ESB population
size, a cornerstone of the plan, had little relevance to planning other than

29(4):277-304, 1990(92)

295

reiterating the obvious. Although the plantations of buckwheat foodplant
and removal of iceplant had a salutary effect on the ESB, the efforts did
not address the basic biology of either ecosystem structure or its restora-
tion. And as mentioned the damage inflicted by the study itself with
extensive and intensive trampling over a small parcel, may have ac-
counted for some reduction of the ESB.

STUDIES CONDUCTED TO DEVELOP A CONSERVATION PLAN:
LAX

As th ede facto major dunes preserve in both land area and habitat
diversity, attention will be focussed on LAX. After the initial 1984
studies, Pratt and Mattoni made observations that provided insights
explaining the low relative ESB numbers at LAX. Their judgement was
based on population density of the foodplant, which implied that several
times the observed number of butterflies were to be expected. Operating
on a grant from California Department of Fish and Game, they deter-
mined that the critical threat to the ESB was high density of two moth
species Lorita scarifica and Aroga species. Abundance of the moths was
the result of the presence of introduced common buckwheat that pro-
vided them foodplant a month prior to blooming of the coastal buckwheat.
Since the moths are multivoltine, and the ESB univoltine, the added
generation provided a direct competitive edge as well as high density of
the parasitoids they share with the ESB. The Airport Commission
consequently provided emergency funding to remove the buckwheat and
otherwise augment the habitat.

A comprehensive restoration and management plan needs information
not only on the biology of the ESB, but of other components of the
ecosystem which impact not only the ESB, but other sensitive species
found at the site. In order to test ESB population responses, a standard
transect was established for annual monitoring of adult ESB. The
transect path was designed to minimize damage to the substrate. The
biological survey of distribution and abundance of all dunes plants and
animals provide data for a model of the ecosystem to document restora-
tion. The study included information collected by Pierce and his col-
leagues in 1938-1939, collections by others, and old aerial and ground
photographs. The study was funded by the Airport Commission a report
now available (Mattoni 1990). A comprehensive restoration and man-
agement plan is in final preparation and funding sources are being
pursued.

EL SEGUNDO BLUE POPULATION MONITORING

A general estimate of butterfly population size is needed to evaluate
both the impact of management techniques and status. Although mark-
release-recapture (MRR) techniques can be useful, deleterious effects of
such handling on so delicate an insect cannot be justified. In addition to
mortality (reported as 10% in the mission blue butterfly by Reid and

296

J. Res. Lepid.

Murphy, 1986) and behavior modification (Morton, 1984), any perceived
precision MRR might provide is unsupportable for studies involving
fragile endangered species. The alternative of visually scoring along a
regular transect is adequate for providing needed data, but even here the
trampling problem must be minimized. Since MRR had been used
simultaneously with a transect in 1984, crude numbers for calibration are
available (Thomas 1983). Because of the unique behavior of adult
Euphilotes butterflies, the transect count method may be more accurate
than MRR. Since adult ESB spend over 90% of their time on flowerheads
of the foodplant, moving less than 10% of the time and then usually only
when travelling to a nearby flowerhead, a direct estimate of the total
instantaneous population size is possible by making a rapid count of
adults on flowerheads from a sample of each colony aggregate, given a
count of total number and distribution of flowerheads.

An accurate estimate of total population size over the entire flight
period, however, depends upon estimates of birth and death rates and
immigration and emigration. Because these parameters rely on esti-
mates only obtainable from MRR, with its faulty assumptions, total
population size estimates from transect counts will have a large error
component (Mattoni, in prep.).

Monitoring is also possible by sampling flowerheads to determine
numbers and species of larvae. The procedure would be useful to estimate
population densities of the two moths which interact with ESB larvae as
well. Lastly, ESB can also be estimated from pupae counted in sifted soil
from the base of foodplants.

DISTRIBUTION

The ESB is limited to the occurrence of its foodplant, but the relation-
ship is not random. Table 2 lists ESB population counts from five major
clusters of buckwheat foodplant in 1984 and 1986-1990 with numbers of
plants and their flowerhead number (Figure 4). The lack of correlation
between butterfly and foodplant is clearest when comparing clusters 1
and 2 (backdune) with 5 (foredune/V OR). The former each yielded 1.6
butterflies per thousand flowerheads, the latter only 0. 14. This distribu-
tion pattern emphasizes the heterogeneity in habitat values, with the
importance of any given plant being its location. The general distribution
of high quality plants largely on the backdune near the toe of the slope
is a key to conservation. The few high quality plants on the foredune only
grow in small depressions with leeward protection.

DISPERSAL

Adult ESB are sedentary animals that spend the bulk of their time
perching and searching for mating opportunities (males) and ovipositing
and feeding (females). From MRR work, a few individuals moved dis-
tances equivalent to the farthest reaches of the habitat (Arnold 1986).
Using a different approach, Mattoni and Pratt (unpub.) set out mature

29(4):277-304, 1990(92)

295

reiterating the obvious. Although the plantations of buckwheat foodplant
and removal of iceplant had a salutary effect on the ESB, the efforts did
not address the basic biology of either ecosystem structure or its restora-
tion. And as mentioned the damage inflicted by the study itself with
extensive and intensive trampling over a small parcel, may have ac-
counted for some reduction of the ESB.

STUDIES CONDUCTED TO DEVELOP A CONSERVATION PLAN:
LAX

As thede facto major dunes preserve in both land area and habitat
diversity, attention will be focussed on LAX. After the initial 1984
studies, Pratt and Mattoni made observations that provided insights
explaining the low relative ESB numbers at LAX. Their judgement was
based on population density of the foodplant, which implied that several
times the observed number of butterflies were to be expected. Operating
on a grant from California Department of Fish and Game, they deter-
mined that the critical threat to the ESB was high density of two moth
species Lorita scarifica and Aroga species. Abundance of the moths was
the result of the presence of introduced common buckwheat that pro-
vided them foodplant a month prior to blooming of the coastal buckwheat.
Since the moths are multivoltine, and the ESB univoltine, the added
generation provided a direct competitive edge as well as high density of
the parasitoids they share with the ESB. The Airport Commission
consequently provided emergency funding to remove the buckwheat and
otherwise augment the habitat.

A comprehensive restoration and management plan needs information
not only on the biology of the ESB, but of other components of the
ecosystem which impact not only the ESB, but other sensitive species
found at the site. In order to test ESB population responses, a standard
transect was established for annual monitoring of adult ESB. The
transect path was designed to minimize damage to the substrate. The
biological survey of distribution and abundance of all dunes plants and
animals provide data for a model of the ecosystem to document restora-
tion. The study included information collected by Pierce and his col-
leagues in 1938-1939, collections by others, and old aerial and ground
photographs. The study was funded by the Airport Commission a report
now available (Mattoni 1990). A comprehensive restoration and man-
agement plan is in final preparation and funding sources are being
pursued.

EL SEGUNDO BLUE POPULATION MONITORING

A general estimate of butterfly population size is needed to evaluate
both the impact of management techniques and status. Although mark-
release-recapture (MRR) techniques can be useful, deleterious effects of
such handling on so delicate an insect cannot be justified. In addition to
mortality (reported as 10% in the mission blue butterfly by Reid and

296

J. Res. Lepid.

Murphy, 1986) and behavior modification (Morton, 1984), any perceived
precision MRR might provide is unsupportable for studies involving
fragile endangered species. The alternative of visually scoring along a
regular transect is adequate for providing needed data, but even here the
trampling problem must be minimized. Since MRR had been used
simultaneously with a transect in 1984, crude numbers for calibration are
available (Thomas 1983). Because of the unique behavior of adult
Euphilotes butterflies, the transect count method may be more accurate
than MRR. Since adult ESB spend over 90% of their time on flowerheads
of the foodplant, moving less than 10% of the time and then usually only
when travelling to a nearby flowerhead, a direct estimate of the total
instantaneous population size is possible by making a rapid count of
adults on flowerheads from a sample of each colony aggregate, given a
count of total number and distribution of flowerheads.

An accurate estimate of total population size over the entire flight
period, however, depends upon estimates of birth and death rates and
immigration and emigration. Because these parameters rely on esti-
mates only obtainable from MRR, with its faulty assumptions, total
population size estimates from transect counts will have a large error
component (Mattoni, in prep.).

Monitoring is also possible by sampling flowerheads to determine
numbers and species of larvae. The procedure would be useful to estimate
population densities of the two moths which interact with ESB larvae as
well. Lastly, ESB can also be estimated from pupae counted in sifted soil
from the base of foodplants.

DISTRIBUTION

The ESB is limited to the occurrence of its foodplant, but the relation-
ship is not random. Table 2 lists ESB population counts from five major
clusters of buckwheat foodplant in 1984 and 1986-1990 with numbers of
plants and their flowerhead number (Figure 4). The lack of correlation
between butterfly and foodplant is clearest when comparing clusters 1
and 2 (backdune) with 5 (foredune/V OR). The former each yielded 1.6
butterflies per thousand flowerheads, the latter only 0.14. This distribu-
tion pattern emphasizes the heterogeneity in habitat values, with the
importance of any given plant being its location. The general distribution
of high quality plants largely on the backdune near the toe of the slope
is a key to conservation. The few high quality plants on the foredune only
grow in small depressions with leeward protection.

DISPERSAL

Adult ESB are sedentary animals that spend the bulk of their time
perching and searching for mating opportunities (males) and ovipositing
and feeding (females). From MRR work, a few individuals moved dis-
tances equivalent to the farthest reaches of the habitat (Arnold 1986).
Using a different approach, Mattoni and Pratt (unpub.) set out mature

29(4):277-304, 1990(92)

297

potted foodplants at sites up to 0.5 km. outside their normal distribution
area with the objective of finding offspring of dispersing females. The
results were negative. All the flowerheads of two isolated plants in the
disturbed foredune area (see map, Figure 4) were sampled with no ESB
early stages found on 184 flowerheads in 1987. These data indicate
dispersal, and/or distant foodplant locating ability across distances as
small as 200 meters, is not common. Although movement between the
main buckwheat clusters is probably more limited than within clusters,
from a practical viewpoint further investigation is not now warranted.
For purposes of population genetics, even with low inter-cluster move-
ment, the population of LAX is a single unit (Forney and Gilpin, 1989).

The small colonies at Chevron and Malaga Cove are isolated with
probably no effective gene flow possible with the LAX “metapopulation”.
Recent data of Pratt (unpub.) found allozyme frequencies of Malaga
indistinguishable from LAX indicating these populations have not been
separated for enough time to permit deviation in frequency of the tested
loci.

Habitat Values

An ecosystem is physically described by the distribution and abun-
dance of all plant and animal species in a circumscribed area, or habitat,
to which they are co-adapted. High natural value habitat can thus be
characterized by the community of species found prior to human interfer-
ence. Only 1 ha backdune and 5 ha foredune at LAX can thus be
considered relatively pristine habitat, with another dozen hectares
moderately undisturbed. These pristine sites provide species area curve
models for restoration and foci from which to carry out restoration based
on species area curves. Alien species require identification with a plan
for their regulation. Particular care must be devoted to assessing
whether apparent natives are historic natives. The common buckwheat
was heretofore believed a native plant of the dunes, for example, yet
proved to be a serious threat to the ESB because it was not.

All extirpated plants must be re-introduced into their proper micro-
habitats. Where precise information is lacking, this can only be done by
over planting and allowing selection to later segregate density and
distribution. Animal re-introduction will require care to avoid excess
herbivory while maintaining assurance of population regulation within
the food web. Re-establishment of the historic mammalian community
is impossible because of isolation and area limitation for megapredator
support. The absence of models for the sequence of introduction of
smaller mammals is another problem.

Biological Program

The biological program outlines the concerns which affect the viability
of the entire plant and animal communities of the dunes ecosystem.
Although focus remains on the ESB, over 30 identifiable species of

Table 3. Species of special interest on the El Segundo sand dunes. Included are extirpated species that may be reintroduced, extinct
species, species with restricted sand dune distributions, and species recognized by institutional listing. Site: location on prairie (P) or dunes
(D). Listing categories: Federal, Endangered, Fed. 1 is awaiting listing, Fed 2 is candidate for listing, Fed 3 is probably not listable;
California Department of Fish and Game, Cal E is endangered, Cal T is threatened, SSC is a species of special concern; CNPS is the
California Native Plant Society listing. Status indicates population condition at the dunes.

298

J. Res. Lepid.

_ a
o- o

be

bo

j •>

be

s

<4-3

T3

T3

rd

p

.s

a

P

0

e .

a

0

0

0

0

‘m

be

o*

■+J

cd

'm

rH

o

rH

cd

cd

td

Id

2

03

P

H->

cd

£

o

cd

0

Sh

a

• rH
2

0

a

vp

a

p

£

cd

0

Sh

0

0

£?

u

b

0

Sh

a

p

&

43

&

*43

£*

43

&

43

cd

rj

XJ

rQ

cd

0

0

0

cd

rD

X!

X

XI

X!

CO

rH
• pH

M
• r-l

rP

0

cd

Sh

T3

>

>

Sh

cd

0

0

0

0

be

a

43

m

3

PQ

i—i

co

Ph

£

o

T3
a )
u
0
be
P

cd ^

« §
w a

<N

a

0 o

a

CN <N

0 0..
a a "d T5

o o 0 0
p a fa

0 0
P P

O O

p a

cd

^ O
co <n

^ 0
O Ph

-73 W

0 -73

Sh cd

p 0

T3

0

0
• rH

CO

Ph

P

pL, Ph

PPQPQPQPPQ

Q P Ph Ph Ph

1-CH

a£-s

<2 2 £?
p-h o cd

H Q J

"d

$

o

'5

if)

0

u

Cfi
0
* !-H

0

0

P

cd

bi

• rH

I

T3

P

cd

CO

£

£

£ ^
5 -S'

CJ ■

e c3
s^ rh4

^ a

e

CO

a

O

a

CO

O
si
o

CO CO

a a

CO
CO

Hi

CO

o
si
a

a a

be bo k

p p d

£ £ a

•o <ih> N Co Co

29(4):277-304, 1990(92)

299

■+j

P

P

d

P

P

■§

d

cd

©

u

a

m

©

d

P

©

d

CO

a

o

a

2 S

d £

3 ^

o

p

p

p

o

a

p

cd

o

P

©

a

P

p

d d
© ©
d d

cd «

3 3

-4-3 rrt

©

u

©

u

©

u

3

s a

© p
a *43

o

a

a

©

8

a

o

O

©

a

a

CO

©

p

p

P d 4-3

&.&S

43 ‘+3 +3

x 3

p

a

p

•4-3

a x

o

P

P

©

d

XXX

© CQ

u

p

u

CO

p ©

o

Sh

P

o

©

© © ©

0 ©
P

PQ

CONN(M

p d d d d
o © © © ©

P fc fc fe fo

Q

co

© © .
P P d
o o ©
P P fe

©©©©©©
p p p p p p

o © o o o o

p p p p p p

pL-j pLf P_| P_) pi j

QQPQQQPPQQQQPPQ

CO

a

44

Q

s*

©

d

o

e

£

©

s

©

5

0
d
d

CO

s •-*

© *e4

e £

1 2

&'C

e

© o

Cl H
d a
O Cq

«
d
d

© o si

CL SS P

fLT“ d ©

b

o d

« Lit

©

&

o

o

d

s

d

S

o
o

_ _ a

-P O Cq

d ..

| §

CO

d

CO

O ,
^ •-

*> |
co d

d d

& m
E ©

© d

s |

© d

> p

ft o

»«f

C5 ^
P

ft id

d -+j
© p

s

•PM

3

a a

© j®

§ fi
^ £
& <
© cd

£ S

CP hP

-4-3
©

a

©

a

p
cd

a

Cfi

si

W

^ §
S P

p ©

PQ «

O cq

©

©

©

©

q3 fQ
© cd
©

_d cd
o

© TO

a

p CO

d cd

© 8

© P3

43 'P

O P

d ©

O m

<$r

d

©

*o

d
1 ><>

d

d

d

d

s

cq

g

&

co

- a

© d

*»3 sg

d g
Cr fc
o ©
o ***

CO

d

"d

-d

§•

co

d

d

50 *o

* a

co
d
©

S

*d

co

d

*©

o

cq ^ < &q cq

£P

>4

a

be

P

>

JD

Sh

©

£

©

5P

d co

§ s

& 3

o T3

&i

© e°

CQ P3

r-H Cd

w o

p

©

*&
cd g

5 S «

g a.-S

TO • i—i pH

*3 rP a

-2 ^ d

CO CO 2
© © w
p p J®
p p p
d d ©

_ ^_| CO
Cd Cd tn

4J -p

CO CO P

cd cd op
OOP

Q o <d

p

©
® pH

©

>

d

B

CO

£h

©

d

P

P

Ih

o

bi

P

cS

d

©

CO

o

p,

©

p

p,

d

©

o

©

p

CO

4h

O

CO

©

• fM

o

©

Pi

CQ

eq cq

d d

© ©

c P

P P Ph

co

d

o

!&

O

d

P

28 -2
d k co
O *42

S’

o

Co ^
d co
d

i,

© .d

CQ P P

CO

©

-ft

cd

u

pfi

©

t te

> £

i1 ®

cd ©

a s

*S P

I o

©

CO

p

o

a

^ j§

O 4h

O ^

-2 P!

^ H

^ P
O ©

s a>

cd o

P P

Burrowing owl Athene cunicularia PD SSC extirpated?

Least tern Sterna antillarum hrowni D Endangered rare (nesting)

San Diego horned lizard Phrynosoma coronatum hlainvillei PD Fed.2, SSC increasing

California legless lizard Amelia pulchra PD in review decreasing

Western Spadefoot toad Scaphiopus hammondi P in review extirpated

300

J. Res. Lepid.

concern and those limited in distribution to southern California coastal
dunes ecosystems will all be aided by re-establishing an optimal environ-
ment for the ESB. Unfortunately emphasis on the butterfly has been and
will remain important since most non-biologists, including politicians,
attorneys, and planners, can more easily deal with a single organism
rather than a complex system that is both conceptually and in reality
vague (Jensen et al. 1990).

The LAX dunes are a leading “hotspot” of biodiversity nationally when
judged by the number of unique species for area size. To date 11 species
have been identified that are endemic to the El Segundo dunes system,
with many more variously listed as species of concern or significance
(Table 3). These include the extinct El Segundo giant flower-loving fly,
Raphiomidas terminatus terminatus ; the San Diego horned lizard,
Phrynosoma coronatus blainvillei; the seaside calandrinia, Calandrinia
maritima; and the beach spectaclepod, Dithyrea maritima.

Principles that will be utilized to develop a plan include:

1. Restoration. The major emphasis for dunes conservation will be
restoration. With only two acres of backdune in pristine condition, re-
establishment the entire backdune area must have highest priority,
since this is the site of highest ESB population density. Effort will also
be directed to maintain and restore habitat diversity on foredune areas
and the prairie area based upon species present for the 1938 survey. The
coastal prairie, though not ESB habitat, is significant by providing
interface to the backdune. This “edge” produces the highest diversity of
any part of the dunes. It is also a buffer area into which ESB extensively
wanders.

2. Free sand. Sufficient area is necessary, including the prairie
deflation plain, to accommodate free movement of sand, by wind and
water. Sand movement is necessary to maintain this physical process to
which the unique sand obligate biota are adapted. The amount of net
movement may be on the order of one centimeter of depth annually.

3. Human activity negatively impacts the ESB and its foodplant. The
sand dune substrate must be protected from trampling or other degrad-
ing contact.

4. Alien plants are a major threat to both the ESB and its foodplant
either directly (shading, choking seedlings) or indirectly (by chemically
and physically modifying the soil or by serving as alternate hosts to
predators or competitors). Most other native dunes species are also
threatened by alien plants. Evidence indicates the majority of the native
flora is not spreading, but is rapidly being replaced. Active management
is required to reverse this trend.

5. Alien animals have had serious effects that are only partially
understood. The European red fox, responsible for the loss of nearly a
dozen small mammals and scrub obligate bird species, must be extir-
pated. Two of the five most abundant ground dwelling insects, the
Argentine ant and the European earwig (Mattoni, 1990a), have displaced

29(4):277-304, 1990(92)

301

native species and are having other effects. Regulation must be at-
tempted.

6. Reintroductions. All extirpated native plant and most animal
species must be reintroduced. These include species known to have
occurred on the dunes and for which similar genetic stock can be
obtained. Globally extinct species such as the El Segundo giant flower-
loving fly, Raphiomidas terminatus terminatus cannot be resurrected.
Extirpated species are all theoretically available, although some may be
rare.

7. Management. All techniques used must minimize damage to native
species while eliminating alien plants and animals.

8. Monitoring, is essential to assess progress of the program and to
provide advance warning of shifting conditions which might otherwise be
unnoticed.

There are a number of uncertainties for carrying out the plan which
suggest continuing research, including pollution effects from JP-4 jet fuel
hydrocarbons and borates, high noise levels, and possibly extra low
frequency electromagnetic radiation. There is clearly potential to per-
form meaningful experiments for community ecology and to provide
contributions to basic scientific knowledge. The plan must include
provision for local universities and colleges that will enable students to
participate through new research programs.

Institutional Program

An institutional program will be developed to establish a committee to
represent all parties with vested interests in land use, biology, and
environmental concerns. The committee will set policy to implement the
biological program. and be charged with administering the habitat con-
servation plan.

A funding mechanism is crucial to the plan for both the restoration
program and continuing management. The recreational facility plan put
forth for the Department of Airports by the Los Angeles city planning
department in 1983 was designed to finance a habitat conservation plan
through fees generated by the facility developer. The privately managed
facility was to include a 27 hole golf course plus an active recreation area.
Nature was to be served through the establishment of a permanent 80
acre conservancy and 12 acre preserve. All undisturbed areas were
included in the 80 acre conservancy.

The matter is unresolved (1991), yet the clear message from data
gathered to date is that without an active program, the dunes ecosystem
will continue to collapse (Mattoni 1990b). At what point the ESB will
disappear cannot be predicted, but the event could be within decades
without further augmentation efforts. Species recently extirpated from
the dunes cannot be restored without manipulated reintroduction be-
cause there are no nearby natural areas. Exotic plants and animals
continue spreading and can only be controlled or eliminated by directed

302

J. Res. Lepid.

intervention. The restoration and management program will require
funding. The 1983 plan was one approach to solving the problem,
although with an unfavorable land area.

Prognosis

Long term survival of the ESB is dependent on a habitat restoration
and management. To best serve the needs of the butterfly, the entire
useable remnant 277 acre El Segundo dunes ecosystem should be
conserved. To conserve the biota, a habitat conservation plan has been
prepared which addresses the following issues:

1. The ESB and its coastal buckwheat foodplant are essentially
restricted to land which has not been disturbed by human activity. Both
species are indicators of habitat quality and conditions which promote
them will serve to restore other components of the dunes ecosystem.

2. Human activity negatively impacts the ESB and its foodplant. The
sand dune substrate must be protected from trampling or other degrad-
ing contact.

3. Alien plants are a major threat to both the ESB and its foodplant.
Other native components of the dunes ecosystem are also threatened by
alien plants. Active management is required to reverse this trend.

4. Patterns of herbivory have been modified by the extirpation of most
mammal species. Alien carnivores must be removed and native herbi-
vores replaced by programmed reintroduction. These actions require
reestablishment of a food web similar to that originally responsible for
regulating the entire community.

5. Any plan should coordinate programs which encompass all former
habitat which can be restored to a condition approaching the historical
state. This includes the habitat fragments at Malaga Cove, Chevron,
DWP right-of-way, and Play a del Rey dune.

Coda

The re-election of Ruth Galanter for the local city council in summer
1991, in part for her strong stand on protecting the LAX dunes, provided
the climate to establish a new Specific Plan with 200 acres devoted to a
preserve while retaining a 100 acre golf course to satisfy constituency
(Figure 3). The rough areas of the golf course will be vegetated as native
habitat. The Specific Plan was adopted by the Airport Board of Commis-
sioners, the City Planning Department, and the City Council. Final
approval of the California Coastal Commission is the last step before
implementation. A detailed habitat restoration and conservation plan
has been completed. The major remaining obstacle is funding.

There are no funds now available and no provision in the LAX charter
for the use of airport funds for either restoration or habitat maintenance.
Although the Department of Airports paid over $300,000 for both the
biological study and initial enhancement programs, their use of general
funds were justified in promoting the 1983 golf course plan.

29(4):277-304, 1990(92)

303

Over 43 acres have been revegetated to a degree that over 90% of the
plant cover is native (1991). Part of the work was performed by a team
of 60 - 140 volunteers. Expansion of the volunteer program is projected
to provide over 30% of the effort to complete the revegetation phase of the
program. The value of volunteers has proved greater than their donated
time alone both in quality and sensitivity of their work and their
influence in the political base.

Among the other parcels: the Hyperion foredune is scheduled for native
revegetation by the Department of Public Works, contact has just been
made for the DWP backdune, the Chevron butterfly garden program
continues, Malaga Cove remains ignored, the Playa del Rey dune resto-
ration project has been delayed by its project developer, but the Ballona
Lagoon restoration and ESB re-introduction plans are proceeding.

Acknowledgments. Gordon Pratt contributed substantial information concern-
ing many aspects of Euphilotes systematics and biology which are invaluable to our
understanding of these animals. The format of this report follows that of T. Reid and
D. D. Murphy (1987) in their report “The Endangered Mission Blue Butterfly.” For
comparative purposes it seemed valueable to use a standard format to present
background information for planning purposes where diverse interests and insti-
tutions are involved. Selected staff of the Los Angeles Department of Airports were
extremely helpful for many favors and support. In particular Paul Principe not only
made fieldwork a pleasure, but provided graphics support. The airport commis-
sioners, in particular Ms. Maria Hummer, generously provided early funding and
demonstrated concern for the broader issues of my biological study of the dunes.
Councilwoman Ruth Galanter, her staff including especially Rubelle Helgesson
and Betty Fisher, Mayor Tom Bradley, and many individuals across several city
departments cannot be commended enough for their efforts and sensitivity to
permanently preserve the ESB, its habitat at three sites, and the entire dunes
biotic community. The California Coastal Conservancy has provided short term
funding until other sources are developed.

Last, but hardly least, I thank Jeremy Thomas, Paul Opler, and Otakar Kudrna
for their comments that helped both enrich and clarify an earlier version of this
paper.

Literature Cited

Arnold, R.A. 1983. Ecological studies of six endangered butterflies (Lepidoptera,
Lycaenidae): Island biogeography, patch dynamics, and the design of habitat
preserves. Univ. Calif. Berkeley Pub. Entomology. 99. 161 pp.

-^1986. Private and government funded conservation programs for endangered

insects in California. Natural Areas Journal 5:28-39.

Brown, D. R. 1986. The effect of human trampling on the dune-mat vegetation of
the Landphere-Christensen Dune Preserve. Rept. to The Natural Conservancy,
San Francisco. 12 pp.

Cooper, W.S. 1967. Coastal dunes of California. Geol. Soc. Amer. Mem. 104:1-131.
Forney, K. A. and M. E. Gilpin, 1989. Spatial structure and population extinction:

A study with Drosophila flies. Conservation Biology 3: 45-51.

Gilpin, M. E. and M. Soule 1986. Minimum viable populations: processes of species
extinction, pp. 19-34 in M. E. Soule ed/ Conservation Biology: the science of
scarcity and diversity. Sinauer. Sunderland, MA.

304

J. Res. Lepid.

Jensen, D., M. Tom, and J. Harte. 1990. In our hands: A strategy for conserving
biological diversity in California. California Policy Seminar, Univ. Calif.
LAX-EIR 1975. Physical environmental studies, Los Angeles International Airport.

Unpublished report by Olson Laboratories.

Mattoni, R.H.T. 1988. Captive propagation of California endangered butterflies.
Report to Calif. Dept. Fish and Game. Contract C-1456.

1989. The Euphilotes battoides complex: Recognition of a species and

description of a new subspecies. Jr. Res. Lepid. 27: 173-185.

1990. Habitat evaluation and species diversity on the LAX El Segundo sand

dunes. Rept. to the LAX board of airport commissioners.

1990a. Unnatural acts: succession on the El Segundo sand dunes in

California. Proc. Soc. Ecol. Restoration and Management: 581-593.

Murphy, D. D. 1984. Book Review: Ecological studies of six endangered butterflies
(Lepidoptera. Lycaenidae): Island biogeography, patch dynamics and the
design of biological Reserves, by R. A. Arnold. Jr. Res. Lepid. 22: 267-269.

1988. Are we studying our endangered butterflies to death. Jr. Res. Lepid.

26: 236-239.

Morton, A. C. 1984. in R. I. Vane-Wright and P. A. Ackery eds. The biology of
butterflies. (Symposium of the Royal Entom. Soc. Lond. No. 11). Academic,
London.

Oppenwall, J. C. 1975. The saving of the El Segundo Blue. Atala 3: 25-28.

Pierce, D. W. 1938-1940. Unpublished notes on the El Segundo sand dune study.

5 vols. Natural History Museum of Los Angeles County.

Pratt, G. 1987. Competition as a controlling factor of Euphilotes battoides allyni
larval abundance. Atala 15: 1-9.

Reid, T. and D. Murphy. 1986. The endangered mission blue butterfly. U. S.

Forestry Service “Syllabus on managing viable populations”.

Shields, 0. 1975. Studies on North American Philotes IV. Taxonomic and biological
notes and new subspecies. Bull. Allyn Museum No. 28. 36 pp.

Shields, O and J. Reveal. 1988. Sequential evolution of Euphilotes (Lycaenidae,
Scolitantidini) on their plant host Eriogonum (Polygonaceae, Eriogonoideae).
J. Linn. Soc. 33:51-91.

Vivret, N. J. and C. H. Muller. 1977. Mechanism of invasion and dominance of
coastal grassland by Mesembryanthemum crystallinum L. Ecol. Monogr. 47:
302-318.

Von Bloeker, J. 1941. Amphibians and Reptiles of the Dunes. Bull. So. Calif. Acad.
Sci. 40:29-38.

White, R.R. 1986. Pupal mortality in the bay checkerspot butterfly
( Lepidoptera :Nymphalidae ) . Jr. Res. Lepid. 25:52-62.

Journal of Research on the Lepidoptera

29(4):305-315, 1990(92)

Decline and Conservation of Butterflies in Japan

Atuhiro Sibatani

Biological Laboratory, Faculty of Humanities, Kyoto Seika University, 137 Kinotyoo,
Iwakura, Sakyo-ku, Kyoto 606, Japan

Abstract. Japanese lepidopterists so far have not defined their
strategies to counter the decline of butterfly fauna in their country,
although extinctions have so far been restricted to local populations.
National and local government policies are to simply promulgate
protective regulations. In most cases this means nothing more than the
prohibition of collecting and has proved ineffective in reversing the
rapid decline and extinction of butterfly populations. The existence
and significance of the Red Data Book, currently in preparation in
Japan, is as yet not widely acknowledged. However, a volume of
collected papers on the history of the decline and protection of Japanese
butterflies has been published by the Lepidopterological Society of
Japan. It shows that traditional agriculture and silviculture practices
in Japan contributed to the dynamic succession of deciduous and non-
deciduous broadleaf forests (the latter being the laurisylvae) and
maintenance of various types of open fields and meadows, all much
needed habitats for the survival of a butterfly fauna with high
biodiversity. In the preagricultural wilderness, this continuing dyna-
mism was probably effected by forest fires, typhoons and floods. Fire,
wind, water, and traditional culture historically served as protective
agents for butterflies, rather than the formal prohibition of collecting.

Introduction

Japan is an archipelago, situated at the eastern margin of the Palearc-
tic Region, and ranging over 3000 km from subboreal to subtropical
climates, which in latitude roughly corresponds to the expanse from
Quebec to Cuba or from Como to Aswan. It has 238 resident butterfly
species, none of which have yet suffered complete extinction. However,
because of the devastating industrialization and recent rapid rise of the
“living standard” in Japan, many local butterfly colonies have become
extinct. It would therefore be useful to compare aspects of Japanese
butterfly survivability with their European counterparts at the other end
of the Palearctic Region.

Climatic and vegetational differences from Europe

Situated equally in the Palearctic temperate zone, the European and
Japanese butterfly faunas show a remarkable difference in the fact that
Japan has a 3- to 5-times higher precipitation than the average in Europe
throughout the year, but especially during the Monsoon season (June
and July). Hence two-thirds of its quite hilly land remains covered by
(mostly secondary) forests even with the recent heavy industrialization.
Japan thus faunistically lacks the so-called Mediterranean element, but

306

J. Res. Lepid.

instead has a Chinese element. The former is characterized by its
sclerophyll forest, and the latter by the presence of a non-deciduous or
evergreen broad-leaf forest with camphor laurels, camelias and tea
shrubs, non-deciduous “oaks” (evergreen Quercus species) and
Castanopsis, etc. This characteristic forest type is called laurisylvae or
Lorbeerwaelder, and I will abbreviate it as LS in this paper. During the
last interglacial period, which started about 12,000 YBP, but especially
since 6,500 YBP, the LS, which had adapted to the warmer temperate
climate of Japan, started to expand to the north, replacing the deciduous
broadleaf forest of the cooler temperate zone. The latter is essentially
similar to the European forest with many common, or vicariant, butter-
flies species: Pyrgus rnalvae, Carterocephalus palaemon, Thymelicus
leoninus; Papilio rnachaon; Colias palaeno, Aporia crataegi, Pieris napi ,
Anthocharis cardamines; Satyrium (or Fixsenia, Strymonidia) w-album,
Scolitantides orion, Maculinea arionides, Lycaeides argyrognomon,
Plebejus argus, Vacciniina optilete; Mellicta britomartis (rather than
athalia) ,Argynnis paphia, Aglais urticae , Nymphalis io, Limenitis populi,
Apatura metis ; Minois dryas, Coenonympha oedippus , and Erebia ligea.
Today the vegetation of the northern half of Japan therefore looks
essentially European, whereas that of the southern half represents the
flora and fauna related to those occurring over the southern part of
China, extending to Southwest China and the mid-slopes of the Himalayas.
Typical representatives include a number of papilionid taxa and the
satyrine genera Lethe and Neope.

Because of the old civilization of rice cultivation, the primeval forests
have long been lost from virtually all of the Japanese plains and most of
the lower montane areas except within certain religious sanctuaries or
enclosures in the southwestern half of the mainland. The secondary
growth LS is mixed with deciduous forest, most likely representing an
earlier stage of succession with LS the climax. This mosaic nature of the
Japanese woodlands extends at sea level to the northern tip of Honsyu,
the largest island of the mainland group, because of the warm north-
bound currents along both the Pacific and Japan Sea coasts. The
presence of grass bamboo, Sasa spp., characterizes Japanese vegetation
over all the country, including the northernmost island of Hokkaido,
distinguishing it from the neighboring part of the Asian Continent
(Northern and Northeast China, Korea, and Primorye or Amur/Ussuri).

General background of butterfly conservation in Japan

The first legislation to protect butterflies was for “Tennen Kinenbutu”
or “natural monumental things” and was promulgated by the National
Government in 1932 for Panchala ganesa , a typical LS species and a
representative of the Oriental Region in Nara City. It was followed by the
second in 1934 for Spindasis takanonis in Tottori City. This is another
species representing the Oriental/African fauna, once thought to be rare,
with myrmecophilous larvae from eggs laid near pine or cherry trees

29(4):305-315, 1990(92)

307

planted in profusion in parks and gardens. Both these protective
regulations , as many others that followed, simply prohibited local collect-
ing. However, since these species did occur in a number of other places
over the southern half of Japan, collectors have turned away from the
original protected areas to look for them elsewhere. After the war,
increases in motor transport and pesticide spraying occurred, and appar-
ently the original protected populations vanished without any records for
an unknown period of time.

These two examples indicate that legislation that prohibits, as a
protective measure, is totally ineffective and that locally limited popula-
tions will suffer extinction even in the absence of any collecting pressure.
In spite of this, patterns of official “protection” have diverged little from
the stereotype, and since then a large number of local populations, or
species as a whole, have been designated by the National and local
governments as Tennen Kinenbutu or natural monumental “things.”
These now amount to a total of 37 species. Although in some cases the
entire habitat or ecosystem is protected, these regulations are usually
not supported by financial and other arrangements for effective manage-
ment. One gets the impression that the ruling bodies wish, by designat-
ing the protected object (be it the area, ecosystem, population, or species
as a whole), to divert public accusations directed to their failure of
providing appropriate measures for protection of declining species (of
which many are vertebrates), with an excuse that they have, after all, not
failed to make their best efforts. Thus, collecting becomes a scapegoat in
the face of the mounting attacks and anger of the general populace. Both
professional and amateur lepidopterists have complained of this lamen-
table situation, but have so far taken little systematic action to counter
the general trend of the mass extinctions of butterflies in many places in
spite of the increases of areas and species protected against collecting.

The latest efforts by scientists and lepidopterists

In recent years the Environment Agency of the Japanese government
has been making an extensive survey of the status quo of the fauna and
flora in Japan, but no results have been published yet. It is expected,
however, that the Agency will do so by the end of 1989. Prof. Takashi
Shirozu, the dean of Japanese butterfly specialists, serves as consultant
for this survey.

Upon returning to Japan in 1985 after an absence of 19 years, I soon
came to realize that Japanese lepidopterists were behind their col-
leagues in Europe and the United States in making efforts to rectify the
trend of modern civilization which is rapidly endangering the butterfly
fauna in Japan in addition to many other life forms. In late 1987 I made
an appeal to the Lepidopterological Society of Japan on this point. It was
received warmly by committee members, who then offered to organize a
working group to publish volumes of collected case studies on the decline
and extinction of individual butterfly populations. Japanese lepidopter-

308

J. Res. Lepid.

ists, with accumulated data and living memories, are encouraged to write
case studies today. These studies would be indispensable to the future
planning of butterfly conservation in Japan. In order to practically cope
with rapidly changing situations, it wras decided to publish certain
selected cases to start with, rather than to aim at ambitious and time-
consuming exhaustive surveys. The first volume of the series, entitled
Decline and Conservation of Japanese Butterflies, was recently publi-
cized under the co-editorship of Eiichi Hama in Matumoto, Nagano-ken,
Minoru Ishii of the University of Osaka Prefecture (Department of
Entomology of the School of Agriculture), and myself. For the present
paper I have extracted much information from the manuscripts of this
volume, written by a number of active field workers from various parts
of Japan.

While this work was in progress, two important books on similar
subjects have been published: one, by Hiroshi Moriyama (1988), a
professional specialist at the National Institute for Environmental
Resources, Tukuba Science City, deals with the real meaning of nature
conservation. He heavily utilized butterfly material for constructing his
theory, even though not a lepidopterist himself. The other has been
written by Kunihiko Sei (1988), a school teacher living near Mt. Fuji, the
renowned beautiful volcano west of Tokyo, where the butterfly fauna has
been seriously impoverished in recent years. Publications of these two
books within a short interval were quite timely, because the former is
mainly concerned with woodland ecosystems, whereas the latter is
concerned with those of open country. Their main conclusions, together
with those of Hama et al. ( 1989), indicate the key importance of J apanese
traditional agriculture and silviculture was the maintenance and well-
being of a substantial part of the Japanese butterfly fauna. The pattern
of biodiversity of these habitats was the result of human culture. By
implication, these books also suggest how the colorful diversity of the
Japanese butterfly fauna evolved and was maintained before human
settlement on the archipelago.

Ironically enough, it can now be inferred that the diversity of butter-
flies must have repeatedly been regenerated by quasi-periodic rejuvena-
tions of the vegetation in various habitats. Thus a dynamic succession
was maintained either by human interference in natural ecosystems
through the labour-intensive, energy-saving, traditional agriculture on
a “human scale,” or by the devastation of natural ecosystems by fire, wind
and water — a paradoxical means of nature conservation. The role of
collecting is also paradoxical. As long as it is not too heavily energy- or
economically-intensive, collecting provides information needed for timely
actions to conserve butterflies. Blind prohibition of collecting would,
because of the absence of constant surveys, pave the road to mass
extinctions without our being aware of what is happening. This last point
is a conclusion of Hama et al. (1989), as well a message of this paper. I
will now turn to the substance of this contention, heavily leaning on
Moriyama (1988) and Sei (1988) as well as Hama et al. (1989).

29(4):305-315, 1990(92)

309

A summary of the case histories of selected populations of Japanese
species that are described in detail in Hama et al (1989) is given in Table
1. Each species is the object of one or more histories given in that report.
Table 2 presents an assessment of the major causal agents having either
positive or negative effects on population viability of the species cited in
Tablet.

The role played by traditional agriculture — woodland
As mentioned above, the Japanese temperate zone may be subdivided
into two parts: the northeastern cooler temperate zone where a Euro-
pean type of deciduous broadleaf forest thrives, with an occasional
admixture of conifers (probably as the component of climax vegetation);
and the southwestern warmer moist temperate zone where, unlike the
Mediterranean region, LS is the adapted climax in this climate with high
precipitation. Moriyama (1988) demonstrated that during the current
interglacial period which started 12,000 years ago, a gradual warming of
the Japanese Archipelago caused the northern advance of LS. He
estimated the speed of the LS advance into the northeast around 5,000
years ago, replacing the deciduous forests. His techniques included
pollen analysis and the estimation of seed dispersal helped by various
animals. The results indicated that the speed of the natural forest
advance was rather slow: 40 km per 1,000-1,500 years. However, local
vegetation today does not represent the climax of such a LS forest. It
contains, or did until 20-30 years ago, conspicuous patches of deciduous
broadleaf forests or coppice, the origin and maintenance of which proved
to be entirely artificial. Moriyama showed, again estimating the speed
of forest spread, that these broadleaf constituents, once so characteristic
of the Japanese countryside, could not have secondarily found their way
down to southwest from its northeastern retreat, quite some time after
the onset of the interglacial period. This means that the northeastward
advance of SL during prehistoric times coincided with the beginning of
primitive agriculture by the prehistoric inhabitants of the islands with
slash and burn methods. This activity promoted the rapid growth of the
deciduous trees, both naturally and artificially, involving relict local
floral components during the vegetational transition period. Thus, some
part of the flora and fauna associated with the cooler temperate forests
could remain, through human interference from the very outset of the LS
invasion into the warmer temperate forest zone. In fact, this deciduous
association was able to remain as coppice, which generations of farmers
must have used for fuel and compost (hence the origin of the word “to
coppice”). The process cleared undergrowth and was reflected in the
annual as well as intermittent rejuvenation of the woodland. The
associated flora and fauna did not immigrate there secondarily, but
remained in situ during all the time since the onset of the current
interglacial episode. Moriyama points out that the coppice corresponds
to the initial stage of arboreal succession in the warmer temperate zone.

310

J. Res. Lepid.

Table 1. Summary of case studies in Hama et al. (1989). For numbers in the columns of
subsequent processes and possible causes of decline, see Table 2; d, decline; e,

extinction; u, unknown.

Taxon

Year

Subsequent

processes

Possible
causes of
decline

Extinct or
(declined)

Protective
actions taken
on population
or (species)

Pyrgus malvae

(1982)

1975

24

26?

Carterocephalus

(1967), 1 987’1

(1975)

24

15, 26

palaemon

Luehdorfia japonica

1979

u2

L. puziloi

(1968)

1975

10

11,26

now proposed

10

11,13

Colias palaeno

no d3

1975

24

no evidence

for 26

(<1975)*4

1975

24 & d5

u 6

Aporia hippia

1966

1975

e before 24

20,14

Anthocharis cardamines

1978

1975

24

u 7

Artopoetes pryeri

1977

16,17,18 s

Coreana raphaelis

(1987-88)

1973

24 d,e9

12,18,26

1973

e, then 10*10

13,14,17

Niphanda fusca

gradual d

17,18

d

17,18

Shijimiaeoides divinus

1950-72

13,17,18,

21,22

1960’s

6 but...

22

Tongeia fischeri

1960

27

1961

17

1964

27+18-19

(1970’s)

14,19

1975

19

(1981)

shift of habitat

6+17

1984

11

Lycaeides subsolanus

(1970)

1975

24

18+20

yarugatakeana

1975

prone to d

5 but 14

Fabriciana nerippe

1963

11,13

Melitaea scotosia

1963

Uu

(1986)

shift of habitat

13‘12

1983

16,17,19,20

d

13,17,19,20

Aglais urticae

1975

24*13

11,12,17

Limenitis populi

1976?

1975

24

14

1975

24‘14

with 5,9 but u

Oeneis noma

1975

24‘15

probably 21

but not 26

Oeneis melissa

(1975-81)

(1965)

24

26?

Erebia ligea

1975

now danger

20

of e

E. niphonica

1975

now danger

20

of e

29(4):305-315, 1990(92)

311

*1 Isolated populations.

*2 Impoverishment of woodlands, increase of orchards, spraying, invasion of hikers,
construction of a golf course are mentioned.

*3 No change of population density before and after designation as Tennen Kinenbutu.

*4 The population had declined prior to the designation of Tennen Kinenbutu.

*5 The population decline has not stopped after the designation of Tennen Kinenbutu.

*6 The reason may or may not be due to the putative overcollecting.

*7 Used to be an unusually large population. Foodplants are plenty after some modification
of habitat due to sightseeing developments.

*8 Foodplants still remaining.

*9 Citizens’ activities for protection, but population perished after the designation of Tennen
^ Kinenbutu .

*10 Foodplants were planted by citizens and school children. The local government and
collectors have collaborated for restoring the population. One of the model cases of
conservation.

*1 1 Other butterflies species are intact at the habitat where this species alone disappeared.

*12 The marsh where this species thrived before has undergone modification of biotope;
hence this species moved out of this habitat and found some alternative spots along
cultivated areas, which are now being abandoned. The new habitats are thus being
threatened of overgrowth.

*1 3 The habitat for larval feeding has vanished.

*1 4 Population number fluctuating.

*15 Population did not change before and after the designation of Tennen Kinenbutu .

Turning to butterflies, Moriyama indicates that Luehdorfia japonica
(Papilionidae), the vicariant in southwest Japan of the south European
Zerynthia (or Parnalius), requires, for adult feeding, early spring flowers
such as Erythronium japonicum (Liliaceae) or violets blooming on the
sunny bed of the coppice then without leaves. Its larvae require
Aristolochia, foodplants that are generally associated with LS, with some
species growing in coppice areas on the periphery of their range. After
vegetational succession causes a denser woodland, L. japonica can no
longer thrive both because of the lack of enough light for adult butterflies
and of adequate food for both adults and larvae.

The coppice is now rapidly vanishing from the rural areas because of
residential, industrial, and tourist developments. It is also disappearing
from hilly slopes owing to the decline of traditional agriculture and
forestry with their lower productivity. Presently there is little rejuvena-
tion of the deciduous forests with abandonment of the regular felling at
about 20 years’ rotation cycle for charcoal production and coppicing for
other purposes associated with the traditional self-sustaining rural life.
Deciduous trees have been replaced either by conifer plantations or by
the advance of succession in the absence of human interference. This
latter process darkens the hillsides with a thick cover of vegetational
growth from high precipitation. Thus, the characteristic butterfly fauna
and conventional agriculture, as integral parts of the traditional Japa-
nese culture, are quickly disappearing.

Some species of arboreal thecline lycaenids such as Japonica saepestriata
and Favonius yuasai that respectively feed on Coreana raphaelis and on
young Quercus acutissima or Fraxinus are heavily dependent on trees

312

J. Res. Lepid.

Table 2. Assessment of various natural and artificial processes for their effects
upon survival and conservation of butterflies. (Processes are numbered for their

references in Table 1 .)

Positive effects

Negative effects

1 . Arrest of vegetational succession.

2. Forest fires of natural origin, typhoons.

3. Traditional forestry: coppicing, removal
of undergrowth, charcoal production,
diverse plantations.

4. Variation of flow rate of rivers.

5. Naturally caused floods.

6. Traditional agriculture: slash and burn,
conversion, crop rotation, plantation, grass-
cutting, animal husbandry.

7. Horticulture and gardening compatible
with high plant diversity.

8. Policy of environmental protection
backed up with adequate financial support.

9. Ordinary collecting with modest
intensity, surveys, observations.

10. Lepidopterists’ advice on protective
measures and volunteer patrolling.

1 1 . Large scale plantation of single species
including conifers and other afforestation.

1 2. Large scale felling of forests.

1 3. Decline of traditional forestry and
agriculture: cessation of coppicing and
grass-cutting; spraying.

1 4. Large scale management of water flow:
inundation by dams, artificial stabilization of
flow routes, technological sealing of slopes,
utilization of dry river-beds for
developments.

1 5. “Super”-forest roads (suupaa rindoo)
for tourism.

1 6. Drainage and other uses of marshes.

17. Road construction and deforestation.

18. Residential development.

19. Industrialization.

20. Resort and sightseeing development:
opening of camping, parking, and skiing
grounds.

21 . Expansion of leisure pursuits and
tourism: golf courses, resort hotels.

22. Modern agriculture: use of mechanical
power, pesticides, large-scale farming.

23. Large scale gardening in artificial
settings.

24. Designation of Tennen Kinenbutu
(Natural Monuments) or other prohibitions
without adequate management, monitoring.

25. Overprotection and excessive
prohibition of human interference.

26. Excessive collecting, especially of
declining populations.

27. Magnifying effects of overconstruction
of mountain roads, etc., on natural disasters
caused by typhoons and floods.

planted by the agricultural community [Moriyama (1988) and others (see
Hama et al. , 1989)]. Because of the decline of the traditional way of life
in rural areas, those trees are now largely being eliminated. Further-
more, in the totally protected areas the native forests are permitted to
grow taller and denser. Here the young growth upon which larvae
depend has disappeared under the thick canopy along with their adult
butterflies stages. It is now evident that traditional agriculture contrib-
uted to maintaining high biodiversity of both flora and fauna.

29(4):305-315, 1990(92)

313

The role played by traditional agriculture — open lands

Meanwhile, Sei (1988) and Hama et al. (1989) have demonstrated that
a parallel situation applies to open country including meadow, marsh,
and rocky areas. Again, the traditional agriculture periodically elimi-
nated overgrown grasses and rejuvenated diverse growth within the low
vegetation. The rotation of cultivated lands also rejuvenated open land
vegetation. Such customs thus artificially created diverse conditions for
various open-country ecosystems. Today these practises have long been
abandoned and the land is now utilized for development including
housing, industry, resorts, modernized agricultural combines, and ani-
mal husbandry. Some of these developments were not even economically
successful. Gone with them were characteristic open-land butterfly
fauna including Shijimiaeoides divinus, Tongeia fischeri (Lycaenidae,
Polyommatini), Phoebe scotosia and Fabriciana nerippe (Nymphalidae).
Sei ( 1988) pointed out that a river, because of variation of its gradient and
hence the velocity of the flow, along its path diversifies the physical
conditions of the bank and encompassed ecosystem and thus maintains
the diversity of butterfly fauna across the surrounding open lands.
According to Hama etal.i 1989), the vegetation along a river is vulnerable
to flooding, yet would regenerate quickly, thus contributing to the
dynamic maintenance of flora and fauna. Current construction technol-
ogy for water flows to prevent landslides and floods and the development
of river banks for modified land use destroys, or at best simplifies,
ecosystems of the bank and stabilizes a river’s route. Coupled with the
decline of traditional agriculture, the diverse habitats of open lands are
lost, resulting in the local extinction of many open country species of
butterflies. Thus, prior to the human intervention, heavy rainfall with
resultant floods and devastating typhoons, along with natural forest
fires, played a critical role in maintaining the diversity of butterfly
habitats in wilderness. The advent of agriculture probably did not
threaten this diversity, largely because it contributed to diversifying the
ecosystem. Only when combined with extensive human interference
resulting in the extensive disappearance of woodland could those natural
disasters produce irreparable damage to the local ecosystems.

Other factors including collecting

Apparently, however, not all factors explaining butterfly extinction are
classified. Hama et al. ( 1989) present two examples in which the reasons
for extinction could not be identified. In both cases land was modified
without a decline in the food plants, yet subalpine Anthocharis cardamines
(Pieridae) became extinct and other butterfly species associated with
Melitaea scotosia in the same grassland habitat declined as well. As for
the effect of collecting on the population of alpine butterflies, there are
two opposite views expressed in Hama et al. (1989). In the central
highlands of Honsyu, Oeneis noma (Nymphalidae, Satyrinae) has not

314

J. Res. Lepid.

declined in the alpine area and did not show any change in population
density before and after the 1975 prohibition of collecting. However, in
small isolated montane areas in Hokkaido, Oeneis melissa and Pyrgus
malvae (Hesperiidae) (both protected species) suddenly started to de-
cline during 1975-1982. Illegal overcollecting is the suspected cause,
although there is no clear-cut evidence.

Comparison with other countries

In urban areas, butterflies are not scarce, but their diversity is always
substantially lower than in nearby natural areas. In this regard, the
current Japanese scene may be comparable to Sydney, Australia (1985)
and Khabarovsk in the Far East USSR (1988), where I found ordinary
butterflies to be numerous in parks and gardens. However, my observa-
tions in Seoul, Korea (1984 and 1986), and Beijing and Kunming in the
People’s Republic of China (1988) indicate that butterflies were quite
scarce in urban and extended agricultural areas while they were fairly
abundant in woodlands. These crude observations may be related to
spraying insecticides, but I have no hard data.

I wish to add that several species endangered in some parts of Europe
have not so far showed signs of decline in Japan. Such species include
Papilio machaon (still a pest in home gardens), Minois dryas and
Coenonympha oedippus .

Strategy for protection and conservation

Butterflies are organisms. They are not simply “things” to be protected
as archaeological or artistic objects and cannot be maintained as such.
They represent life processes. It is thus fundamentally wrong to protect
them as tangible objects. Continuing population maintenance requires
dynamic processes. The centerpiece of a dynamic system requires
constant attention and monitoring. In today’s society this means money.
To designate a population or species of butterfly or an area to be protected
is meaningless unless supported by budgetary action. This includes
defending the habitat against invasions by exotic competitors, constant
management using traditional agriculture or forestry practise as a
model, and monitoring and ranging activities. We urge Japanese
authorities to consider these points. At the same time, the trend of
establishing green zones in urban areas should be directed towards
restoring ecosystems native to these areas. Human residences should be
rearranged and removed from some larger habitat areas, simply because
commensal animals as sparrows and rodents might be factors affecting
some sensitive species. A large effort in politics and education are needed
from butterfly specialists and collectors alike.

Acknowledgements. I express my hearty thanks to those who provided me with the
information needed to write this article. They include the authors and co-editors of
Hama et al. (1989), T. Hirowatari, Akiko Oohata, and Y. Shoji. I also express my

29(4):305-315, 1990(92)

315

gratitude to Rick Davis, Rudi and Leona Mattoni for linguistic corrections, and to
O. Kudrna and P. S. Wagener for encouraging me to participate at the Wageningen
Congress.

Literature Cited

Hama, E., M. Ishii, and A. Sibatani (Eds.), 1989. Nipponsan tyoorui no suiboo to
hogo (Decline and conservation of butterflies in Japan), Part 1. Nippon Rinsi
Gakkai (Lepidopterological Society of Japan), Osaka. In Japanese.

Moriyama, H., 1988. Sizen o mamoru towa dooiu kotoka (What is the meaning of
nature conservation?). Noobunkyoo, Tokyo. In Japanese.

Sei, K, 1988. Huzisan ni sumenakatta tyoo tati (Butterflies which failed to live on
Mt. Fuji). Tukizi Syokan, Tokyo. In Japanese.

Editor’s note: This paper is a modified version of the presentation at the Interna-
tional Congress “Future of Butterflies in Europe: Strategies for Survival” organized
by the Agricultural University Wageningen and held at the International Agricultural
Centre, Wageningen, Netherlands during 12-1 4. IV. 1989. A portion of the paper
comprised the English summary of Hama, Ishii, and Sibatani (1989) and is printed
with permission of the Lepidopterological Society of Japan.

INSTRUCTIONS TO AUTHORS

Manuscript format: Two copies must be submitted, double-spaced, typed, with wide
margins. Number all pages consecutively. If possible italicize rather than underline scientific
names and emphasized words. Footnotes are discouraged. Do not hyphenate words at the right
margin. All measurements must be metric. Time must be cited on a 24-hour basis, standard
time. Abbreviations must follow common usage. Dates should be cited as: day-Arabic numeral;
month-Roman numeral; year- Arabic numeral (ex. 6. IV. 1984). Numerals must be used for ten
and greater e.g. nine butterflies, 12 moths.

Electronic submission: The Journal is now being produced via desktop publishing, allowing
much shorter publication times. Although typewritten manuscripts are acceptable, those
submitted on computer disc are highly preferred. After being notified of your paper's
acceptance, submit either a Macintosh or IBM disc version. Include on your disc both the fully
formatted copy from your word processing programand a text-only (ASCII) copy. The two most
preferred formats are Microsoft Word for the Macintosh and either Microsoft Word or Word
Perfect for the IBM, although translation utilities will allow conversion from most formats. Put
returns only at the ends of paragraphs, not at the end of each line. Use one tab to indent each
paragraph. Even if your printer is incapable of outputting italics, please specify italics rather
than underline in your disc copy. Please note any special characters that are used in either the
body of the text or the tables {e.g.. e, u, °, §, p, 6 , 9 ).

Title material: All papers must include the title, author's name, author’s address, and any
titular reference and institutional approval reference. A family citation must be given in
parenthesis (Lepidoptera: Hesperiidae) for referencing.

Abstracts and Short Papers: All papers exceeding three typed pages must be accompanied
by an abstract of no more than 300 words. Neither an additional summary nor key words are
required.

Name citations and Systematic Works: The first mention of any organism should include
the full scientific name with unabbreviated author and year of description. There must be
conformity to the current International Code of Zoological Nomenclature. We strongly urge
depositing of types in major museums, all type depositories must be cited.

References: All citations in the text must be alphabetically listed under Literature Cited in
the format given in recent issues. Abbreviations must conform to the World List of Scientific
Periodicals. Do not underline or italicize periodicals. If four or less references are cited, please
cite in body of text not in Literature Cited. For multiple citations by the same author(s), use
six hyphens rather than repeating the author’s name.

Tables: When formulating tables, keep in mind that the final table will fill a maximum space
of 11.5 by 19 cm either horizontally or vertically oriented. Number tables with Arabic
numerals. When submitting tables on disc, use tabs between columns rather than multiple
spaces.

Illustrations: Color can be submitted as either a transparency or print, the quality of which
is critical. Black and white photographs should be submitted on glossy paper, and, as with line
drawings, must be mounted on stiff white cardboard. Authors must plan on illustrations for
reduction to page size. Allowance should be made for legends beneath, unless many consecutive
pages are used. Drawings should be in India ink. Include a metric scale. Each figure should
be cited and explained as such. Each illustration must be identified by author and title on the
back. Indicate whether you want the illustration returned at your expense. Retain original
illustrations until paper is accepted. Legends should be separately typed on pages entitled
"Explanation of Figures." Number legends consecutively with separate paragraph for each
page of illustration.

Review: All papers will be read by the editor(s) & submitted for formal review to two referees.

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

Volume 29 Number 4 Winter 1990(1992)

IN THIS ISSUE

Date of Publication: September 1, 1992

Conservation of butterflies in Australia 237

T. R. New

Changes of Distribution of thermophilous butterflies in Slovakia 254

Miroslav Kulfan and Jan Kulfan

Effects of a Microbial Insecticide, Bacillus thuringiensis kurstaki, 267

on nontarget Lepidoptera in a Spruce Budworm-infested Forest
Jeffrey C. Miller

The Endangered El Segundo Blue Butterfly 277

Rudolf H. T. Mattoni

Decline and Conservation of Butterflies in Japan 305

Atuhiro Sibatani

Cover Illustration: Photographs of El Segundo Blue and stages in its life
history by R. Mattoni.

He journal

| RESEARCH
ON THE LEPIDOPTERA

i , i mm
i x * ;

Volume 30
Number 1-2
Spring 1991

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

ISSN 0022 4324

Published By: The Lepidoptera Research Foundation, Inc.

9620 Heather Road
Beverly Hills, California 90210
(310) 274-1052

Founder: William Hovanitz

Editorial Staff: Rudolf H. T. Mattoni, Editor

Scott E. Miller, Assistant Editor

Associate Editors: Emilio Balletto, Italy

Henri Descimon, France
Philip DeVries, U.S.A.

Thomas Emmel, U.S.A.
Lawrence Gall, U.S.A.
Hansjuerg Geiger, Switzerland
Otakar Kudrna, Germany
Arthur Shapiro, U.S.A.

Atuhiro Sibatani, Japan
Karel Spitzer, Czechoslovakia

Manuscripts and Notices Material may be sent to the Editor at:
9620 Heather Road, Beverly Hills, CA 90210 (310) 274-1052

The JOURNAL is sent to all members of the FOUNDATION.

Classes of Membership:

Regular (Individual)
Contributing

Student/Retired-Worldwide
Subscription Rate/Institutions
Life

$ 20.00 year (vol.)

$ 30.00 or more, year (vol.)
$ 18.00 year (vol.)

$ 28.00 year (vol.)

$ 250.00

STATEMENT OF OWNERSHIP AND MANAGEMENT

THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA is published four times a year by THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The office of publication and the general business
office are located at 9620 Heather Road, Beverly Hills, California 90210. The publisher is THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The Editor is R. H. T. Mattoni at the above address.
The Secretary-Treasurer is Leona Mattoni at the general business office. All matters pertaining to
membership, dues, and subscriptions should be addressed to her, including inquiry concerning mailing,
missing issues, and change of address. The owner is THE LEPIDOPTERA RESEARCH FOUNDA-
TION, INC., a non-profit organization incorporated under the laws of the State of California in 1965. The
President is R. H. T. Mattoni, the Vice President is John Emmel, the Secretary-Treasurer is Leona
Mattoni. The Board of Directors is comprised of Barbara Jean Hovanitz, Leona Mattoni, and R. H. T.
Mattoni. There are no bond holders, mortgages, or other security holders.

Journal of Research on the Lepidoptera

30(1-2):1-13, 1991

The distribution of radiolabeled pigment
precursors in the wing patterns of nymphalid
butterflies

H.F. Nijhout

Abteilung Allgemeine Zoologie, Universitat Ulm, D-7900 Ulm, Germany

Abstract. The incorporation of radiolabeled tyrosine and tryptophan
into the wing patterns of six species of Nymphalidae ( Precis coenia,
Inachis io,Aglais urticae,Araschnia levana,Heliconius charitonia, and
Agraulis vanillae) was studied. Tyrosine is the precursor for the
pigment melanin (black to brown), while tryptophan is the precursor
for the pigments 3-hydroxykynurenine (yellow) and ommochromes
(orange to brown). Tyrosine always and exclusively labeled the dark
brown and black scales. The red, light brown, and tan portions of the
color pattern never labeled with tyrosine suggesting that the browns
and earth tones of the nymphalid wing patterns are not melanins.
Tryptophan labeled the yellow portions of the wing pattern in H.
charitonia , and the red, orange, brown and tan portions of the pattern
in the other species we studied, suggesting that the pigments in these
regions are all members of the tryptophan-ommochrome pathway in
the Nymphalidae. In most cases these two pigment pathways appear
to exclude each other in single scales, and only one precursor is
incorporated in any given scale. But in a few cases, particularly in A.
levana, both pathways may be active in some scales.

Introduction

The pigments that make up the color patterns of butterflies belong to
relatively few chemical families. The vast majority of wing patterns are
made up of melanins and ommochromes. Pterins are common in the
Pieridae, while papiliochromes, bile pigments and flavonoids occur
sporadically in several families (Needham, 1974; Nijhout, 1985). The
most common of the pattern pigments, the melanins and ommochromes,
are also among the most difficult of all animal pigments to isolate and
identify. The melanins are indefinitely large polymers, complexed with
proteins, and almost completely insoluble. Attempts to solubilize mela-
nins generally destroy their molecular structure (Needham, 1974). The
ommochromes are fairly large organic molecules. Many of them are also
complexed with proteins in situ and thus rendered poorly soluble unless
harsh methods are used for their extraction (Linzen, 1974).

2

J. Res. Lepid.

Melanins range in color from black, through brown and red, to yellow.
The colors of ommochromes largely overlap those of melanins, ranging
from brown through red to orange-yellow. As a consequence of the
considerable overlap in the color of these pigments, and their general
resistance to solubilization and characterization, it has proven ex-
tremely difficult to determine whether a specific patch of color on a
butterfly wing is melanin or ommochrome. Solubility properties cannot
distinguish between melanins and poorly soluble ommochromes and are
therefore inadequate for their identification. Chemical characterization
by chromatography after vigorous extraction is possible (Umebachi,
1980), but this requires sufficiently large amounts of starting material
that it cannot be used to identify a pigment in a specific small area of a
wing.

In the present paper we explore a method for detecting melanins and
members of the tryptophan-ommochrome pathway in situ, by taking
advantage of the fact that they are synthesized from different organic
precursor molecules (tyrosine and tryptophan, respectively). When a
radiolabeled precursor specific for one or the other pigment families is
injected into pupae, that pigment becomes radioactively labeled, and its
location in the intact wing can be visualized by autoradiography. While
this method does not allow one to distinguish among the various
ommochromes or among the various molecular forms of melanin, it can
identify a given patch of color as containing either melanin, or
ommochrome, or both. Autoradiography thus allows us to study the
pattern of specific pigments and to determine whether more than one
pigment can occur in a given element of the color pattern.

Materials and Methods

The following species were used for these studies, all members of the family
Nymphalidae: Precis coenia , Inachis io, Aglais urticae, Araschnia levana,
Heliconius charitonia, and Agraulis vanillae. Two radiolabeled amino acids were
used in these studies: L- [methyl-14 C] -Tryptophan and L-[14C(U)]-Tyrosine (New
England Nuclear). These compounds were dissolved in an insect saline and their
pH was adjusted to between 6.8 and 7.0 with phosphate buffer. Volumes of 2 pi,
containing approximately 0. 1 pCi of radiolabel were injected into pupae 12 to 24
hours before the onset of pigment synthesis. In a few cases (A. urticae and A.
levana ), 1.0 pCi of tryptophan was fed to larvae during their final instar. Adults
were allowed to harden their wings for about 12 hours after emergence and wings
were then prepared for autoradiography as follows. The radiolabeled amino acids
were incorporated into the general proteins of the wing cuticle at a low rate and
this resulted in a modest but bothersome amount of non-specific background
noise in autoradiograms. We were able to remove nearly all of this background
radioactivity by stripping the scales onto adhesive plastic. Wings were pressed
tightly between two pieces of book-cover plastic (obtainable from office supply
stores) with the sticky sides facing each other. When the two pieces of plastic
were peeled apart the scales of dorsal and ventral wing surfaces adhered tightly
to the sticky surfaces of the plastic while the wing itself was easily peeled away.

30(1-2):1-13, 1991

3

The adhesive surface was then covered with thin plastic food wrap (Handi-Wrap
II, DOW Chemical). The food wrap side of this plastic sandwich was the pressed
against X-ray film (Kodak Diagnostic Film, X-Omat AR) and stored in a freezer
at -70 °C for 1 to 3 weeks. After this exposure period the film was developed in
an automatic film processor (Konica QX-60A). Prints were made from these
autoradiograms. Hence the bright (white) portions of the prints correspond to
areas of radiolabel incorporation.

Results

Figures 1-6 show the tyrosine and tryptophan incorporation patterns
in the six species of butterflies. In all cases tyrosine appears to be
incorporated primarily into the black and dark brown portions of the
pattern while tryptophan is incorporated into the red, brownish orange,
and yellow portions of the pattern.

In the dorsal fore wing of Precis coenia (Fig. 1A, B), for instance,
tyrosine incorporation is strongest in the black eyespot and in the black
bands that flank the two red bars in the discal cell. There is a fainter
incorporation in the outer ring of the eyespot and in the areas that flank
the broad white band on the wing, both of which are dark brown. In the
hind wing radiolabel incorporation is primarily in the distal half of the
central disk of the two eyespots (which is the only portion of the central
disk that is black), in the rings around the eyespots, and in one of the
narrow submarginal bands. By contrast, tryptophan incorporation in
the dorsal fore wing is strongest in the two red bars within the discal cell,
and in the white band. Evidently the red is a tryptophan derivative,
possibly an ommatin (Butenandt et al., 1960), and the white regions
contain a significant amount of a tryptophan derivative as well. Nijhout
(1980) noted that the “white” regions of the fore wing of P. coenia is
actually not pure white but a light buff color and contains a low
concentration of what appears to be the same pigment as that which
occurs in the red bars. The tryptophan label is also incorporated in all the
orange scales in both fore and hind wing, and also in the yellow-brown
scales in between the inner and outer ring of the eyespots on the hind
wing. Finally, the light brown ground color in the basal half of the wing
labels moderately with tryptophan but not with tyrosine, suggesting
that, in contrast to the dark brown regions, these light brown areas
contain an ommochrome and not a phaeomelanin, as suggested by
Nijhout (1980).

Radiolabel incorporation in the ventral pattern ofP. coenia (Fig. 1C, D)
resembles that seen on the dorsal side: Tyrosine is incorporated prima-
rily in the black portions of the pattern and tryptophan in the red-orange
portions. The incorporation pattern in the brown portions of the pattern
is a bit more difficult to interpret. There is a slight incorporation of both
the tyrosine and tryptophan label throughout the background, although
the pattern elements of the nymphalid ground plan (Nijhout, 1985, 1991)
label only with tryptophan. The actual pattern on the ventral hind wing

4

J. Res. Lepid.

Figure 1 . Radiolabel incorporation pattern in the wings of Precis coenia. These are
negative autoradiograms, so that white areas indicate regions of radiolabel
incorporation. A and B, dorsal wing surfaces. C and D, ventral wing
surfaces. A and C, tyrosine incorporation pattern. B and D, tryptophan
incorporation pattern.

30(1-2):1-13, 1991

5

consists primarily of light brown pattern elements superimposed on a tan
background. Evidently the pattern elements are a tryptophan derivative
(again, probably an ommochrome), while the background contains a
mixture of tyrosine and tryptophan derivatives.

The dorsal wing pattern of Inachis io (Fig. 2A, B) likewise shows
tyrosine labeling in the black portions of the pattern and a very strong
tryptophan labeling of the reddish brown portions. The long hairlike
scales in the dorsal hind wing label particularly intensively with trypto-
phan. The white dots on the dorsal fore wing that are the homologs of the
foci of the border ocelli (Nijhout, 1985, 1991) do not label with either
amino acid, as was the case with the white scales at the center of the fore
wing eyespots in P. coenia (Fig. 1). The ventral wing pattern consists of
a fine black ripple pattern on a dark brown background. Both pattern
and background label only with tyrosine, the black more intensely than
the dark brown. There is little or no tryptophan labeling of the ventral
pattern.

In the wings of Aglais urticae (Fig. 3) there are also clear demarcations
between areas of tyrosine and tryptophan label incorporation, corre-
sponding to the black and reddish-brown areas, respectively. As in the
previous species, the intensity of labeling corresponds to the intensity of
pigmentation.

For our studies with Araschnia leuana we used the summer form,
prorsa (Fig. 4). The dorsal pattern of this form is mostly black, with a
dislocated whitish band running across both fore and hind wing, and with
one or two narrow reddish-brown submarginal bands on the hind wing
(Fig. 4). The black portions incorporate tyrosine, while the reddish-
brown submarginal bands label intensely with tryptophan. In some
specimens there is a weak tryptophan labeling of the whitish band (as in
the fore wing of P. coenia ), especially when reddish scales are scattered
within the white band. This effect is strongest in forms intermediate
between the normal spring and summer forms (Koch, 1991). Again, the
small white spots on the dorsal and ventral surfaces that correspond to
the foci of the border ocelli are not labeled with either tyrosine or
tryptophan.

The ventral color pattern of A. levana is more complex with areas of
reddish brown, dark brown, black, and white pigmentation. The white
portions of the pattern incorporate a small amount of the tryptophan
label, as in P. coenia , and are not labeled by tyrosine, while the black
portions of the pattern are labeled by tyrosine and not by tryptophan.
Most of the reddish-brown and dark brown portions of the wing pattern
incorporate both the tyrosine and the tryptophan label, although the two
labeling patterns do not overlap precisely. Tyrosine labels the reddish-
brown areas weakly, and the dark brown areas strongly (and the black
areas strongest of all). This is best visible in the ventral fore wing in the
region between R and M: and 9. Thus tyrosine labeling intensity corre-
sponds to the degree “darkness” of the color. Tryptophan labeling, by

6

J. Res. Lepid.

Figure 2. Radiolabel incorporation patterns in the wings of Inachis io. Arrangement
as in Fig. 1 .

30( 1-2): 1-13, 1991

7

Figure 3. Radiolabel incorporation patterns in the wings of Aglais urticae . Arrange-
ment as in Fig. 1 .

8

J. Res. Lepid.

Figure 4. Radiolabel incorporation patterns in the wings of Araschnia levana.
Arrangement as in Fig. 1 .

30( 1-2): 1-13, 1991

9

contrast, is particularly
intense in the reddish-
brown areas of the pat-
tern, best visible in the
proximal third of the dor-
sal wing pattern and in a
broad band correspond-
ing to the field of the bor-
der ocelli. This latter field
is bordered on both sides
by bands that label only
with tyrosine (this is par-
ticularly easily visible on
the hind wing autoradio-
grams), and correspond
to bands of black scales in
these areas. By eye, these
bands are barely distin-
guishable from the dark
brown field in which they
occur. Thus autoradiog-
raphy is a tool that helps
to reveal a differentiation
in this portion of the color
pattern. It should be in-
teresting to investigate
the relation between this
biochemical differentia-
tion and the extreme sea-
sonal polyphenism of this
Figure 5. Radiolabel incorporation patterns in the species.

dorsal wing surface of Helicon i us The dorsal wing pattern

charitonia. A, tyrosine incorporation pat- 0f Heliconius charitonia
tern. B, tryptophan incorporation pattern. (Fig 5) is a crisp and bold

pattern of black and yellow bands. The black bands are melanin while
the yellow bands owe their color to 3-hydroxykynurenine, an intermedi-
ate in the ommochrome pathway (Linzen, 1974). Radiolabel incorpora-
tion reveals that tyrosine is incorporated exclusively in the black por-
tions of the pattern while tryptophan incorporation is restricted to the
yellow bands and the red dots on the ventral side at the base of each wing.
There is no overlap of labeling. Thus the black portions of the pattern
evidently do not mask a more broadly distributed yellow color. Yellow
and black are true alternatives.

Agraulis vanillae shows intense incorporation of tryptophan label in all
the reddish-brown areas of the dorsal fore and hind wings, and in the

10

J. Res. Lepid.

reddish areas of the ventral fore wing (Fig. 6B, C). Tyrosine labels all the
black portions of the pattern. Particularly striking in this species is the
intense tyrosine labeling of the black scales that overlie the major veins
on the dorsal fore wing (Fig. 6A). The ventral hind wing labels fairly
homogeneously but not very heavily with tyrosine. The silver spots on
the ventral hind wing do not show any differential label incorporation.
The silver spots label as intensely as the light brown “background”, which
suggests a homogeneous pigmentation of this wing surface behind the
structural coloration that is responsible for the silver.

Discussion

It is evident from the foregoing that radiolabeled tyrosine and trypto-
phan are incorporated differentially into the wing patterns of butterflies.
Tyrosine, a precursor for melanin, is associated with the black and dark
brown portions of the color pattern. In the Nymphalidae there are no
other known pigments for which tyrosine is a precursor, and since in
several nymphalid species the tyrosine label is uniquely associated with
the black portions of the pattern, we may adopt the working hypothesis
that radioactive tyrosine labels only the melanins. Tyrosine is also a
precursor for the papiliochromes (Umebachi, 1985), but these pigments
are restricted to the Papilionidae. In the species that we studied tyrosine
labels only the black and dark brown portions of the pattern (and the
reddish-brown portions in A. levana), but not the light brown, red, or tan
portions, which suggests that the browns and earth tones in butterfly
color patterns are not all melanins.

Tryptophan is a precursor for pigments in the tryptophan-
ommochrome pathway (Linzen, 1974; Needham, 1974; Umebachi, 1980).
Radiolabeled tryptophan becomes incorporated in the red, light brown,
tan, and off-white portions of the pattern, suggesting that their pigments
are members of that pathway, a fact that has been confirmed by pigment
extraction and identification in A. levana (Koch, 1991). The reds and
browns could be ommatins, rhodommatin or ommatin D, all of which
have been identified in the wings of Nymphalidae, or it could be
xanthommatin which occurs in butterfly scales as a degradation product
of (unknown) labile ommatins (Butenandt et ah, 1960). In A. levana the
reddish-brown portions of the pattern are also known to incorporate
radiolabeled sulfur (Liidicke and Plesse, 1970) and glucose (Koch, 1985),
which supports the hypothesis that both ommatin D and rhodommatin
may be present in these areas. The yellow of Heliconius is 3-
hydroxykynurenine, (Tocuyama et al., 1967; Brown, 1981; Koch, 1991),
a key metabolite in the tryptophan-ommochrome pathway.

In most species we studied the tyrosine and tryptophan labeled non-
overlapping portions of the color pattern. This suggests the operation of
discrete biochemical and developmental switches for each of these
regions of the wing. To switch a region of the color pattern from yellow
to black in Heliconius , or from red to black in most other species, for

30( 1-2): 1-13, 1991

11

Figure 6. Radiolabel incorporation patterns in the wings of Agraulis vanillae.
Arrangement as in Fig. 1.

12

J. Res „ Lepid.

instance, would have to involve the inactivation of critical enzymes in the
ommochrome pathway and the synthesis and/or activation of melanin-
forming phenoloxidase and its associated proteases. This discrete
separation of label is perhaps nowhere better visible than in H. charitonia
(Fig. 5).

Complex color patterns in butterflies are always composed of a finely
tiled mosaic of monochrome scales. Certain color tints, such as the
“greens” in Pieridae, as well as graduated changes from one color to
another are obtained by locally adjusting the ratios or proportions of
scales of different colors (Nijhout, 1985, 1991). In view of this mosaic
nature of the color pattern we have long suspected that each scale might
only contain a single kind of pigment. Our observations on H. charitonia
tend to support this notion. We have found several instances in which
single yellow scales that developed well within a black region of the wing
were labeled with tryptophan, and also instances in which individual
black scales within a yellow region were labeled with tyrosine. Thus in
this species there appears to be discrete switching of biochemical path-
ways for pigment synthesis at the level of the individual scale cell. Our
autoradiograms do not have sufficient resolution to allow us to determine
whether this is true also in any of the other species we studied.

Overlapping tyrosine and tryptophan labeling was found only in A.
leuana and possibly in P. coenia. This observation suggests that in these
species both melanins and ommochromes can occur in the same scales.
We can, however, not at present exclude the possibility that A. levana
may contain novel pigments that incorporate metabolites of both trypto-
phan and tyrosine, as is the case in the papiliochromes of the Papilionidae
(Umebachi, 1985)

Acknowledgement : This work was supported by a grant from the Duke Univer-
sity Research Council, and, in part, by Grant DCB-8517210 from the National
Science Foundation.

Literature Cited

Brown, K.S. 1981. The biology of Heliconius and related genera. Ann. Rev.
Entomol. 26: 427-456.

Butenandt, A., E. Biekert, H. Kubler and B. Linzen. 1960. Uber Ommochrome. XX.
Zur Verbreitung der Ommatine im Tierreich. Neue Methoden zu ihrer
Identifizierung und quantitativen Bestimmung. Hoppe-Seyler’s Z. Physiol.
Chem. 319: 238-256.

Koch, P.B. 1985. Diehormonale Steuerungdes Saisondimorphismus vonAraschnia
levana L . (Nymphalidae, Lepidoptera). Thesis, Univ. Ulm.

1991. Tryptophan and 3-hydroxykynurenine are hemolymph-carried

precursors of pattern-specific ommatin formation in red wing scales of the
butterfly Araschnia levana L. (Nymphalidae, lepidoptera). Insect Biochem. (in
press).

Linzen, B. 1974. The tryptophan-ommochrome pathway in insects. Adv. Insect
Physiol. 10: 177-246.

Ludicke, M. and D. Plesse. 1970. Die Darstellung des Pigmentwechsels in den

30( 1-2): 1-13, 1991

13

Fliigeln der saisondimorphen Formen von Araschnia levana-prorsa L.
(Nymphalidae) nach oraler Application von [35S]-Natriumsulfat im
Raupenstadium. Z. Naturforsch. 25: 399-406.

Needham, A.E. 1974. The Significance of Zoochromes. Springer Verlag, New York.
Nijhout, H.F. 1980. Ontogeny of the color pattern on the wings of Precis coenia
(Lepidoptera: Nymphalidae). Dev. Riol. 80: 275-288.

1985. The developmental physiology of color patterns in Lepidoptera. Adv.

Insect Physiol. 18: 181-247.

1991. The Development and Evolution ofButterfly Wing Patterns. Smithsonian

Institution Press, Washington.

Tocuyama, T., S. Senoh, T. Sakan, K.S. Brown and B. Witkop. The photoreduction
of kynurenic acid to kynurenic yellow and the occurrence of 3-hydroxy-
kynurenine in butterflies. J. Amer. Chem. Soc. 89: 1017-1021.

Umebachi, Y. 1980. Wing-pigments derived from tryptophan in butterflies. In:
Biochemical and Medical Aspects of Tryptophan Metabolism (O. Hayashi, Y.
Ishimura and R. Kido, eds.), pp. 117-124. Elsevier, New York.

Umebachi, Y. 1985. Papiliochrome, a new pigment group of butterfly. Zool. Sci. 2:
163-174.

Journal of Research on the Lepidoptera

30(1-2):14-18, 1991

Atepa , a new Sonoran Euliini genus (Lepidoptera:
Tortricidae)

Jozef Razowski

Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Slawkowska
17, 31-016 Krakow, Poland.

Abstract. The new genus Atepa is erected to contain three species, two
of which are described as new ( cordobana , sinaloana).

Introduction

Obraztsow ( 1967) transferred Walsingham’s species Tortrix triplagata
described in 1914 from Tabasco, Mexico from Tortricinae to Phaloniinae
(=Cochylini) basing probably on its external habit. However, reexamina-
tion of the type allowed the following correction. In the material kindly
provided by Prof. Dr. Jerry A. Powell, Berkeley I have found two other
congeneric species .described below in a new erected genus. The types of
those species are in the collection of the University of California, Berke-
ley. I would like to express my thanks to Prof. Powell for providing the
Mexican material for study.

Atepa gen. n.

Type-species: Atepa cordobana sp.n.

Forewing weakly expanding posteriorly in male, without costal fold,
rather uniformly broad in female. In forewing all veins separate: M3 -
Cul strongly approached to one another at median cell, R5 to beyond
apex; in hindwing Rr - Ml stalked to middle, M3 - Cul originating in one
point or on a very short stalk. Coloration: ground colour pale, usually
yellowish, pattern brown, consisting of oblique postbasal blotch at
dorsum, slender median fascia atrophying in dorsal part, or interrupted
subdorsally, and subapical blotch. Foreleg without scale tuft.

Male genitalia: Tegumen fairly long, with broad shoulders; uncus long,
slender, well sclerotized; gnathos typical tortricine; socius drooping, with
sharp tip; vinculum well developed, forming short saccus; valva very
long, with costa well developed; sacculus fairly long, folding dorsally;
pulvinus indistinct; transtilla provided with pair of strong distal pro-
cesses; juxta small, with dorsal lobe; aedeagus rather stout; cornuti
present.

Female genitalia: Papilla analis typical of subfamily; sterigma fairly
large, distinctly sclerotized, with lateral arms long, broad basally; ap-
ophyses posteriores very delicate, apophyses anteriores thick, short;
ductus bursae long, in major part sclerotic; corpus bursae small, mem-
branous; signum absent.

Biology: Moth collected in VII (several examples) and VI and X (3
specimens only) at altitudes 800 - 1010 m above sea level.

30(1-2):14-18, 1991

15

Distribution: Mexico (states Veracruz, Tabasco, Sinaloa).

Comments: The supposed autapomorphies of this genus are: strongly
elongate, upcurved base of costa of valva; presence of sclerotic, thorn like
end of valva; sharp termination of socius; presence of slender, spined area
of mid-part of anellus membrane above aedeagus; presence of large,
bulbous base of ductus bursae; semi-coiled broad basal part of ductus
seminalis; long process of eighth tergite connecting the apophysis;
presence of scent organ in base of subgenital segment of female. The new
genus has a distinct position within Euliini and belongs in a group of the
undesribed genera characterised by long, slender uncus. Three species
are known to this date. The generic name is an anagram of Teapa, the
type-locality of triplagata.

Atepa cordobana sp.n.

Alar expanse 11-13 mm; labial palpus about 1.5, orange cream, cream
terminally; remaining parts of head cream, suffused with orange later-
ally; flagellum of antenna pale brownish. Thorax cream, orange anteri-
orly. Forewing costa weakly convex; apex very short, rather rounded;
termen somewhat oblique, tolerably straight. Ground colour cream, some-
what glossy, suffused with orange or ochreous especially among pattern
elements in posterior half of wing. Pattern brown; dorsal blotch slender,
with sharp dorsal prominence at the end, accompanied with dull brown
suffusion at base of dorsum; costa similarly suffused in basal third;
median fascia originating beyond middle of wing where replaced by
ochreous brown, indistinct suffusions; subapical blotch reaching beyond
middle of termen, brown at costa, ochreous in remaining parts. Distinct
ochreous suffusions along pattern edges and small groups of brown scales
in dorsal part of median fascia present. F ringes ochreous cream . Hindwing
gray, with more white fringes.

Variation rather distinct; in dark specimens ground colour strongly
suffused with chreous or brownish yellow, so the whitish, glossy surface
remains in the posterior part of wing only; median fascia ocasionally
complete.

Male genitalia (figs 1 - 4): Base of uncus broad; terminal plate of
gnathos moderate, acute; socius gradually expanding to beyond middle,
tapering terminally; costa of valva with very long upcurved, bent base,
concave in middle, convex subterminally, with apex directed dorsally;
sacculus strongly sclerotized, provided with small prominence before
middle of dorsal edge, reaching beyond middle of valva; disc rather
sparcely hairy; transtilla broad but weakly sclerotized medially and
anteriorly, with sclerotic posterior edge armed with a pair of strong,
spoon like sublateral processes; juxta small, with dorsal prominence
directed distally; aedeagus stout, with well developed coecum penis,
concave in region of caulis, terminating in two broad lobes, membranous
beyond opening for ductus ejaculatorius; cornuti - three groups of short,

16

J. Res . Lepid.

Figs 1-10. 1 - 5 - male and female genitalia of Atepa cordobana sp.n., paratype; 6
- female genitalia of A sinaloana sp.n., holotype; 7, 8 - female genitalia of
A. triplagata (Walsm.); 9, 10 - abdominal scent organ (ventral and lateral
view) of same specimen.

30( 1-2): 14-18, 1991

17

thick, non-capitate spines; elongate area of minute spines in middle of
anellus just beyond zone (in figs 3, 4 connected with aedeagus).

Female genitalia (fig. 5): Lateral arms of sterigma broad, upcurved,
angulate sublaterally; anterior portion of sterigma tubular, with pair of
delicate submedian concavities ventrally; posterior edge deeply concave,
with small central concavity; dorsal wall of sterigma broadly concave in
middle; pair of sublateral, membranous convexities beyond sterigma.
Antrum as broad as anterior part of sterigma, sclerotic, separated from
the latter by means of very short membrane, provided with long ventral
fold armed with numerous spines; corpus bursae small, rounded; a large
dorsal lobe of ductus seminalis extending to the left just beyond corpus
present; in base of the right side of that lobe extends broad ductus
seminalis directed dorso-posteriorly to surround ductus bursae.

Holotype, female: “Cordoba, Mex [ico], Veracruz, VII- 13 - 1966; J. S.
Buckett, M.R. & R.C. Gardner Coll.”, Genitalia Slide 11785. Paratypes,
6 males and 6 females with similiar data but dated VI - 29, VII - 20, 25,
and 28, and 1 specimen collected 6 mi S.E. Rinconada, Vera Cruz.

Atepa sinaloana sp.n.

Alar expanse 11 mm. Head and thorax as in preceding species; ground
colour of forewing more yellow, suffused with ochreous yellow; pattern
yelowish brown, consisting of diffuse, rather broad median fascia and
elongate-trinagular subapical blotch; dorsal blotch ill-defined; apical
portion of wing tinged ochreous. Hindwing greish, with weak ochreous
hue; fringes whiter. Male (with abdomen missing) paler than female,
with ochrous brownish, delicate pattern.

Female genitalia (fig. 6) as in preceding species but with different distal
incisure of sterigma, shorter apophysis anterior and strongly slerotized,
smaller dorsolateral lobe of ductus bursae; base of ductus seminalis also
somewhat different.

Holotype, female: “27 mi E. Villa Union, 800', Sin./aloa/, Mex./ico/, VII-
26- 64; J. Powell, Black & white lights”, G. S. 6035. Paratype, probably
male, with abdomen missing; same label.

Atepa triplagata (Walsingham), comb.n.

Tortrix triplagata Walsingham, 1914, Biologia Centr.-emer., Lepidopt.
Heterocera, 4: 282, pi. 8, flg.22. Holotype, female: “Teapa, Tabasco,
Mexico, III. 18.., (H.H. Smith) Gdm. Slv. 66449”, G.S. 4726 /BM/, in coll.
British Museum (Natural History), London.

Externally very similiar to cordobana and showing similiar pattern
variation. Four specimens from Mexico, Vera Cruz: Fortin de las Flores
collected 7 - 12 VII 1974 by J. A. Chemsak and J. A. Powell before me.
They do not differ from the type in the female genitalia.

Female genitalia (figs 7, 8): Sterigma deeply incised in middle posteri-
orly, with large lateral lobes rounded subterminally; concavity of dorsal
wall slender, distinctly extending beyond middle of distal edge of ventral

18

J. Res. Lepid.

wall; anterior part of sterigma slenderer than in preceding species,
membrane between it and sclerite of ductus bursae longer; the latter
long, curved, expading in anterior part laterally, provided with large fold
of left side; base of ductus seminalis broad; corpus bursae elongate; no
spines in bursa copulatrix.

Scent organ (figs 9, 10): A lateral sack at base of subgenital sternite,
partially entering the preceding segment with transverse, elongate
opening.

Reference

Obraztsov, N.S. 1967. Some apocryphal species of the Tortricinae (Lepidoptera:
Tortricidae). J.N. Y. Ent. Soc.,75(l): 34.

Journal of Research on the Lepidoptera

30(l-2):19-37, 1991

Larval Foodplant and other Effects on Troidine
Guild Composition (Papilionidae) in Southeastern
Brazil

Ana Beatriz B. de Morais* and Keith S. Brown Jr.

Departamento de Zoologia, Institute de Biologia, Universidade Estadual de Campinas, C.P.
6109, Campinas, Sao Paulo 13081 Brazil

Abstract. In two neighboring forest sites (Amarais and Monjolinho) in
Campinas, Sao Paulo, Brazil, with different Aristolochia host plants,
two guilds of the same six troidine swallowtail butterflies showed major
differences in their relative species proportions. Parides agavus repre-
sented over half the adults marked in Monjolinho but only 3% of those
in Amarais; P. proneus included over half the adults in Amarais, versus
3% in Monjolinho. These contrasting proportions remained essentially
unaltered during a major macroclimatic anomaly (El Nino-Southern
Oscillation) in 1982-1983, which however significantly altered the
proportions of three other species, and permitted the occasional P.
neophilus to become established and common in both sites. Troidine
juveniles (a total of 1802) were located and followed during two years
in Monjolinho. Introduction of Aristolochia melastoma, abundant in
Amarais, into Monjolinho caused larvae and adults of the melastoma-
specialist P. proneus to become more common there. No hostplant
effects could be clearly identified, however, to account for the contrast-
ing abundances of P. agavus in the two sites, which seem better
correlated with dense vegetation structure and reduced understory
illumination, affecting the microclimate in Monjolinho.

Introduction

Most species of the cosmopolitan butterfly family Papilionidae (swal-
lowtails) occur in the tropics (Slansky, 1972; Collins & Morris, 1985).
Larvae of some Papilio species are pests of Citrus (Rutaceae), and other
swallowtails attack cultivated Lauraceae, Magnoliaceae, Annonaceae
and Umbelliferae (Feeny et al., 1983; Scriber, 1984). The tribe Troidini
(in the subfamily Papilioninae) specializes on plants in the family
Aristolochiaceae (Slansky, 1972; Scriber, 1984); Neotropical records
from any other plant family in nature have not been confirmed, though
may be possible since some Magnoliaceae are accepted by both Battus
and Parides in the laboratory (Brown and Klitzke, unpublished).
Aristolochiaceae species are also mostly tropical (Hoehne, 1942) and
most contain aristolochic acids and benzylisoquinoline alkaloids, chemi-
cals with noted pharmacological properties, including abortifascient and
irritant of the gastro-intestinal tract of vertebrates (Hoehne, 1942; Von
Euw et al., 1968; Chen & Zhu, 1987). Recently it has been shown that
secondary compounds other than aristolochic acids and alkaloids are
found in the roots, stems and leaves of south Brazilian Aristolochia

* Present address: Departamento de Biologia, CCNE, Universidade

Federal de Santa Maria, Faixa de Camobi Km 09, Santa Maria RS, 97119-900 Brazil

20

J. Res. Lepid .

species: lignans (Rucker et al., 1981), diterpenes (Habib & El-Sebakhy,
1981; Lopes et al., 1987; Lopes & Bolzani, 1988; Luiz et al., 1988), and
essential oils (Leitao et al., 1988). Few studies have reported the effects
of these plants or compounds on phytophagous insects (but see Miller &
Feeny, 1983, 1989); very few herbivores other than Troidini are found
feeding onAristolochia (Rausher & Feeny, 1980;Rausher, 1981:6; Brown
et al., 1981). Troidine larvae store toxic chemicals from their hostplants
and pass them to the adults (Von Euw et al., 1968; Rothschild et al., 1970;
Brown et al., 1981, 1991; Urzua et al., 1983, 1987; Urzua & Priestap,
1985), that are unpalatable to vertebrate predators (Brower & Brower,
1964; Brower et al., 1967; Rothschild, 1973; Brower, 1984; Chai, 1986,
1988). Both larvae and adults of troidines show aposematic coloration
with Mullerian convergence, and serve as models for mimicry rings that
include other Papilionidae and further Lepidoptera such as Pericopinae
and Castniidae; some adults also resemble unpalatable ithomiine butter-
flies, and large Pepsis wasps (Brown, 1988; Aiello & Brown, 1988).

The natural history of American troidine communities has been stud-
ied in Brazil (Moss, 1919; D’Almeida, 1922, 1944, 1966; Brown et al.,
1981; Otero & Brown, 1986) and western Mexico (Spade et al., 1988).
Brown et al. ( 198 1) discussed five sympatric species and their foodplants
in the region of Campinas, interior of Sao Paulo state, and suggested the
possible influence of host plant palatability and other factors in causing
the different species proportions of troidines in different localities. The
species showed varying acceptance for oviposition, and tolerance for
larval feeding on the available Aristolochia hostplants in the field and
laboratory (Brown et al., 1981; Otero & Brown, 1986). Other factors such
as parasitism, predation, competition, phenology, adult resources, physi-
cal environment and evolutionary history were implicated in the variable
composition of the guilds.

In this paper we seek to expand the data-base of Brown et al. ( 1981) by
addressing the following questions raised there:

(1) Which environmental factors are important in determining the
relative abundance of troidine species within the community? Is this
abundance predictable? Is the hostplant involved?

(2) To what extent do different troidine species make use of different
hostplants in different habitats? What are the causes and consequences
of this?

In doing so, we hope to achieve a better understanding of the processes
involved in community composition and dynamics at the Troidini/
Aristolochia interface.

Materials and Methods

Studies were undertaken in a 3-ha well-watered forest garden (arboretum),
“Monjolinho,” and a larger (25-ha) tall Eucalyptus forest with a native middle-
and understory, “Amarais,” both in the Fazenda Santa Elisa of the Campinas
Agronomical Institute, Sao Paulo (22°54’ S., 47°05’ W., about 650 m elevation),
from 1980 to 1984. The two areas are separated by 1.0 km of cultivated fields

30(1-2): 19-37, 1991

21

(Brown et al., 1981; Figure 1). Both contained many troidines during most of the
year, in microhabitats resembling those occupied by the same species in fully
natural systems (partly disturbed moist forests with abundant flowers).

The vegetation of Monjolinho (Figure 1), where juveniles were followed in 198 1-
1983, was mostly ornamental shrubs and trees, both native and introduced, with
a dark and humid but well-developed understory in many parts of the woods, and
an adjacent sunny garden with abundant flowers. The two major Aristolochia
hostplants occurred in separated patches, consisting of 80 accessible A. elegans
Masters and 42 A. esperanzae O. Kuntze, each population including individual
plants ranging from small herbs to large climbers; five small plants of A. arcuata
Masters were also present in Monjolinho. A single plot of 20 m2 of A. melastoma
Manso (about 40 rooted stolons) was experimentally introduced in 1981 from
nearby Amarais, where A. melastoma and A. arcuata were abundant, and A.
elegans and A. esperanzae absent. All these Aristolochia plants in Monjolinho
were regularly examined for troidine juveniles, found on all of them at least once
during this period (five high A. elegans vines remained inaccessible).

The troidine populations occurring in both study areas were Battus polydamas
(L., 1758), Parides proneus (Hiibner, 1825), P. b. bunichus (Hiibner, 1822), P.
agavus (Drury, 1782), P. anchises nephalion (Godart, 1819), and P. neophilus
eurybates (Gray, 1852), as described in Brown et al. (1981). In both areas, adults
were censused by weekly capture-mark-release-recapture periods, one hour in
Monjolinho and two in Amarais (enough to visit at least twice all areas usually
frequented by the butterflies), in the warmer part of the day, marking with
indelible soft point pens of various colors (Sharpie, Sanford Corp.), with minimal
handling to reduce trauma. Each individual was given a unique number, and note
was taken of the species, sex, wing wear, location and time of day, and any
unusual condition or behavior of the insect.

Two or three times a week, Troidini juveniles were sought in Monjolinho by
thorough examination of leaves and stems of foodplants. Each egg or larva
discovered was recorded and left on the original plant, to be observed on each
subsequent visit until disappearance. The general yellow-orange color and
similar size of the eggs of P. bunichus , P. agavus and P. a. nephalion precluded
sure identification as to species until the end of the first instar, when the
characteristic pattern of tubercles on the abdominal segments became conspicu-
ous (Brown et al., 1981); the five larval instars were recognized by the width of
their head capsules (Brown et al., 1981; Otero & Brown, 1986).

The minimum number of days that each larva stayed in each instar was
determined as the period between the first and last observation in that instar.
Means of these “minimums” were compared between species of Troidini and
Aristolochia , using Student’s t-test and one factor ANOVA. A chi-square test (or
the Fisher exact probability when expected frequencies were five or fewer) was
used to compare, on each species of hostplant, the numbers of larvae that
survived to the third instar, in relation to those that “disappeared” before the end
of the second instar (Sokal & Rohlf, 1981).

Results and Discussion

COMPARISONS OF THE SPECIES PROPORTIONS OF ADULT
TROIDINES IN MONJOLINHO AND AMARAIS

Up through mid- 1982, the species proportions of the troidine commu-
nities in Monjolinho and Amarais remained basically the same as those

22

J. Res . Lepid.

Figure 1 . Map of yonjolinho, showing the location of individual Aristolochia plants followed in this study, and spatial relationship to Amarais.

30(l-2):19-37, 1991

23

reported in Brown et al. (1981) (Table 1). Thus, while P. proneus
represented over half the adults marked (1324 of 2420) in Amarais, and
P. agavus represented only 3% (82) of that community, the abundance of
the two was precisely reversed in nearby Monjolinho, where over half the
adults marked (190 in 350) were P. agavus , and P. proneus represented
but 3%. In Amarais, nine times as many P. bunichus and twice as many
B. polydamas were marked as P. a. nephalion', in Monjolinho, the latter
species was commoner than either of the former two. P. neophilus was
only occasionally seen in both areas. By next-day recapture sessions,
capture rates in Amarais were found to sample about 25-33% of the
various Parides present and 10% of the Battus in two hours, while in
Monjolinho, essentially all troidines present could be recorded in one
hour (see Brown et al., 1981: 217). Exchanges of adults between the two
sites occurred regularly (Table 1), indicating the possibility of choice of
habitat and preferred resources in each.

In continued marking in the two localities from mid- 1982 to mid- 1984
(Table 1), these patterns remained relatively stable, though P. proneus
doubled its abundance in Monjolinho, P. a. nephalion doubled its propor-
tion in Amarais, B. polydamas became only half as common there and
disappeared from Monjolinho, P. bunichus became rarer in Amarais, and
the previously occasional visitor P. neophilus suddenly became a promi-
nent resident (more than 20% of both communities), which it continues
to be up until today (1991). All these changes were statistically signifi-
cant (Table 1), with P. agavus remaining near its former proportions in
both sites, and P. proneus stable in Amarais.

These alterations corresponded to a marked change in the rainfall
patterns in 1982-1983, linked to a major El Nino-Southern Oscillation
episode in the equatorial Pacific region (Rasmussen & Wallace, 1983;
Barber & Chaves, 1983; Caviedes, 1984; Canby, 1984; Glynn, 1988). In
south Atlantic manifestations of this climatic anomaly, over 2000 mm of
rain fell in both years in Campinas, with very few dry months, represent-
ing an event never before recorded (only one other year since 1890
showed over 2000 mm — 1970, with 2564 — and it was flanked by two
fairly dry years). Large areas of Amarais were flooded as a major new
stream cut across the area, draining a nearby plateau. Elsewhere in
Campinas, long-term flooding in the interior of old forests killed trees
hundreds of years old, suggesting the rarity of such events. Such
continuously rainy weather, promoting high humidity and extensive
plant growth, could have reduced activity and population levels of B.
polydamas , whose adults prefer dry, sunny habitats (Rausher, 1979;
Brown et al., 1981), and whose gregarious larvae are especially prone to
diseases which prosper under high humidity (Moss, 1919). These condi-
tions are, however, typical of habitats preferred by P. neophilus in central
Brazil and the Amazon Basin and Andean foothills. Such humidity may
also have favored P. a. nephalion in Amarais, and been unfavorable to P.
bunichus , a dry-area species (Brown et al., 1981; Otero & Brown, 1986).

Table 1. Adult individuals marked and recaptured in two adjacent Troidini communities during two periods, 1981-1984.

24

^ co
£ 'co

c *

§ E

lT <

Q.

CO „_L
O CO >

£ O T5

,± ©
c o

03 C
C/) o

CM

— ra CO

TO ^ ^ 2

<5 S2.S ?
2 Em

£

o' © o
CM

LL I CO
£ r- O)
00 r-

o)Jp
— E

CL

CC

O

CD

DC

JZ

2 ef

03 CD °

0 03

E

1 CO
“ 03
^ =3
.9 T3

4E cd
C o

03 E

C/5 o

co

03

° §
. cm
00 °g

CD 03
T3 P

-I E

v*_ 05

° O

6 CXJ
00 00
03 03

CO

©

'o

©

Q.

CO

O 00 LO

TO

V,

CO

E

10

00

T—

-or

co

T"

CM

CM

CO

CO

jx.

TO

CO

O

TO

r-

f-

19

130

33

32

228

TO , TO *

c c I

O CD 03 00 CD T"

^ t~ CM

CO CO CO

CM

00 CM CM 03 CD

CM CM CM t- CM

03 xt CM CM 03

t- O i- CO CO O

LO CM -i-

03 *
C *

CO CM CO LO CO
LO t— CM

CM LO •'rf LO CO t~

LO CM LO ^ 00 O

03 CM

00 LO O CO CO
LO CO

03 ^ O CM CM
00 CM CM CO 00
r- 00 N

TO

CO

6

I

Q

a

TO

£

to

00

TO

§ .§

TO O -C

D) c a

Q. -Q TO TO TO

TO TO TO TO ^

P ^ P P

C C: c; C;

TO TO TO TO

a. a. a, a.

1 1
■§ TO

o

TO

O ©

c n

m|
CO c

CT S

to jg

o "
5 g5
co .==

+- TO

>, 3
Q ^
C O
© JZ
03 C

■B o

LO

O

00

o

CM

c

LO

LO

CD

CM

X

T—

CO

CM

TO

CO

TO

>-,

03

’M”

CD

T-

o

JD

TO

LO

o

~o

0

CL

TO

0

TO

'CO

O

CD

LO

o

0

CO

03

LO

LO

O

CO

c

E

CO

o

<

c

c

O)

TO

O

'to

0

c

TO

o

ID

c

©

Q,

©

CL

CL

o

2

©

£

c

©

©

2

©

JD

CL

.E i

©

a

c

©

•1=f o
§0
E v

E CL

O II

o 11

CO

©

1 .

o

c

d

TO

V

©

Q_

'a

II

©

a.

*

TO

■*-/

if)

C

CO

O

TD

d

CZ

V

o

CL

12

II

CO

-t— >

*

TO

„

O

E

c

CO

©

_n

.9

c

03

TO

C

'to

'co

o

03

c

CO

II

c

TO

.9

C

a

TO

0

Js£

o

o

CO

E

J„ Res . Lepid.

30(l-2):19-37, 1991

25

Thus, many of the small but significant changes in the community of
adults during 1982-1984 could be tentatively correlated with a major
macroclimatic change in that period. The contrasting abundances of P.
proneus and P. agauus in the two sites remained evident, however, and
reasons for this persistent difference in the two adjacent communities
must be sought in other factors.

CHARACTERISTICS OF THE JUVENILE COMMUNITY IN
MONJOLINHO

Between December 1981 and November 1983, 1802 troidine eggs and
larvae were found and followed in Monjolinho (Table 2); 668 of these
disappeared before the end of the first instar without being identified as
to species. Only three pupae were encountered, in April 1982 and
February-March 1983, all ofP. a. nephalion, parasitized, and near the A.
elegans on which they had been followed as larvae. While this could
merely emphasize the usual dispersal of the larva far from the hostplant
before pupation (thus escaping detection in the census periods), it also
reflects the great mortality of juveniles in these species (Figure 2). Only
3% of all juveniles registered reached the final instar.

The numbers of juveniles found varied through the year, sometimes
but not always showing peaks corresponding to, or a couple of months
before the peaks in the adult population; both were markedly reduced in
winter (June-August), when most of the population were in pupal
diapause (D’Almeida, 1966; Brown et al., 1981). One fifth instar P. a.
nephalion was followed in mid-winter, with little rainfall and tempera-
tures between 5-25 °C, for eight weeks without showing notable growth,
nor disappearing.

In 1981-1982 (Table 3), Parides proneus was the rarest of the juveniles,
and P. a. nephalion was the commonest, with 50% more records than P.
agauus (whose adults were four times as common as those of nephalion ,
Table 1) and three times as many as P. hunichus (equally common as
adults). The period 1982-1983 was marked by large numbers of P.
neophilus (Table 3), which led to an initial problem in separating its
juveniles from those of P. a. nephalion , resolved only in mid-1983.
Juveniles of P. a. nephalion and P. neophilus (probably about equal in
number) both greatly outweighed those of P. agauus , while P. proneus
juveniles increased significantly, presumably due to the introduction of
their foodplant (A. melastoma ), becoming just as frequent as those of P.
hunichus and the diminished B. polydamas. These three were still less
than half as common as P. agauus ; unidentified eggs and first instar
larvae increased from 24% to 38% of all juveniles. These juvenile
proportions probably reflect both the abundance of adult females (Table
1) and of acceptable oviposition sites, as well as increased mortality of
eggs and first instars due to heavy rain (Blau, 1980).

The relative deficiency of P. agauus juveniles in relation to the adult
abundance (Table 3) could be due to their undetected presence on the

Table 2. Total eggs and larvae of Troidini encountered on A ristolochia species in Monjolsnho, XII-1981 to XI-1983

26

W T-

CO

CD

TO

00

cm

CO OJ

OJ

CO

ID

CO

CD

o

OJ T-

T—

CO

T—

T“

CD

00

T—

•T—

CO o

TO

in

o

CO

ID

00

T-

CXI

OJ

TO

S

to

8

to

.TO

■TO

O

o

§

TO

CM

TO

5

s

TO

.TO

TO

.TO

■§

O

8

.TO

TO

TO

a

TO

&

TO

.TO

-5

o

o

§

.TO

TO

TO

TO

CTO

TO

TO

.TO

•C

TO

O

8

.TO

m

co

U_ C

I — CD 3
CO CTO O

O B c

X CO CD

TO 00 CM TO
CM

CO

CO o

OJ i-

CO

03

OJ

CM

CM

CM

o

CO

OJ

CM

CO

o

T-

T-

T~

CM

o

CO

CM

T_

T-

\—

TO

T—

o

00

1— •

03

TO

00

TO

OJ

CD

00

T—

o

TO

CO

-r-

CO

CM

LO

T—

7-

TO

1-

h-

T-

CO

7 —

m

CD

TO

OJ

If)

CO

in

7—

CD

l-

o

00

o>

o

in

1-

CM

TO

CO

o

CO

to

CM

TO

o

T—

CO

CO

T-

CO

co

T—

TO

CO

00

7—

1 —

CO

or

03

r--

o

TO

CM

CM

03

TO

TO

CD

CD

CO

o

03

03

o

o

T—

T-

to

TO"

LD

CM

CM

TO

CM

TO

CO

CM

T-

co

-

00

r-

CO

to

CD

OJ

OJ

03

TO

o

OJ

to

T“

CD

^ n in n

T- OJ U) T-

03 CD
OJ T-

« S

TO .r-

?|
8 TO

Cl-Q

TO TO
TO TO
"TO ID

03 CO
00

CTO TO
TO TO
TO TO
TO TO
^3 ^5

TO TO TO TO
0. 0, CL cl

CD

TO

TO TO

I ■§.

1°
TO

TO

OJ

00

CD

TO TO .3
TO TO 7=
■£- TO "5

TO
TO

- 5

S- TO

O 13

TO sL

CD

c

V.'

TO

c

TO

TO

o

TO

CD

TO

+

a

§

■8

§

"C:

"TO

TO

TO

TO

a.

CL

QQ

<

H

O

o

o

G)

OJ

TO

TO ^
CL TO

11

D)

C

CD

C

TO

CL

JC

o

TO

TO

C

O

TO

_TO

"E

TO

>

<

I—

o

J. Res. Lepid.

a 0 = eggs, 1 to 5 = first to fifth instar larvae; sT = subtotal for each species on each hostplant.

PARIDES PRONEUS PARIDES ANCHISES NEPHALION

30(1-2): 19-37, 1991

27

r-~

^r

cc

<

I—

c n
z

cc
<
I —

c n
z

ONIAIAdflS %

ONIAIAdflS %

ONIAIAdflS %

Figure 2. “Survival” of larvae (first to fifth instars) of various Parides species (and of indistinguishable P. a. nephalion and P. neophilus) on
different Aristolochia species in Monjolinho. The probabilities represent survival to the molt to the third instar (see text), which is always
higher in the right-hand graph of the pair; note that more eggs are always placed on A. elegans, the species with the greatest biomass
(left-hand graphs), except for P. a. nephalion (upper right pair), but this plant always gave lower survivorship than the alternative
hostplant, for all species of troidines.

28

J. Res. Lepid.

Table 3. Juveniles of Troidini discovered in Monjolinho, 1981-1983.

1 981 -mid 1 982 % all mid 1 982-end 1 983 % all Signifi-
Species Eggs Larvae Total juv. Eggs Larvae Total juv. cance*

Battus polydamus

79

42

121

23

66

12

78

6

**★

Parides proneus

0

6

6

1

43

28

71

5

***

Parides b. bunichus

4

40

44

8

29

41

70

5

ns

Pariedes agavus

31

66

97

18

98

70

168

12

-

P. anchises nephalion

43

99

142

26

6

7

13

n

ns

P. neophilus eurybates

0

0

0

0

76

47

123

9r

***

P. (anchises nephalion 0
+ neophilus eurybates)

0

0

0

88

248

336

24 J

Unidentified Parides 101
species (less proneus)

28

129

24

408

131

539

38

***

Totals

258

281

539

100

814

584

1398

100

* Difference between census periods, compared against P. agavus (most abundant
species as adults). The clustered eggs of p olydamas reduce the validity of the
statistical test, which nevertheless would not lose significance. Non-significance for
P. a. nephalion occurs for all proportions between 0.49 and 0.97 in the mixed batch
(336 juveniles) with P. neophilus. ns = not significant, *** = P<0.001 .

highest A. elegans vines; P. a. nephalion preferred the small, easily
inspected plants of A. esperanzae for oviposition, where the larvae were
easily discovered and followed.

DIFFERENTIAL SURVIVORSHIP OF TROIDINE JUVENILES ON
THE VARIOUS FOODPLANTS

Troidine larvae usually stay on the hostplant nearest to where they
hatch from the egg as long as suitable leaves are available, only leaving
it to seek other foodplants in later instars if necessary (Rausher, 1981;
Brown et al., 1981). Even so, 97% of all juveniles observed in 1981-1983
(1748 of the 1802) disappeared before attaining the fifth instar; larger
larvae were sometimes found on neighboring plants ( Aristolochia and
others), but in general “disappearance” meant death.

Arthropods (ants and small spiders) were the principal larval preda-
tors in Monjolinho, also observed in the laboratory by Brown et al. (1981)
(see also Watanabe, 1976, 1981; Hirose et ah, 1980; Rausher, 1981; and
Feeny et al. , 1985). Attack on later instars by vertebrates, observed by all
these authors, was not recorded in Monjolinho; fifth instar mortality in
P. anchises nephalion and P. neophilus occurred through parasitoids,
also recorded for eggs, larvae and pupae by Moss (1919) and the above
authors. Parasitism and cannibalism by congeners, indicated by most of
these studies, were not quantified in Monjolinho.

The “quality” of the hostplant can make important contributions to
juvenile survivorship and mortality (Scriber & Feeny, 1979; Price et al.,

30(l-2):19-37, 1991

29

1980; Scriber & Slansky, 1981; Courtney, 1981; Williams et al., 1983;
Rhoades, 1985). Troidine juveniles were encountered in different num-
bers and proportions on their four available foodplants (Table 2). Parides
agavus predominated on A. elegans , on which P. neophilus was more
common than P. a. nephalion, but the last species was dominant on A.
esperanzae. P. proneus predominated on A. melastoma but was absent
from A. arcuata , on which B. polydamas was also absent, possibly due to
its very small biomass in Monjolinho (five small, scattered plants).

Table 4 shows the mean minimum duration (in days) for each instar of
Parides on each Aristolochia species in Monjolinho. The apparent gen-
eral tendency for more rapid development on A. esperanzae in relation to
A. elegans was significant only for the first larval instar of P. agavus , and
in any case may have been due to the softer, perhaps more nutritious
leaves on the generally younger plants of A. esperanzae. The lower means
of P. proneus on A. melastoma did not reach significance. Not enough
larvae were marked and followed continuously to permit simultaneous
comparisons of larval survivorship on the four species of hostplants.

Figure 2 shows the percentage of larvae of each Parides species
surviving on each hostplant (in pairs), from first to fifth instars (fourth
in P. proneus ). While P. agavus did equally well on both A. elegans and
A. esperanzae , larvae ofP. bunichus andP. a. nephalion survived better
to the third instar on A. esperanzae (with marginal significance, P= 0.06
and 0.09), andP. neophilus and the mixed (—1:1) lot ofP. a. nephalion and
P. neophilus had significantly higher survivorship on A. esperanzae than
on A. elegans. This might indicate a better “quality” for the former plant,
at least in Monjolinho where it is represented mostly by young, soft
plants. P. proneus did much better (P= 0.003) on its natural foodplant A.
melastoma than on the far more abundant A. elegans , where females
deposited the majority of their eggs.

These different reactions of larvae of each species to the various
foodplants chosen by their mothers are better seen in Figure 3, in which
the ability to survive is emphasized as percent of total larvae remaining
on each plant species as the larvae grow through each instar. The figure
reveals the superior relative survival ofP. proneus on A. melastoma over
other hosts, of P. agavus and younger P. neophilus on A. esperanzae
(though the more abundant A. elegans was preferred for oviposition, see
Table 2 and Figure 2), and of B. polydamas and P. bunichus on A. elegans.
Some patterns are easier to discern in Figure 3 than in Figure 2; note
especially, in the later instars of P. bunichus , P. a. nephalion and P.
neophilus , the reversion of the lower survival patterns on A. elegans in
the early instars.

A preliminary investigation of hostplant chemistry (Morais, 1986)
showed a general absence of aristolochic acids in the leaves of A. elegans
and A. esperanzae , confirming results of Hussein & El-Sebakhy (1974)
and Urzua & Priestap (1985) (these substances are common in the roots
of these same plants — see Priestap et al., 1971 — and are known to have

30

J. Res. Lepid.

Table 4. Mean number of days for each larval instar of Parides species on different
Aristolochia species in the field (Monjolinho).

Species and

Instar

A. elegans
x± s

(n)

A. esperanzae A. melastoma “t” or
x±s (n) x±s (n)ANOVAa

P. proneus egg

6.60 ±2.93

(19)

5.67± 1.15

(3)

3.67 ±1.51 (6)(all 3) NS

(elxmelP*

first larval

7.37 + 6.67

(8)

6.67 ± 2.34 (6)

NS

second larval

5.83 ± 4.07

(6)

3.50 ±2.14 (8)

NS

P. bunichus egg

6.71 ±2.63

(7)

6.0 ±2.51

(8)

NS

first larval

7.97 ±4.98

(31)

6.29 ± 1.72

(17)

6.50 ±3.54 (2)

NS

second larval

5.50 ±3.25

(8)

5.50 ±2.17

(10)

NS

third larval

3.60 ±2.61

(5)

3.60 ± 2.07

(5)

NS

fourth larval

4.0 ±2.64

(3)

3.50 ±2.12

(2)

5.33 ±1.53 (3)

NS

fifth larval

8.0 ±5.57

(3)

-

P. agavus egg

7.08 ±3.30

(73)

5.91 ±1.90

(22)

*

first larval

8.46 ± 4.68

(116)

5.91 ±3.14

(32)

**

second larval

5.60 ±2.72

(42)

5.18 ±3.52

(11)

NS

third larval

4.95 ±2.76

(19)

5.17 ± 3.19

(6)

NS

fourth larval

4.73 ± 2.53

(11)

3.0 ±0.0

(2)

-

P. anchises egg

6.40 ±2.51

(5)

first larval

6.75 ±5.68

(4)

6.0 ±3.54

(8)

NS

second larval

6.20 ± 3.77

(5)

5.70 ± 3.33

(10)

NS

third larval

5.0 ±2.83

(2)

3.0 ±0.0

(3)

-

fourth larval

6.40 ± 1.95

(5)

5.0 ±2.52

(7)

NS

fifth larval

10.83 ± 5.12

(6)

4.75 ± 1.98

(8)

★

A. arcuata

P. neophilus egg

5.63 ±3.48

(38)

4.44 ± 1.81

(9)

5.67 ±2.31 (3)

NS

first larval

6.29 ±4.28

(55)

5.62 ±2.20

(8)

4.67 ± 2.08 (3)

NS

second larval

5.50 ±3.39

(14)

4.0 ±0.0

(2)

-

third larval

10.0 ±4.32

(4)

-

fourth larval

4.25 ±3.30

(4)

2.0 ±0.0

(2)

-

a NS = Not significant, * = p<0.05, *

*=p<0.01.

effects on troidine larval performance — Miller & Feeny, 1989).
Benzylisoquinoline alkaloids (El-Sebakhy & Waterman, 1984), also
important in larval feeding (Miller & Feeny, 1983, 1989) were variable,
but most abundant in A. elegans. Several other secondary chemicals
could be detected in the leaf fractions, different in each species. Extracts
of Aristolochia leaves were seen to turn very dark with great rapidity in
the presence of air, indicating phenolic polymerization; the compounds
responsible for this were not isolated.

It is clear that evaluation of the possible influences of leaf secondary
chemistry on the variation in larval performance (Figures 2-3) and
preference for different species must await more detailed investigations
and bioassays (these authors, P. E. R. dos Santos & C. Klitzke, in
progress).

Since the larvae were not all observed simultaneously in the field, it is
not possible to extrapolate the relative global performance of the species

30(1-2): 19-37, 1991

31

over two years on each plant to draw any conclusions about competitive
ability. Maps of the use of individual A. elegans and A. esperanzae (based
on Figure 1) by each troidine species showed appreciable concentration
and overlap of species on certain plants over the period. These were often
individuals of greater biomass or permanence in the system. Globally, A.
esperanzae was occupied predominantly by final instar larvae of P. a.
nephalion , and A. elegans by fifth instar Battus polydamas (Figure 3).
Although these data might indicate a reasonable potential for competi-
tive interaction among the species on both plant populations in Monjolinho,
the results did not permit rejection nor sustaining of the hypothesis of
resource partitioning due to competitive interaction, sufficient to affect
species proportions in the community, distinguishable from the effects of
tolerance, oviposition preference, growth rates and other factors men-
tioned.

Finally, a density-independent factor causing juvenile death, espe-
cially on young A. esperanzae plants near the entrance to Monjolinho
(Figure 1), was human interference during the observations and in
periodical cutting of these “weedy” plants by the caretakers. These plants
were always greatly preferred by P. a. nephalion juveniles, that could
have suffered excessive reductions in the larval stage due to this factor,
leading to lower adult abundance (Table 1).

INFLUENCE OF THE PHYSICAL ENVIRONMENT ON P. AGAVUS

While variations in species proportions of five of the six troidines in
Campinas have been tentatively correlated with macroclimate and
larval foodplants, no factor yet discussed was adequate to explain the
consistent great difference in abundance of P. agavus between Monjolinho
and Amarais (Table 1).

Microclimate is also important to troidines, including in the laboratory
where it can determine adult activity, larval survivorship and pupal
diapause (Brown etal., 1981). Because P. agavus is more restricted to the
dense humid Atlantic forests than the other members of the community
in Campinas, it could be more sensitive than these to light structure and
humidity. Microclimatic measurements in the understories of Monjolinho
and Amarais showed no significant difference in humidity, but apprecia-
bly more illumination in the latter site, especially in the midday hours
(Morais, 1986). Thus, the different vegetation structures in the two areas
affect the quality and quantity of light in the understory, recognized as
an important physical factor for insect species (Warren, 1985), including
troidines (Moss, 1919; Rausher, 1981; Brown et al., 1981).

The influence of vegetation/light structure versus larval foodplant on
the proportion of P. agavus in the community might be tested in
additional sites, such as ones with the foodplants of Amarais and the dark
understory of Monjolinho, or with a brightly lit understory and the plants
of Monjolinho. The latter habitat, often with abundant A. elegans and/or
A. esperanzae , is very frequent in the disturbed interior region of Sao
Paulo; P. bunichus and B. polydamas are predominant in these sites,

PLANTS OCCUPIED PLANTS OCCUPIED PLANTS OCCUPIED

32

J. Res . Lepid.

PARIDES PRONEUS
N= 43 35 20 12 5 0

P. NEOPHILUS EURYBATES
N= 77 52 14 11 5 3

P. BUNICHUS
N= 45 35 23 9

BATTUS POLYDAMAS

N= 110 82 39 48 22 12

EGG/LARVAL INSTAR

P. ANCHISES NEPHALION
N= 69 55 47 22

LARVAL INSTAR

Figure 3. Percentage of eggs and larvae (by instar) of each troidine species found on
each different Aristolochia hostplant. Eggs and first-instar larvae of P.
bunichus, P. agavus and P. a. nephalion could not always be surely
separated from each other, and therefore are omitted.

30(1-2): 19-37, 1991

33

with P a. nephalion and P. neophilus variably present in more humid
riverine areas, andP. agavus at best very occasional. In contrast, six well-
studied sites in the interior of Sao Paulo with reduced understory
illumination, from 400 to 1300 m elevation and with a wide variety of
Aristolochia foodplants, all showed predominance of P. agavus in the
troidine community. A notable example was the dense riverine tangle
800 m NNW of the main colonies of the Horto Florestal de Sumare (Area
“D” of Figure 1A in Brown et ah, 1981: 201, 217), which contained
principally P. agavus , mobile between this preferred area and the little-
chosen, more open woods of the Horto, where it represented only 6% of the
community — a situation almost parallel to that of Monjolinho and
Amarais. While correlation does not mean causation, the circumstantial
evidence suggests a strong influence of understory light structure and
small effect of foodplant species on the abundance of P. agavus in the
community.

CONCLUSIONS, SUMMARY AND PERSPECTIVES

In relation to the questions posed in the Introduction, this study
showed that:

(1) Larval hostplant effect on the proportions of six species in the
troidine guilds in southeastern Brazil was evident only in the case of
Parides proneus , whose abundance corresponded to that of its principal
hostplant, Aristolochia melastoma.

(2) An unusual period of increased and continuous rainfall, associated
with a very strong El Nino episode in 1982-1983, was accompanied by
significant reduction in abundance of Battus polydamas in two commu-
nities, and a decrease of Parides hunichus and increase of P. anchises
nephalion in one of them. The same factor may have encouraged a range
expansion ofP. neophilus , to become the second most common species in
both communities, in which it was previously only an occasional visitor.

(3) A denser structure of the vegetation, with corresponding reduction
of light in the understory, correlated with the abundance of P. agavus in
diverse habitats with a variety of potential hostplants available.

(4) All of the species except P. proneus could feed and grow on all four
common Aristolochia hostplants present in the two habitats studied;
some differences in larval performance detected may be due more to
chance or the age and size of the foodplants used, rather than to plant
chemistry or general “quality.”

(5) No evidence was found for partitioning of hostplants or their parts
between the species (except for P. proneus), even though this phenom-
enon may exist in natural systems where Aristolochia is rarer (Moss,
1919). This suggests that Troidini are usually quite opportunistic in
foodplant usage within the Aristolochia available, as found by Spade et
al. (1988) in western Mexico.

Present work is directed towards details of foodplant chemistry, includ-
ing controlled feeding experiments with fractions and isolated com-

34

J. Res. Lepid.

pounds, especially from the least acceptable A. galeata (- brasiliensis )
and A. gigantea (Brown et al., 1981), not present in Monjolinho or
Amarais. The importance of parasitism and predation will also be
evaluated experimentally with split cohorts of juveniles on various
foodplants in different habitats. Further data may come from experi-
ments and observations over long periods in natural habitats where the
Troidini and their foodplants are rarer, and perhaps thereby subjected to
large effects from small variations in various environmental factors.

Acknowledgments. We are grateful to Iria B. Baldessari and Cecilia T. Teradaira,
who first worked on Aristolochia/Troidini interactions in Monjolinho, for location
of many of the plants and earlier juveniles. Marlies Sazima and Marcia Siqueira
identified the plants and encouraged us to study their biology, chemistry and
systematics. Paul Feeny stimulated work on the Troidini swallowtails; he read and
criticized an earlier version of this paper, along with Mark Scriber, Marc Rausher,
and two unidentified reviewers, whom we thank for their helpful comments.
Angelo Pires do Prado, Hermogenes F. Leitao Filho, Mohamed E. E. M. Habib and
W. W. Benson made valuable suggestions on the part of this work which was
presented as a Master’s thesis to the Post-Graduate Course in Ecology of the
Universidade Estadual de Campinas. Silvana A. Henriques helped in the labora-
tory studies, and Jose R. Trigo in these and also in the field. The Instituto
Agrondmico de Campinas gave permission to work on the Fazenda Santa Elisa;
Hermes Moreira de Souza labored many years to make Monjolinho a pleasant place
for Troidini and for scientific research. CAPES and FAPESP provided fellowships
to ABBM, and the CNPq to KSB, for research on chemical aspects of insect/plant
interactions.

Literature Cited

Aello, A. &K. S. Brown Jr., 1988. Mimicry by illusion in a sexually dimorphic, day-
flying moth, Dysschema jansonis (Lepidoptera: Arctiidae: Pericopinae). J. Res.
Lepid. 26: 173-176.

Barber, R. T. & F. P. Chaves, 1983. Biological consequences of El Nino. Science 222:
1203-1210.

Blau, W.S., 1980. The effect of environmental disturbance on a tropical butterfly
population. Ecology 61: 1005-1012.

Brower, L. P., 1984. Chemical defense in butterflies. IN: Vane-Wright, R.I. & P.R.
Ackery (Eds.), The Biology of Butterflies, pp. 109-134. Symp. Royal Ent. Soc.,
London, 11. Academic, London.

Brower, L.P. & J.V.Z. Brower, 1964. Birds, butterflies, and plant poisons: a study
in ecological chemistry. Zoologica (New York) 49: 137-159.

Brower, L.P., Cook, L.M., & H.J. Croze, 1967. Predator response to artificial
Batesian mimics released in a neotropical environment. Evolution 21: 11-23.
Brown Jr., K.S., 1988. Mimicry, aposematism, and crypsis in Neotropical
Lepidoptera: the importance of dual signals. Bull. Soc. Zool. France. 113: 83-
101.

Brown Jr., K.S., Damman, H.J. & P. Feeny, 1981. Troidine swallowtails (Lepidoptera:
Papilionidae) in southeastern Brazil: natural history and foodplant relationships.
J. Res. Lepid. 19: 199-226.

30(1-2): 19-37, 1991

35

Brown Jr., K.S., Trigo, J.R., Francini, R.B., Morais, A.B.B. & P.C. Motta, 1991.
Aposematic insects on toxic hostplants: coevolution, colonization and chemical
emancipation. IN: Price, P.W.; Lewinsohn, T.M.; Fernandes, G.W. & W.W.
Benson (eds.), Plant-Animal Interactions: Evolutionary Ecology in Tropical
and Temperate Regions, pp. 375-402. John Wiley & Sons, New York.

Canby, T.Y., 1984. El Nino’s ill wind — global weather disaster. Nat. Geogr. 165:
144-184.

Caviedes, C.N., 1984. El Nino 1982-1983. Geogr. Rev. 74: 267-290.

Chai, P., 1986. Field observations and feeding experiments on the responses of
rufous-tailed jacamars ( Galbula ruficauda) to free-flying butterflies in a
tropical rain forest. Biol. J. Linn. Soc. 29: 161-189.

— , 1988. Wing color of free-flying neotropical butterflies as a signal learned by

a specialist avian predator. Biotropica 20: 20-30.

Chen, Z.L. & D.Y. Zhu, 1987. Aristolochia alkaloids. IN: Brossi, A. (ed.), The
Alkaloids: Chemistry and Pharmacology, vol. 31, pp. 29-65.

Collins, N.M. & M.G. Morris, 1985. Threatened Swallowtail Butterflies of the
World: the IUCN Red Data Book. IUCN, Gland, Switzerland, 401 pp.

Courtney, S.P., 1981. Coevolution of Pierid Butterflies and their cruciferous
foodplants. III. Anthocharis cardamines (L.) survival, development and
opposition on different hostplants. Oecologia (Berlin) 51: 91-96.

D’Almeida, R.F., 1922. Melanges Lepidopterologiques. I. Etude sur les Lepidopteres
du Bresil. R. Friedlander & Sohn, Berlin, 226 pp.

, 1944. Estudos biologicos sobre alguns lepidopteros do Brasil. Arq. Zool. (Sao

Paulo) 4: 33-72.

, 1966. Catalogo dos Papilionidae Americanos. Sociedade Brasileira de

Entomologia, Sao Paulo, 366 pp.

El-Sebakhy, N. & P.G. Waterman, 1984. (-) - (R,R’)-7’-0-Methylcuspidaline from the
leaves of Aristolochia elegans. Phytochemistry 23: 2706-2707.

Feeny, P., Blau, W.S. & Kareiva, P.M., 1985. Larval growth and survivorship of the
black swallowtail butterfly in central New York. Ecol. Monogr. 55: 167-187.

Feeny, P., Rosenberry, L. & M. Carter, 1983. Chemical aspects of oviposition
behavior in butterflies. IN: Ahmad, S. (ed.), Herbivorous Insects: Host Seeking
Behavior and Mechanisms, pp. 27-76. Academic: New York.

Glynn, P.W., 1988. El Nino-Southern Oscillation 1982-1983: nearshore population,
community, and ecosystem responses. Ann. Rev. Ecol. Syst. 19: 309-345.

Habib, A.M.M. & El-Sebakhy, N.A., 1981. Ett£-kaurane- 16a, 17-diol and (-)-cubebin
as natural products from Aristolochia elegans . Pharmazie 36: 291-294.

Hirose, Y. , Suzuky, Y. , Takagi, M. , Hiehata, K. , Yamasaki, M. , Kimoto, H. , Yamanaka,
M., Iga, M. & K. Yamaguchi, 1980. Population dynamics ofthe citrus swallowtail,
Papilio xuthus Linne (Lepidoptera: Papilionidae): mechanisms stabilizing its
numbers. Res. Pop. Ecol. 21: 260-285.

Hoehne, F. C., 1942. Aris toloquiaceas . Flora Brasflica 15: 3-141.

Hussein, F.T. & N.A. El-Sebakhy, 1974. A phytochemical investigation ofthe leaves
of Aristolochia elegans. Planta Medica 25: 310-314.

LeitAo, G.G., Souza, F., Kaplan, M.A.C., Damasceno, L.P. & A.A. Craveiro, 1988.
Composifao qufmica de especies do genero Aristolochia. I. Oleos essenciais.
Resumos Encontro Brasileiro-Alemao de Qufmica de Produtos Naturais, Rio de
Janeiro: 88.

Lopes, L.M.X. & V.S. Bolzani, 1988. Lignans and diterpenes of three Aristolochia
species. Phytochemistry 27: 2265-2268.

36

J. Res . Lepid.

Lopes, L. M.X., Bolzani, V.S. & L.M.V. Trevisan, 1987. Clerodane diterpenes from
Aristolochia species. Phytochemistry 26: 2781-2784.

Luiz, V., Bolzani, V.S., Lopes, L.M.X. & L.M.V. Trevisan, 1988. Alguns constituintes
quimicos de caules de Aristolochia elegans. Mast. Resumos X RESEM, Joao
Pessoa, Parafba: 59.

Miller, J.S. & P. Feeny, 1983. Effects of benzylisoquinoline alkaloids on the larvae
of polyphagous Lepidoptera. Oecologia (Berlin) 58: 332-339.

, 1989. Interspecific differences among swallowtail larvae (Lepidoptera:

Papilionidae) in susceptibility to aristolochic acids and berberine. Ecol. Entom.
14: 287-296.

Morais, A.B.B. De, 1986. InterafaoentreTroidini(Lep.: Papilionidae) e Aristolochia
(Aristolochiaceae) em Campinas, Sao Paulo. Master’s thesis, Instituto de
Biologia, UNICAMP, Campinas, SP. 153 pp.

Moss, A.M., 1919. The Papilios of Para. Novit. Zool. 26: 295-319.

Otero, L.S. & K.S. Brown Jr., 1986. Biology and ecology of Parides ascanius
(Cramer, 1775) (Lep., Papilionidae), a primitive butterfly threatened with
extinction. Atala 10-12: 2-16.

Price, P.W., Bouton, C.E., Gross, P., McPheron, B.A., Thompson, J.N. & A.E. Weis,
1980. Interactions among three trophic levels: influence of plants on interactions
between insect herbivores and natural enemies. Ann. Rev. Ecol. Syst. 11:41-65.

Prestap, H.A., Ruveda, E.A., Mascaretti, O.A. & V. Deufoleu, 1971. The aristolochic
acids from Aristolochia argentina Gris, and Aristolochia esperanzae O. Kuntze.
Anales Assoc. Quim. Argent. 59: 245-250.

Rasmussen, E.M. & J.M. Wallace, 1983. Meteorological aspects of the El Nino/
Southern Oscillation. Science 22: 1195-1202.

Rausher, M.D., 1979. Larval habitat suitability and oviposition preference in three
related buterflies. Ecology 60: 503-511.

, 1981. Hostplant selection by Battus philenor butterflies: the roles of

predation, nutrition and plant chemistry. Ecol. Mongr. 51: 1-20.

Rausher, M.D. & P. Feeny, 1980. Herbivory, plant density and plant reproductive
success: the effect of Battus philenor on Aristolochia reticulata. Ecology 61: 905-
917.

Rhoades, D.F., 1985. Offensive-defensive interactions between herbivores and
plants: their relevance in herbivore population dynamics and ecological theory.
Amer. Natur. 125: 205-238.

Rothschild, M. 1973. Secondary plant substances and warning colouration in
insects. IN: Van Emden, H.F. (ed.), Insect/Plant Relationships, pp. 59-83.
Symp. Royal Ent. Soc., London, 6.

Rothschild, M., Reischstein, T., Von Euw, J., Aplin, R. & R.R.M. Harman, 1970.
Toxic Lepidoptera. Toxicon 8: 293-298.

Rucker, G., Langmann, B. & N.S. Siqueira, 1981. Inhaltsstoffe von Aristolochia
triangularis. Planta Medica 41: 143-149.

Scriber, J.M., 1984. Larval foodplant utilization by the world Papilionidae (Lep.):
latitudinal gradients reappraised. Tokurana 6-7: 1-50.

Scriber, J.M. & P. Feeny, 1979. Growth of herbivorous caterpillars in relation to
feeding specialization and to the growth form of their food plants. Ecology 60:
829-850.

Scriber, J.M. & F. Slansky Jr., 1981. The nutritional ecology of immature insects.
Ann. Rev. Entom. 26: 183-211.

30( 1-2): 19-37, 1991

37

Slansky Jr., F., 1972. Latitudinal gradients in species diversity of the New World
swallowtail butterflies. J. Res. Lepid. 11: 201=217.

Sokal, R.R. & F. J. Rohlf, 1981. Biometry, 2nd. ed. Freeman, San Francisco, 859 pp.

Spade, P., Tyler, H. & J.W. Brown, 1988. The biology of seven troidine swallowtail
butterflies (Papilionidae) in Colima, Mexico, J. Res. Lepid. 26: 13-26.

Urzua, A. & H. Priestap, 1985. Aristolochic acids from Battus polydamas. Biochem.
Syst. Ecol. 13: 169-170.

Urzua, A., Rodrigues, R. & B. Cassels, 1987. Fate of ingested aristolochic acids in
Battus polydamas. Biochem. Syst. Ecol. 13: 169-170.

Urzua, A., Salgado, G., Cassels, B.K & G. Eckhardt, 1983. Aristolochic acids in
Aristolochia chilensis and the Aristolochia- feeder Battus archidamas
(Lepidoptera). Collect. Czech. Chem. Comm. 48: 1513-1519.

Von Euw, J., Reichstein, T. & M. Rothschild, 1968. Aristolochic acid-I in the
swallowtail butterfly Pachlioptera aristolochiae (Fabr. ) (Papilionidae). Israel J.
Chem. 6: 659-670.

Warren, M.S., 1985. The influence of shade on butterfly numbers in woodland
rides, with special reference to the wood white Leptidea sirtapis. Biol. Conserv.
37: 147-164.

Watanabe, M., 1976. A preliminary study on population dynamics of the swallowtail
butterfly, Papilio xuthus L. in a deforested area. Res. Pop. Ecol. 17: 200-210.

, 1981. Population dynamics of the swallowtail butterfly Papilio xuthus L.,

in a deforested area. Res. Pop. Ecol. 23: 74-93.

Williams, K.S., Lincoln, D.E. & P.R. Ehrlich, 1983. The coevolution of Euphydryas
chalcedona butterflies and their larval hostplants. II. Maternal and hostplant
effects on larval growth, development, and food-use efficiency. Oecologia
(Berlin) 56: 330-335.

Journal of Research on the Lepidoptera

30(l-2):38-44, 1991

An Annotated List of Moths Recorded at Florissant
Fossil Beds National Monument, Colorado.

Charles V. Coveil Jr.

Department of Biology, University of Louisville, Louisville, Kentucky 40292, U.S.A.

Abstract. The author lists, with dates of capture, 201 species of moths
collected or seen and positively identified at Florissant Fossil Beds
National Monument, Colorado, from 1976 to 1988.

During the summers of 1976 and 1977, Dr. F. Martin Brown collected
moths by UV light at the Florissant Fossil Beds National Monument in
Teller County, Colorado. Some specimens (indicated in the text) were
also collected by National Park Service personnel Ellen W. Elder, R.
Hodgson, and John R. Lancoa. Moth specimens were spread, labelled,
and sent to the American Museum of Natural History, New York City, for
identification. Most of these specimens are now in the collection at the
Florissant Fossil Beds National Monument. A small portion of them —
particularly the microlepidoptera — are in the collection of the Univer-
sity of Louisville, Kentucky.

On June 18, 1988, the author, with Patrick Lawler and Troy Payne
assisting, collected at blacklight near the Monument headquarters. All
of the 1988 records included here resulted from that collection, which is
incorporated into that of the University of Louisville, Kentucky.

The Florissant Fossil Beds National Monument occupies slightly more
than nine square miles - more than 5,500 acres - in Teller County. For
the most part it was grazing land with a few hundred acres devoted to
tourist attractions of fossil stumps and pits in which to dig - for a fee.
Structurally the park occupies a valley following branches of Grape
Creek, flanked with low granite hills covered by an open stand of
ponderosa pines. In a few of the northward facing valleys are small
clusters of firs and spruces. The greater area of the park is grassland,
some of it marshy, the rest dry.

The names of the species collected are numbered and constructed in
accordance with Hodges et al. (1983) although Grammia geneura (no.
8180) follows a more recent combination (Ferguson, 1984). Dates of
collection follow the names. Most of the moths are medium and larger
species - particularly Pyralidae, Geometridae, and Noctuidae - and were
determined by F.H. Rindge, E.L. Quinter, and the late A.B. Klots of the
American Museum of Natural History, New York City, to whom I am
grateful for their help. Identifications of the 1988 material were done by
D.C. Ferguson, R.W. Hodges, and R.W. Poole, all of the U.S. National
Museum of Natural History, and J.A. Powell, University of California,
Berkeley. To them I am very grateful. Special thanks are due to F.

30(l-2):38-44, 1991

39

Martin Brown and Grace Kemper for providing valuable information for
this manuscript.

While the 201 species listed here represent only a small portion of the
total Lepidoptera fauna of the Monument, and a few specimens are still
undetermined, I hope these data will provide a useful baseline of
information upon which a more complete survey can later be undertaken.

OECOPHORIDAE

980 Ethmia discostrigella (Chambers). 6-18-88

1011 Antaeotricha schlaegeri (Zeller). 6-18-88

GELECHIIDAE

1968+ Gnorimoschema sp. 6-18-88

MOMPHIDAE
1423+ Mompha sp. 6-18-88

TORTRICIDAE

3014 Eucosma ridingsana (Robinson). 8-20-76

3015 E. fernaldana (Grote). 8-20-76

3030 E. ragonoti (Walsingham). 7-30-76

3070 E. Franclemonti Powell. 6-18-88

3502 Croesia albicomana (Clemens). 6-18-88

3608 Argyrotaenia coloradana (Fernald). 7-30-76

3609 A. provana (Kearfott). 6-18-88

3618 A. dorsalana (Dyar). 6-18-88

3638 Choristoneura fumiferana { Clemens). 7-30-76. Spruce Bud-

worm adults were also observed by the author near the Park
headquarters on ponderosa pines, 7-16-84 and 7-23-84.
3644a C. lambertiana (Busck), ssp . ponderosana Obraztsov. 7-27-
76; 7-30-76. A series of 7 highly variable specimens.

3647 Cudonigera houstonana (Grote). 6-18-88

COCHYLIDAE

3794 Hysterosia aureoalbida Walsingham. 6-18-88

PYRALIDAE

4781 Petrophila avernalis (Grote). 6-18-88

4919 Prorasea fernaldi Munroe. 6-21-76; 7-30-76

5017 Loxostege cereralis (Zeller). 6-21-76; 7-13-77; 6-18-88

5034 Pyrausta signatalis (Walker). 6-18-88

5060 P. subsequalis (Guenee), ssp. plagalis Haimbach. 7-30-76;
8-20-76

40

J. Res. Lepid.

5347 Crambus awemellus McDunnough. 8-20-76

5352 C. richusalis (Hulst). 9-9-76

5388 C.dimidiatellus Grote. 7-14-76; 7-30-76; 8-20-76
5403 Agriphila vulgivagella (Clemens). 7-30-76; 8-20-76

PTEROPHORIDAE

6109 Platyptilia carduidactyla (Riley). Questionably this spe-
cies. 8-20-76

GEOMETRIDAE

6290 Itame loricaria (Eversmann). 7-14-76

6304 7. bitactata (Walker). 7-30-76

6339 Semiothisa transitaria (Walker) (subspecies or undescribed
sibling species). 6-18-88

6346 S. unipunctaria (W.S. Wright). 6-21-76

6373 S. delectata Hulst. 6-18-88

6381 S. colorata Grote. 6-18-88

6383 S. pervolata (Hulst). 6-18-88

6393 S. yavapai (Grossbeck). 6-18-88

6396 S. neptaria (Guenee). 6-21-76

6415 S. cyda (Druce). 6-18-88

6448 Glena nigricaria (Barnes & McDunnough). 6-21-76
6463e Stenoporpia pulmonaria (Grote), ssp. vicaria Rindge. 8-23-

77

6470 S. macdunnoughi Sperry. 6-21-76; 7-14-76; 6-18-88
6474 S. ex celsaria (Strecker). 6-21-76

6631 Galenaria lixarioides McDunnough. 6-21-76; 6-18-88

6635 Vinemina opacaria (Hulst). 6-18-88

6640a Biston betularia (Linnaeus), ssp. cognataria (Guenee). 6-
18-88

6684 Drepanulatrix bifilata (Hulst). 7-13-77

6737 Euchlaena tigrinaria (Guenee). 6-18-88

6761 Pero occidentalis (Hulst). 6-18-88

6819 Metanema inatomaria Guenee. 6-18-88

6852 Epiplatymetra coloradaria (Grote & Robinson). 7-30-76

6857 Lychnosea helveolaria (Hulst). 7-30-76

6860 Neoterpes trianguliferata { Packard). 6-21-76

6865 Caripeta aequaliaria Grote. 6-18-88

6865.1 Caripeta suffusata Guedet. 7-30-76

6866 C. interalbicans Warren. 6-21-76

6875 Snowia montanaria Neumoegen. 6-18-88

6876 Nemeris speciosa (Hulst). 7-14-76

6879.1 Meris alticola Hulst. 7-30-76

6883 Destutia excelsa (Strecker). 6-18-88

6895 Lambdina vitraria (Grote). 6-18-88

6905 Nepytia swetti Barnes & Benjamin. 7-30-76
6964 Tetrads cachexiata Guenee. 6-18-88

7000 Sabulodes niveostriata (Cockerell). 7-30-76

30(l-2):38-44, 1991

41

7006 Enypia griseata Grossbeck. 7-13-77

7065 Cheteoscelis bistriaria (Packard). 7-13-77; 7-14-76
7096 Lobocleta quaesitata (Hulst). 7-14-76

7140 Cyclophora nanaria (Walker). 7-16-76

7170 Scopula luteolata (Hulst). 7-30-76

7191 Dysstroma formosa (Hulst). 7-13-77

7199 Eulithis propulsata (Walker). 8-20-76

7216 Plemyria georgii Hulst. 9-24-76

7220 Thera latens Barnes & McDunnough. 8-20-76

7236 Hydriomena renunciata (Walker). 6-18-88

7243 H. albimontanata McDunnough. 7-13-77

7264 H. morosata Barnes & McDunnough. 7-13-77
7275 H. similaris Hulst. 6-30-76; 7-30-76

7301 Entephria multivagata (Hulst). 8-20-76

7328 Perizoma custodiata (Guenee), form “ polygrammata ”
(Hulst). 6-21-77

7347 Stamnodes formosata { Strecker). 7-14-76; 7-30-76; 6-18-88

7385 Xanthorhoe alticolata Barnes & McDunnough. 8-20-76

7386 X. defensaria (Guenee). 6-18-88

7391 X. dentilinea Barnes & McDunnough. 7-30-76; 8-20-76
7405 Loxofidonia acidaliata (Packard). 8-20-76

7445 Horisme intestinata (Guenee). 7-13-77

7513 Eupithecia nabokovi McDunnough. 7-30-76

7529 E. coagulata Guenee. 7-30-76

7539 E. multistrigata (Hulst). 7-13-77

7540 E. perfusca (Hulst). 6-21-76

7594 E. anticaria Walker. 6-21-77; 6-18-88

7641 Lobophora montanata Packard. 6-18-88

LASIOCAMPIDAE

7687 Phyllodesma americana (Harris). 6-18-88

7702 Malacosoma californicum (Packard). 7-30-76

SATURNIIDAE

7726 Coloradia luski Barnes & Benjamin. 6-21-77; 6-18-88

SPHINGIDAE

7822 Smerinthus cerisyi Kirby. 6-18-88

7828 P achy sphinx mode sta { Harris). 6-21-76; 6-18-88

7894 Hyles lineata (Fabricius). 7-13-77; 7-22-78

NOTODONTIDAE

7900 Clostera brucei (Hy. Edwards). 6-21-76

7924 Odontosia elegans i Strecker). 6-18-88

7931 Gluphisia septentrionis Walker. 6-18-88

42

J. Res. Lepid.

7940 Furcula scolopendrina (Boisduval). 6-21-76; 6-18-88
8014 Oligocentria pallida (Strecker). 6-21-77

ARCTIIDAE

8043 Eilema bicolor (Grote). 7-30-76

8050 Crambidia impura Barnes & McDunnough. 9-10-77

8083 Lycomorpha grotei (Packard). 8-6-83 (R. Hodgson)

8089 Hypoprepia miniata (Kirby). 7-13-77

8125 Holomelina fragilis (Strecker). 7-16-77

8126 Leptarctia californiae (Walker). 5-11-77

8144 Turuptiana permaculata (Packard). 7-14-76; 7-30-76; 7-28-

83 (E.W. Elder); 6-18-88

8180 Grammia geneura (Strecker). 6-21-76

8218 Aemilia ambigua (Strecker). 6-21-76; 6-18-88

8261 Ctenucha cressonana Grote. 8-1-83 (E.W. Elder)

NOCTUIDAE

8622 Synedoida inepta Hy. Edwards. 6-21-76; 6-18-88
8819 Catocala pura Hulst. 9-24-76

8907 Autographa biloba (Stephens). 6-21-77

8924 Anagrapha falcifera (Kirby). 6-20-77; 6-18-88

9111 Therasea augustipennis (Grote). 8-23-77

9193 Raphia frater Grote. 6-21-76; 6-18-88

9195 R. coloradensis Putnam-Cramer. 6-18-88

9205a Acronicta lepusculina Guenee, ssp. felina (Grote). 6-21-76;
6-18-88

9348 Apamea amputatrix (Fitch). 7-30-76

9356 A. spaldingi (Smith). 6-18-88

9358 A. parcata (Smith). 6-21-76

9364a A. finitima Guenee, ssp. cerivana (Smith). 6-21-77

9365 Agroperina lateritia (Hufnagel). 7-14-76

9370 A. conradi (Grote). 7-30-76; 8-20-76

9374 Protagrotis niveivenosa (Grote). 8-20-76

9382 Crymodes devastator (Brace). 7-30-76; 8-20-76

9383 C. longula (Grote). 9-6-77

9459 Amphipoea senilis (Smith). 9-24-76

9511 Hydroecia medialis Smith. 9-5-77

9537 Aseptis dilara (Strecker). 6-18-88

9601 Pseudanarta flavidens (Grote). 7-30-76

9609 P. exasperata Franclemont. 9-6-77

9610 P. perplexa Franclemont. 9-3-77

9617 Pseudanthoecia tumida (Grote). 8-19-76

9656 Platyperigea extima (Walker). 7-20-76; 9-17-76; 6-18-88

9681 E lap hr ia festivoides (Guenee). 6-18-88

9813 Achytonix epipaschia (Grote). 8-23-77

9883 Homoglaea carbonaria (Harvey). 9-24-76

10001 Lomilysis discolor (Smith). 8-20-76

10085 Oncocnemis terminalis Smith. 9-10-77

10093 O. levis Grote. 8-20-76

43

30(l-2):38-44, 1991

10190a Cucullia speyeri Lintner, ssp. dorsalis Smith. 8-20-76
10194a C. intermedia Speyer, ssp. Cinderella Smith. 6-21-76; 7-14-
76

10206 Copicucullia antipoda (Strecker). 6-21-76

10224 Discestra mutata (Dod). 6-21-77

10280 Polia purpurissata (Grote). 7-16-77

10290 P. obscura (Smith). 6-21-77; 6-18-88

10307 Lacinipolia lilacina (Harvey). 7-30-76

10313 Papestra brenda (Barnes & McDunnough). 6-18-88

10325 Miselia glaciata { Grote). 6-20-77; 6-18-88

10326 M. variolata (Smith). 6-21-76

10372a Lacinipolia anguina (Grote), ssp. larissa (Smith). 6-21-76
10373 L. incurva (Smith). 6-21-76; 7-14-76; 6-18-88.

10379 L. umbrosa (Smith). 7-30-76

10381 L. naevia (Smith). 6-21-76; 7-14-76

10394 L. vicina (Grote). 7-14-76; 8-20-76

10401 L. lepidula (Smith). 6-21-76; 6-18-88

10406b L. olivacea (Morrison), ssp. megarena (Smith). 7-14-76

10416 L. illaudabilis (Grote). 6-18-88

10431c Faronta diffusa (Walker), ssp. neptis (Smith). 6-21-76; 6-
18-88

10436 Aletia oxygala (Grote). 7-30-76; 8-20-76
10438 Pseudaletia unipuncta (Haworth). 9-23-76; 9-24-76
10446 Leucania multilinea Walker. 7-14-76; 7-30-76
10449b L. insueta Guenee, ssp. megadia Smith. 6-21-76; 6-18-88
10523.1 Tholera undescribed sp. 8-23-76

10524a Nephelodes minians Guenee, ssp. tertialis Smith. 7-30-76;
8-20-76

10528 N. carminata (Smith). 8-23-77

10541 Homothordes reliqua (Smith). 7-30-76

10610 Neleucania patricia (Guenee). 7-14-76

10616 Trichorthosia parallela (Grote). 7-30-76

10617 T. tristis (Barnes & McDunnough). 7-14-76

10651a Agrotis venerabilis Walker, ssp. arida (Cockerell). 9-3-77; 9-
5-77

10659 A. volubilis Harvey. 6-21-76; 7-14-76; 6-18-88

10670 Feltia jaculifera (Guenee). 8-20-76

10671 F. hudsoni Smith. 8-20-76

10702 Euxoa divergens (Walker). 6-21-76; 7-14-76; 7-30-76

10731 E. auxiliaris (Grote). 6-21-76; 7-14-76

10755 E. declarata (Walker). 8-20-76

10758 E. flavidens (Smith). 8-20-76

10764 E. stigmatalis (Smith). 8-23-77

10780 E. comosa (Morrison) (questionably this species). 9-5-77

10805 E. tessellata (Harris). 7-13-77

10824 E. brevipennis (Smith). 9-24-76

10826 E. idahoensis (Grote). 7-30-76

10830 E. quadridentata (Grote & Robinson). 9-8-76

10833 E. olivalis (Grote). 7-30-76

10849 E. dodi McDunnough. 8-23-77

44

J. Res. Lepid.

10850 E. infracta (Morrison). 8-20-76

10853 E. arizonensis Lafontaine. 8-20-76

10854 E. servita (Smith). 7-30-76

10858 E. cooki McDunnough. 8-20-76

10860 Euxoa sp., near perolivalis (Smith). 6-21-77

10861 E. ridingsiana (Grote). 8-20-76; 9-17-76

10882 Richia parentalis (Grote). 8-20-76

10915 Peridroma saucia (Hiibner). 6-21-77

10977 Setagrotis atrifrons (Grote). 9-3-77

10992 Paradiarsia littoralis (Packard). 6-21-76
11041 Abagrotis placida (G rote). 8-24-77

11062 Eutricopis nexilis Morrison. 6-1-77

11068 Heliothis zea (Boddie). 9-3-77

11072 H. phloxiphagus Grote & Robinson. 8-25-77
11103 Schinia persimilis (G rote). 7-14-76

Literature Cited

Ferguson. D.C. 1984. Two new generic names for groups ofHolarctic and Palearctic
Arctiini (Lepidoptera, Arctiidae). Proc. Ent. Soc. Washington 86 (2): 452-459.
Hodges, R.W. etal. 1983. Check list of the Lepidoptera of America north of Mexico.
E.W. Classey Ltd. & Wedge Entomological Research Found., London. 284 p.

Journal of Research on the Lepidoptera

30(l-2):45-81, 1991

Invited Paper

Speciation: A Review of Concepts and Studies with
Special Reference to the Lepidoptera

Michael M. Collins1

Research Associate, Invertebrate Zoology Section, The Carnegie Museum of Natural
History

PART I: Concepts
Introduction

Darwin introduced his concept of speciation through natural selection
as the central process in evolution, not just as a process within popula-
tions of organisms but also as the ultimate origin of biological diversity.
And while modern genetics has provided an understanding of the overall
mechanism for evolution, from the molecular level to population biology,
the genetic changes during speciation remain the subject of intense study
and controversy.

The evolutionary significance of speciation for sexually reproducing
organisms such as Lepidoptera is the origin of reproductive isolation,
since only after the cessation of gene flow can related populations
genetically diverge toward separate evolutionary fates. The genetics of
speciation is thus a more narrowly defined topic than overall genetic
differences among closely related species. Many points of debate have
centered around this distinction between the genetics of speciation and
the genetics of species differences (Templeton, 1982).

Three important and general findings in population biology over the
last three decades relate to theories and models of speciation. First,
sexually reproducing organisms contain surprisingly high levels of
genetic variation within and among populations. This result calls into
question early concepts of limits on population divergence which held
that homogenizing gene flow and the supposed adaptive value of highly
integrated gene systems must maintain similar gene frequencies among
populations. Second, various parts of the genome evolve discordantly
such that morphology, enzyme variation, and ecological adaptations, for
example, may be poor predictors of reproductive isolation. Third, repro-
ductive isolation is neither an intrinsic nor inevitable result of overall
genetic divergence, but may arise only under special circumstances of
population structure and selection.

These findings reveal the limitations of traditional taxonomy based
solely on morphological criteria (Byers & LaFontaine, 1982; Collins,
1984; Gall & Sperling, 1980; Hafernik, 1982; Lorkovic, 1985; Mayr &
Ashlock, 1991; Oliver, 1972, 1978, 1979ab, 1980; Sperling, 1987; Wake,
1981; Zink, 1988), including male genitalia in Lepidoptera (Porter and
Shapiro, 1990). As an expedient in classification and identification,
Lepidoptera field guides customarily treat all species as discrete units of

bailing address: 11901 Miwok Path, Nevada City, CA 95959

46

J. Res. Lepid.

equal taxonomic rank, each with distinctive morphological and ecological
characters. Evolutionarily “old” species may have accumulated many of
these differences either before or after the origin of reproductive isola-
tion, thus creating the impression that many classes of taxonomic
characters change concordantly during the speciation process. The
evolutionary biologist necessarily concentrates not on these so-called
“good” species, but on taxa which lack discrete boundaries with related
entities such as overlapping phenotypic variation or incomplete repro-
ductive isolation - i.e. supposed examples of speciation in progress.
Taxonomic decisions are especially difficult to make for allopatric popu-
lations whose reproductive isolation cannot be determined directly, or for
morphologically distinct taxa which hybridize in nature. Assigning
taxonomic rank in these cases should be secondary to collecting and
evaluating evidence relating directly to genetic compatibility and repro-
ductive isolation, and other factors regulating (or potentially regulating)
gene exchange between the taxa.

A current trend in population biology is to view the component popu-
lations and demes of a species as genetically diverse, more or less
independent ecological and evolutionary units (Murphy & Ehrlich,

1984) . Speciation models must distinguish between the overall, inevi-
table tendency of separate populations of a given taxon to diverge
genetically, and the origin and significance of reproductive incompatibil-
ity among related populations. Although this question is usually phrased
in terms of the origin of “reproductive isolating mechanisms”, this may
not be (as discussed in section 3 below) the proper conceptual context for
understanding speciation.

Practical difficulties arise because speciation cannot be studied di-
rectly. The rate of speciation in sexual animals is too slow to be observed
and the complex interaction of population-level genetics with the natural
environment prevents a realistic laboratory simulation of speciation.
(Severe artificial selection on interbreeding populations in the lab has
improved incipient reproductive isolation [Thoday & Gibson, 1962; Rice,

1985] but these experimental results are difficult to extrapolate to
speciation in nature). Three basic approaches can be used to identify the
genetic changes accompanying speciation, especially those concerned
with reproductive compatibility among closely related taxa:

1.) We can study genetic differentiation among populations and recog-
nized subspecies within a species. If we assume that such intraspecific
differentiation represents incipient speciation we can extrapolate the
changes accompanying speciation. This assumption is open to the criti-
cism that speciation may be a rare and unique population genetic
process. Arguments concerning the limitations of the subspecies concept
(Arnold, 1983,1985; Brittnacher et al.,1978; Forbes, 1954; Futuyma,
1986, ch. 4; Gillham, 1956; Hammond, 1985,1990; Wilson & Brown,
1953) are a part of this criticism.

30(l-2):45-81, 1991

47

2) . We can look back in time and catalog the observed differences among
taxonomically very similar species, assuming that they represent the
result of recent speciation. However, we cannot know if divergence in a
given trait occurred before or after speciation. Comparisons can be made
in methods 1) and 2) among populations with respect to population size,
geographic distribution and degree of geographic isolation, rates of gene
flow, ecological associations, etc. - factors affecting population genetic
processes.

3) . We can employ laboratory and natural hybridization to study the
genetics of traditional taxonomic characters, life history traits, and
ecological adaptations. Experimental hybridization is especially reveal-
ing because it is also a functional assay of reproductive and developmen-
tal compatibility of the parental taxa as expressed in the hybrid genome.
Yet, only certain organisms-happily including many Lepidoptera-lend
themselves to this kind of experimental manipulation. Unfortunately,
perhaps in the quest for perfect specimens, experimental hybridization
is often omitted from taxonomic studies, even when comparative mate-
rial is reared in the lab. When hybrids are obtained, quantitative data on
fertility, viability, developmental rates, and diapause should always be
recorded.

I summarize in this paper major advances in speciation theory from
various viewpoints. My purpose is not to evaluate this entire body of
work; comprehensive reviews of speciation theory include Barton ( 1989);
Barton and Charlesworth (1984); Bush (1975, 1982); Futuyma (1986);
Otte and Endler (1989); Templeton (1981, 1982, 1989); other papers are
cited in the text. With this background I will then discuss representative
studies of speciation in Lepidoptera.

Lepidoptera have served as subjects for many important evolutionary
studies (Vane-Wright & Ackery, 1984). Much of the underlying natural
history literature is the result of amateur studies and I hope to encourage
the continued contribution of amateurs, especially in mutually reward-
ing cooperation with professional biologists. I have endeavored to
present this paper to a broad range of readership.

Species Concepts

1. The Biological Species Concept dates back to the synthesis by
Dobzhansky (1937) of Darwinian theory, classical genetics, and math-
ematical models of population genetics into a genetic theory of evolution.
As currently applied, the “biological species” (Dobzhansky, 1970; Mayr,
1963) is composed of populations contributing to a gene pool, united by
actual or potential gene exchange, adapted to a unique range of ecological
niches, and “protected” from disruptive interbreeding with related spe-
cies by means of reproductive isolating mechanisms. This concept recog-
nizes the considerable phenotypic and genetic variability of the popula-
tions comprising a species, in contrast to the “typological” or morphologi-
cal species characterized by an idealized, uniform phenotype. Species

48

J. Res. Lepid.

are thought to arise in allopatry, either through intervention of a
physical barrier or by colonization, when cessation of gene exchange
allows genetic divergence. According to this view, speciation is most
likely when populations are small, such as newly founded populations at
the periphery of a species range. Here the effects of genetic drift (reduced
genetic variability due to random “sampling” of the parental gene pool)
and inbreeding could produce novel genotypes in the founding population
(Carson, 1968; Carson & Templeton, 1984; Mayr, 1982; Templeton,
1980). The interaction of these genotypes and their subsequent recombi-
nants with a novel selection regime in a new environment may lead to a
“genetic revolution” - a new, stable equilibrium in gene interaction -
which may accompany a shift to a new ecological niche. Incipient
reproductive isolation, such as altered mating times, may develop inci-
dentally to these genetic and ecological adaptations.

Upon secondary sympatry, the two differentiated populations may
interbreed but their hybrid offspring may be subvital, sterile or other-
wise unfit. Overall genetic divergence in allopatry can reduce the fitness
of hybrids by disrupting embryo and larval development, diapause, or
adult development (postzygotic isolation). Selection would then favor the
improvement of prezygotic (premating) isolating mechanisms which
would reduce the gametic waste of interspecific matings. Selection
cannot, of course, improve postzygotic isolation since hybrids are less fit.
Morphologically distinct taxa which interbreed in nature do not chal-
lenge the Biological Species Concept but serve as examples of this final
stage of speciation where reproductive isolation is perfected (Remington,
1985).

The origin of reproductive isolation accompanied by a period of genetic
reorganization or “revolution” both stem from the concept of a highly
integrated genome unique to each species. The gene pool of the biological
species is assumed to be composed of an array of “coadapted gene
complexes”, groups of genes acting together in a highly coordinated
manner and adapting the organism to its environment. This model is
based partly on studies of chromosome inversions in natural populations
of Drosophila (Dobzhansky, 1970, ch. 5,9; Lewontin, 1974, ch.3). Such
inverted gene sequences (identified by characteristic banding patterns in
polytene chromosomes) are thought to be of adaptive value because their
frequency often varies in predictable geographic and seasonal patterns
in specific populations. (Their true adaptive value is still to be deter-
mined.) The inversion acts to prevent crossing over and recombination in
meiosis and is therefore thought to be favored by selection as a mecha-
nism to maintain the gene complexes intact. This interpretation appears
to corroborate the mathematical models of Wright (1931) who described
the stable equilibrium of specific combinations of alleles for interacting
genes. The allele frequencies for given loci in a population reflect the
fitness that specific allele combinations confer; other populations in
different ecological settings are characterized by different allele frequen-

30(l-2):45-81, 1991

49

cies and combinations. Unfavorable combinations reduce fitness and
thus selection controls allele frequency by eliminating less fit genes and
gene arrangements. In a small population selection and fortuitous gene
frequency changes due to drift (or - at the extreme - local extinction and
recolonization) can shift the population to more favorable gene combina-
tions, or “adaptive peaks” in Wright’s graphic depiction.

Intraspecific test crosses between Drosophila populations with differ-
ing inversion types have produced progeny with reduced fitness. By
extension, interspecific lab hybrids are often subvital or barren presum-
ably because of the disruption of their respective coadapted genomes.
These observations provided the basis for the idea of the highly inte-
grated gene pool unique to each species. Speciation would then necessar-
ily require revolutionary changes to achieve a new set of harmonious
gene complexes appropriate for new adaptations. Selection in turn would
favor perfection of isolating mechanisms preventing disruptive interspe-
cific hybridization. If this view of the genetics of speciation is true, one of
the constraints on any “genetic revolution”, and on overall speciation
rates, must be the accompanying loss of fitness during such transitions.
This is especially true for traits directly relating to reproduction, where
independent gene systems in the two sexes control separate but compat-
ible aspects of mate location and fertilization. Any mutation affecting
mate recognition in one sex would probably be disruptive unless an
unlikely complementary change occurred fortuitously in the opposite
sex.

During the last two decades researchers have critically reexamined the
Biological Species Concept, aided by computer modeling and mathemati-
cal analysis.

2. Criticism of the Biological Species Concept.

a. Gene flow and population structure.

The origin of species in allopatry remains a widely accepted model, both
because biogeographic patterns of variation support it and because well
established theory shows that even small rates of gene flow between
populations can effectively prevent genetic divergence. Our understand-
ing of the factors which maintain species integrity among separate
populations has changed over the last three decades. In the Biological
Species Concept gene flow and a highly coadapted genome tend to unify
the gene pool. Ehrlich and Raven (1969) point to practical problems in
testing the potential for gene exchange and question the assumption that
gene flow rates among populations are high enough to prevent significant
genetic differentiation. Populations are often widely separated and
distance alone can act as a barrier to gene flow. They view species as
genetic mosaics of variable populations whose relative reproductive
isolation may be untestable. No one population can characterize a species
and reproductive isolation may be poorly developed in certain groups,
making taxonomic boundaries arbitrary. The population or deme is the

50

J. Res. Lepid.

ecological and evolutionary unit, not the taxonomic species. Entire
species do not evolve uniformly geographically nor synchronously (Ehrlich
& Murphy, 1981; Murphy & Ehrlich, 1984).

Recent concepts (Endler,1977; Slatkin, 1973, 1987; Templeton, 1980)
de-emphasize the unifying effects of gene flow and a highly integrated
genome. The genetic makeup of a given population is a function of not
only the stabilizing effects of gene flow but also the opposing effects of
mutation, genetic drift, and diversifying selection. In addition, the
present genetic structure of a population reflects major historical events
(e.g. glaciation, past periods of selection) affecting its present demography
and genetic variability. The organization of genes within the genome is
relevant to speciation processes (Templeton, 1982). Genes may occur as
many loci throughout the genome, each with a small effect, or at the other
extreme may exist as one major controlling gene with a few modifier loci.
In a founder population genetic drift more likely would be important for
a single large-effect gene than in the case of the polygene system.

Various classes of genes respond differently to selection. Thus, even in
the absence of significant gene flow, geographic variation may be mini-
mal for genes controlling critical developmental and metabolic path-
ways, mating and reproduction, and other characters under strong
stabilizing selection. Other traits, such as morphology and life history
traits, may exhibit abundant geographic variation. The adaptive value of
much morphological variation - including that used to characterize
subspecies -is often unclear, and while such characters may at first
appear taxonomically useful, they may not be well correlated with overall
genetic differentiation among populations within a species. The wide-
spread view that subspecies are necessarily incipient species (e.g.
Hammond 1985,1990) is in general unfounded. As will be discussed
below, reproductive isolation is likely to arise under special circum-
stances of population genetics.

In Lepidoptera examples of adaptive geographic variation include
regional host plant specialization (Bowers, 1986; Fox and Morrow, 1981;
Scriber, 1983) , voltinism (Rabb, 1966; Waldbauer, 1978), polyphenism
(Janzen, 1984; Shapiro, 1984ab), polymorphisms related to mimicry
(Gilbert, 1983), etc. Unfortunately, the genetic basis for these adapta-
tions is generally poorly known. Conversely, it has been difficult to
establish the adaptive value of the considerable geographic variation in
allozymes (Ehrlich & White, 1980; Johnson, 1976;Kingsolver& Wiernasz,
1991; McKechnie et al., 1975; Watt, 1968).

Some of the best studied examples of the adaptive value of geographic
variation in life history traits are in the frog genus Rana . Moore (1957)
revealed a north-south cline in genetic adaptations regulating larval
development in response to water temperature in the R. pipiens complex
(Leopard frogs). Experimental hybrid tadpoles showed distortions in
body size and shape which increased as a function of geographic distance
and difference in ambient water temperature experienced by the paren-

30(l-2):45-81, 1991

51

tal populations. This work eventually led to a systematic revision of the
group (Sage & Selander, 1979 and cited refs.). In a similar study (Berven
& Gill, 1983) found egg size, larval growth rates, and mature larva size
to be critical adaptive variables, yet their relative genetic correlations
and heritabilities varied within and among species as geographically
separate populations were compared. In other words, the underlying
gene systems were not just “fine tuned” in a clinal fashion, but were
fundamentally reorganized in response to diverse environmental selec-
tion.

The evolutionary significance of these examples is that adaptive
geographic variation among populations within a species is not qualita-
tively different from that distinguishing ecological and physiological
characters among related species. Shifts to new adaptive modes or
ecological niches can occur within species, even along continuous dines,
and often without accompanying changes in gross morphology or repro-
ductive compatibility.

b. Genetic revolutions and speciation in founding populations.

The early concept of a genetic revolution in a founding population
leading to speciation has been revised by its advocates (Carson, 1982;
Carson & Templeton, 1984; Mayr, 1982; Templeton, 1981,1982). These
authors promote the theory of rapid speciation in small isolated popula-
tions, but differ with respect to the number and kind of genes involved,
and the type of selection acting on them. The Carson model of reorgani-
zation of polygene balances involves a relaxation of selection (during a
phase of rapid population growth) allowing a major reorganization of the
genome. Templeton describes the action of strong selection on a few
major genes in favoring a shift to a novel Wrightian adaptive peak. Both
authors cite the example of the Hawaiian Drosophilids where isolation
of small founder populations by lava flows seems to be involved in the
extraordinary adaptive radiation of these flies into more than 700
endemic species (Carson & Kaneshiro, 1976).

Barton (1989), Barton & Charlesworth (1984), and Felsenstein (1981)
give critical, population genetic analyses of the restricted conditions
under which genetic changes leading to reproductive isolation are actu-
ally likely to occur. Barton & Charlesworth (op.cit.) conclude that the
probability of speciation during a single founder event is extremely low.
Reduction of variability due to drift, which could promote a reorganiza-
tion through the uniting of rare recessives, actually reduces the likeli-
hood of shifts to new adaptive peaks. Templeton (1981,1982) advocates
the role of a few major genes in the origin of reproductive isolation, yet
the substitution of a new allele in a single gene with major effects on
reproduction will be especially opposed by selection. Conversely, the
sequential change in many genes with smaller effects, as in the Carson
model, is unlikely to occur quickly in a founder event. Barton &
Charlesworth (op. cit.) support the model of a change in selection favoring

52

J. Res. Lepid.

gradual genetic change over many loci which eventually results in
reproductive isolation. While this process does not necessarily require a
small, isolated population, novel selection regimes are in general more
likely at the periphery of a species range, on islands, or in mountain
ranges isolated by desert, etc.

The many surveys by biochemical geneticists have found abundant
allozyme variations within and among populations, but have not found
evidence for a genetic revolution. Allozyme types and frequencies vary
progressively through a hierarchy of intraspecific populations, conge-
ners, and related genera (Avise, 1976). It may be generally true that most
of the genetic differences (based on allozyme studies) between closely
related species are present as polymorphisms within these species
(Lewontin, 1974 ch.4).

c. Genetic changes during speciation.

Electrophoresis reveals variation in “structural” genes coding for
metabolic enzymes, which are unlikely to relate directly to speciation.
Electrophoretic studies are very useful in constructing phylogenies and
in detecting gene exchange between populations (Geiger and Scholl,
1985; Geiger & Shapiro, 1986; Porter and Geiger, 1988), but show only
correlation, not causation with regard to speciation.

The “regulatory” aspects of the genome controlling reproduction and
development are more relevant to speciation but are also more difficult
to study than allozyme variation, especially in organisms for which we
lack gene sequence or linkage data. Experimental hybridization can
reveal reproductive and developmental incompatibilities, but cannot
reveal directly the underlying genetic basis (Collins, 1984; Hafernik,
1982). By crossing members of geographically isolated populations of
Phyciodes , Oliver (1972, 1978, 1979ab, 1980) demonstrated genetic in-
compatibilities (disrupted diapause and emergence schedules, etc.) within
species which increased with distance of separation but differed only in
degree from developmental problems seen in interspecific crosses. In
assessing genetic differences between closely related species it is impor-
tant to remember that some of the incompatibility seen in hybrids may
be due to post-speciation genetic divergence rather than resulting from
the speciation process. Geiger (1988) and Lorkovic (1986) discuss and
debate the proper application of experimental hybridization and enzyme
assays to taxonomic problems in Lepidoptera. Indices of enzyme similar-
ity and hybrid compatibility are in general agreement in comparisons
ranging from intergeneric to subspecific. Discrepancies may occur in
specific cases near the species level. Parker et al. (1985) employed both
hybridization and electrophoresis to demonstrate developmental incom-
patibilities among related Centrarchid fish.

The products of regulatory genes are likely to be relatively small
molecules present in low concentrations, which regulate the action of
other genes distributed throughout the genome (Britten & Davidson,

30(l-2):45-81, 1991

53

1969; Hedrick & McDonald, 1980). If such a system more accurately
describes “coadapted gene complexes” than the tightly linked inversion
sequences of the Biological Species Concept, this finding would bolster
the criticisms of speciation through genetic drift in founder populations
because changes at one or a few loci would have less effect in an
interacting, distributed system of genes each with small effects. When
many loci are involved new, favorable combination are less likely to occur
(see (b) above).

Studies with Drosophila are the most detailed examples of the genetic
basis of reproductive isolation (Dobzhanski, 1970, p. 340 ff; Ehrman,
1962; Schafer, 1978; Sved, 1979). Typically, effective prezygotic isolation
is based on the additive effect of many genes controlling a variety of
physiological, morphological, and behavioral traits, none of which is
effective by itself. In an important synthesis of published data, Coyne and
Orr (1989a) analyzed pre- and postzygotic isolation in Drosophila hy-
brids with respect to genetic similarity (based on allozyme surveys) and
sympatry vs. allopatry of given species pairs. They found a strong
correlation between strength of prezygotic isolation and sympatry, which
suggests the evolution of mating barriers due to selection against
hybridization. Postzygotic isolation seemed to evolve in allopatry, in-
creasing in severity with genetic distance, which is in turn an index of
evolutionary time. The Drosophila seem to be an exception to the general
lack of evidence in other animals for improved premating isolation in
sympatry (see discussion in Section 3 below).

Coyne and Orr ( 1989b) relate their findings to the work of Charlesworth
et al. (1987) on the occurrence of hybrid incompatibility in the heteroga-
metic sex (Haldane’s Rule); in Lepidoptera females are XY and sterility
and in viability is almost always confined to hybrid females. In this
model, beneficial recessive mutations arising during divergence in allo-
patry tend to accumulate faster on the sex chromosomes than on
autosomes. But upon hybridization following secondary contact the
expression of these genes is disrupted in the hybrid genome, and
especially in the heterogametic sex where no dominant gene would be
present to mask the recessive allele on the X chromosome. The
Charlesworth model thus provides an explanation both for the origin of
postzygotic isolation and for Haldane’s Rule in hybrids between closely
related species (Coyne & Orr, 1989b). There is recent evidence that key
regulatory genes controlling the expression of a disparate collection of
reproductive traits may reside on the sex chromosomes (Carde & Baker,
1984; Charlesworth, et al., op. cit.; Grula & Taylor, 1980ab; Hagen, 1990;
Taylor, 1972; Tuskes & Collins, 1981). Further research, perhaps em-
ploying new techniques in gene sequencing and mapping, will reveal if
this phenomenon is a general mechanism in speciation.

Yet, in general we still lack a detailed knowledge of the genetic basis
for traits associated with reproductive isolation (Lewontin, 1974) . If an
overall genetic revolution does not accompany speciation, how much

54

J. Res . Lepid.

genetic divergence is necessary to achieve reproductive isolation? In
closely related Lepidoptera, for example, how many genes control respec-
tive differences in wing pattern and associated courting behavior, or
interspecific variation in pheromones and their time of release? How
much geographic variation exists in genes controlling reproductive traits
within nominal species? What are the factors involved in the origin and
maintenance of this variation? Is this variation of the same order of
magnitude as that found in allozymes, morphology, or ecological adapta-
tions? Some geographic variation is known to occur in the component
chemicals of pheromones, although in only a relatively few cases can
differences be interpreted as antihybridization mechanisms (Garde,
1987; Garde & Baker, 1984). Rutowski (1984) reviews the role of sexual
selection in the evolution of butterfly mating behavior.

3. The Recognition Species Concept.

Paterson (1986) formulated this concept in an attempt to correct
perceived shortcomings in the Biological Species Concept. He defines
species as “the most inclusive population of individual biparental organ-
isms which share a common fertilization system”. The Recognition
Concept looks at “reproductive isolating mechanisms” from a different
point of view. Traditionally, labeling traits involved in reproduction as
isolating mechanisms implied that they evolved through natural selec-
tion for this adaptive “purpose”. Paterson stresses that their origin is
based on their adaptive value in identifying mates, regulating the
initiation and full expression of courtship in both sexes, and achieving
successful mating and insemination. In this context, any subsequent role
such traits might play in isolating a gene pool from hybridization is
irrelevant, especially if they arose in allopatry.

This theory of the origin of mate recognition traits has generally been
acknowledged as useful and has been supported by the general failure to
document the improvement of reproductive isolation in areas of sympa-
try between species, or in stable hybrid zones (Butlin, 1989; Spencer et
al., 1986). See Endler (1989) and Endler & McClellan (1988) for a more
general discussion of the role of mate recognition/isolation traits in
evolution.

Other aspects of the concept have been extensively criticized (Coyne et
ah, 1988; Templeton, 1989). The major points of criticism can be summa-
rized as follows: 1) The true adaptive origin of reproductive traits does not
invalidate the role isolation plays in allowing species to evolve indepen-
dently. 2) Paterson ignores the role postzygotic isolation plays in regulat-
ing gene exchange between taxonomic entities. Hybrid incompatibility is
not an intellectual abstraction, as Paterson suggests, even if this unfit-
ness arose from genetic divergence unrelated to selection favoring
species recognition. Postzygotic incompatibility acts to stabilize genetic
processes in hybrid zones by regulating gene exchange. 3) The Recogni-
tion Concept is not superior in correctly assigning species status since it

30(l-2):45-81, 1991

55

is burdened with many cases of geographic variation in prezygotic
isolation (or recognition) traits within species. This criticism responds to
the symmetrical argument by Paterson that many apparently distinct
species lack effective prezygotic isolation and thus cannot be distin-
guished by isolation criteria.

4. Geographic dines and parapatric speciation.

Clines are formed when phenotypic characters vary over a geographic
gradient. This kind of variation is well known for Lepidoptera and other
organisms; many subspecies are connected to other populations through
character dines. Most dines are thought to be the phenotypic response
to environmental selection gradients varying geographically in inten-
sity. Slatkin (1973) and Endler (1977) have analyzed dines in terms of
formal, mathematical models. Clinal variation will be minimal when
gene flow rates are high and selection is weak. Abrupt dines occur when
gene flow is weak and selection gradients are steep. Endler (op. cit.)
believes that gene flow rates have traditionally been overestimated and
that most populations are only weakly connected by gene flow. Thus,
strong selection can geographically differentiate a species. If critical
genes adapting a population to its environment are closely linked to those
controlling mate choice, natural selection could favor the origin of
reproductive isolation if gene flow rates were below some critical value.
This model is often referred to as parapatric speciation.

Endler (op. cit.) also demonstrates that in the absence of good biogeo-
graphic evidence this kind of primary differentiation would be indistin-
guishable from secondary contacts between previously isolated popula-
tions. The term “dine” is often used too casually in the Lepidoptera
literature to refer to anecdotal examples of phenotypic variation. Many
other phenomena cause taxonomic confusion by mimicking dines: mosa-
ics of small populations where variation is due to founder effects rather
than selection gradients; stable polymorphisms; narrow bands of sympa-
try between noninterbreeding but phenotypically very similar species;
and hybrid zones where two taxa regarded as separate species are
hybridizing. Critical analysis is required to distinguish these situations.

5. The Subspecies concept and its relevance to speciation mod-
els.

Subspecies are recognizably different geographic populations or sets of
populations assigned formal taxonomic rank. The use of the subspecies
category has proliferated in the taxonomy of Lepidoptera, partly because
it is a catch-all for difficult problems in classification, but also because of
the natural tendency of taxonomists to split intensely studied groups into
new named entities. Many evolutionary biologists feel the naming of
morphological subspecies is largely an arbitrary decision, which on the
one hand may help catalog variation within a species, but on the other
hand may actually mislead further study by implying a specific genetic
status and by ignoring other significant patterns of divergence.

56

J. Res. Lepid.

Four specific objections can be raised against the subspecies concept.
First, there is no testable criterion to assign the subspecies rank, such as
reproductive isolation as a test for species status. How different must a
population be to be called a subspecies? Second, a subspecies may be a
member of a clinal or mosaic array of variable populations. Geographic
variation in several characters may not be congruent, making geo-
graphic limits for a subspecies quite arbitrary. Third, other populations
undistinguished by phenotype may actually have diverged in more
significant ways, such as adaptations to new hostplants, shifts in mating
time, etc. Fourth, traditional morphological subspecies are not necessar-
ily undergoing incipient speciation, even when isolated in unique envi-
ronments. As has been discussed above, the origin of reproductive
isolation is conservative and may require special circumstances of
population genetics.

6. Hybrid Zones as natural experiments in speciation.

From the standpoint of the collector, hybrid zones are usually first
detected as areas where two taxonomically distinguishable populations
interbreed to produce a population of hybrids. With further analysis,
hybrid zones are seen as narrow character dines maintained by hybrid
unfitness. This definition is a practical way of distinguishing hybrid
zones from intraspecific dines and dines or blend zones between recog-
nized subspecies. Subspecies, by definition, should interbreed without
loss of fitness, but the subspecies concept is arbitrary since the genetic
and geographic boundaries of subspecies are themselves arbitrary. The
distinction between hybrid zones and blend zones is predicated on
knowledge of genetic compatibility; morphological analysis cannot reli-
ably distinguish these phenomena nor determine if a zone is the result
of primary or secondary intergradation (Endler, 1977; Section 4, above).

The structure of hybrid zones suggests a dynamic equilibrium between
gene flow, which would tend to widen the zone, and hybrid unfitness
which would tend to narrow the zone. Prezygotic isolation is either
lacking or ineffective such that the two taxa cannot exist in sympatry
without interbreeding. The two parental phenotypes are rarely found
together in most zones unless hybrids are effectively sterile in both sexes.
Studies of hybrid zones provide evidence against the concept of the origin
or perfection of premating isolating mechanisms in response to hybrid-
ization (Butlin, 1989; Paterson, 1986; Spencer et al., 1986). The historical
context of hybrid zones is difficult to determine, but many appear to have
resulted from secondary contact following Pleistocene range changes. If
so, this would point to a long term stability in the equilibria maintaining
the zone. Unlike many plant hybrid zones (Stebbins, 1974), most hybrid
zones in animals are not associated with ecotones, which suggests they
are maintained by hybrid unfitness, rather than differential ecological
adaptations.

30(l-2):45-81, 1991

57

The structure and complexity of hybrid zones, and their obvious
relevance to species and speciation concepts, are intellectually very
seductive and a large body of work has been done by both empirical and
theoretical evolutionary biologists. This literature has been reviewed by
Barton & Hewitt (1981,1985), Bigelow (1965), Harrison & Rand (1989),
Hewitt (1989), and Woodruff (1973,1981). The most comprehensive
summary of mathematical models and computer simulations of hybrid
zones are to be found in Barton & Hewitt (1985) and the works of Barton
cited therein.

Postzygotic developmental incompatibilities appear to be the most
important factor limiting gene flow across hybrid zones. If at least one sex
is fertile in hybrid zone populations, a great deal of genetic variation will
result from backcrossing, the uniting of individuals of mixed genetic
background, and the effects of recombination. Indeed, hybrid zone
phenotypes may be poorly correlated with other indices (allozymes,
fertility, etc.) of hybridity. Many hybrid zones contain “hybrids” with
complex genotypes (or this is inferred from overall phenotypic variation)
and yet the zones are narrow with respect to estimates of dispersal rates.
One might expect natural selection to act on this range of genotypes to
increase compatibility and thus lead to a fusion of the interbreeding taxa.
The narrow width of some hybrid zones suggests that incompatibilities
result from disharmonious gene interaction at many loci such that
recombination tends to break up harmonious gene combinations as they
arise. Gene flow from outside the zone would have the same effect.

In the absence of a detailed knowledge of the genetics of incompatibili-
ties, most workers have used allozymes and morphological characters as
markers to measure zone width and to construct character dines away
from the zone. If only a few loci with large effect control hybrid fitness,
neutral or advantageous alleles from one interbreeding taxon should
introgress across the zone and form long dines into the other population.
This will not occur if such alleles are closely linked to any deleterious
gene. Thus, if many fitness loci are involved, hybrid zones can form strong
barriers to gene exchange, even in the complete absence of premating
barriers. This would be especially true for species with low dispersal
rates. Barton & Hewitt (1989) and Hewitt (1989,1990) argue that a
species could be dramatically subdivided into genetically distinct popu-
lations through the action of environmental disturbance (e.g. glaciation)
and the consequent formation of parapatric hybrid zones.

Bigelow ( 1965) points out that when selection favoring improvement of
reproductive isolation is strongest, i.e. when hybrid unfitness is highest,
then gene flow actually will be lowest into bordering parental popula-
tions. The great majority of matings either side of the hybrid zone will be
between “pure” genotypes. The hybrid zone itself, through incompatibili-
ties, isolates the two interbreeding forms.

In summary, hybrid zones can reveal much about the genetics of
speciation. They provide a laboratory for testing theories about the

58

J. Res. Lepid.

nature and origin of premating barriers. Genetic differences between the
interbreeding taxa causing disruption of postzygotic development prob-
ably represent the kinds of critical genetic changes leading to speciation.
Unfortunately, we cannot know the time span between such differentia-
tion and the formation of a given hybrid zone. We must also bear in mind
that speciation can occur by the formation of strong premating isolation
with little or no postzygotic incompatibility. Closely related species in
this class will not likely form hybrid zones. The great deal of genetic
variation generated in hybrid zones could potentially lead to novel
adaptations. The ability of hybrid zones to block introgression, however,
may limit the role this variation can play in evolution.

7. Sympatric speciation.

Models of sympatric speciation describe conditions under which repro-
ductive isolation could theoretically arise on a distance scale comparable
to the average dispersal of breeding individuals in a population (Bush,
1969; Diehl and Bush, 1989; Tauber and Tauber, 1989). The model
requires a close association between mating and host choice, both
behaviorally and genetically. Mating would occur on the host and genes
controlling mate choice and host preference would consist of single loci
closely linked, thus reducing the effects of recombination. If a new allele
arose by mutation which adapted the organism to a new host (by altering
its digestive enzymes, for example) selection would tend to favor compat-
ible variation in the gene controlling host association and thus mate
choice. Homozygous matings would be favored if heterozygotes were ill-
adapted. A new host race or species could arise in this manner. The model
has been criticized on the grounds that the genetic system is unrealisti-
cally simplified, and that when the new allele first appears matings with
parental genotypes would occur and tend to break up gene combinations
favoring the host shift (Futuyma and Mayer, 1980; Butlin, 1987) . While
some organisms, such as the Rhagoletis (Diptera:Tephritidae) fruit flies
studied by Bush, seem to fit the model, most Lepidoptera in natural
environments do not mate in such close association with their hosts and
many exhibit some degree of dispersal in response to pheromones or
during courtship behavior. Rapid shifts in host choice could accompany
speciation in Lepidoptera in the context of range expansion or coloniza-
tion. It may not be possible to know if change in host plants is a cause or
an effect of range expansion. In evaluating the many putative examples
of sympatric speciation it is difficult to define criteria which would
exclude allopatric models.

Summary

No single process controls the genetic changes leading to reproductive
isolation between closely related animal populations. Reproductive iso-
lation may arise incidental to the evolution of the ecological, morphologi-
cal and physiological traits we use to characterize species. Alternatively,

30(l-2):45-81, 1991

59

reproductive isolation may be poorly developed among taxa otherwise
rich in taxonomic character differences. Strict application of the isolation
criterion of the Biological Species Concept is inappropriate in such cases.

To study speciation we must compare genetic differences between
closely related taxa and infer which of these accompanied the origin of
reproductive isolation. Experimental hybridization is an important tool
in this regard. One of the challenges for the evolutionary biologist is to
distinguish the genetic differences among species from the genetics of
speciation.

Populations of species may become differentiated through the action of
selection, genetic drift and historical events affecting demography; the
roles of gene flow and coadapted gene complexes in countering this
divergence and maintaining species integrity may be less important than
once believed. Under conditions of reduced gene flow, strong selection
gradients may produce character dines or otherwise act to differentiate
populations. Theoretically, it is impossible to distinguish primary from
secondary differentiation. Genetic compatibility may decrease within a
species when test crosses are made with individuals from increasingly
distant populations. Although taxonomy must deal with discrete species,
single populations are the ecological and evolutionary units of change.
Speciation does not necessarily require a major restructuring of the
genome; many of the fixed differences in allozymes between species are
present as polymorphisms within closely related species.

Features which traditionally have been termed “reproductive isolating
mechanisms” may actually arise as products of selection favoring in-
creased reproductive fitness for “mate recognition” within a population.
Subsequent isolation from related taxa is then a byproduct of this
process. Contrary to the predictions of the Biological Species Concept,
theoretical arguments and empirical data do not support the routine
improvement of isolation between species which have come into second-
ary sympatry or which have formed hybrid zones. The Drosophila may be
an exception to this rule.

Populations lacking effective prezygotic isolation may become signifi-
cantly different for traits affecting postzygotic development in their
hybrids. Upon secondary contact such taxa form hybrid zones, the
structure of which is largely determined by the opposing effects of hybrid
unfitness and gene flow. The study of hybrid zones has been profitable
in understanding the genetics of speciation.

PART II. Representative Studies.

I have chosen especially well-documented studies in five taxonomic
groups to illustrate the concepts of population differentiation and specia-
tion presented in Part I. These genera or species groups share at least
some aspect of natural hybridization, and by this criterion deal with taxa
near the species boundary of evolution. All the studies involve extensive

60

J. Res. Lepid.

field and laboratory research using various methodologies, as contrasted
to taxonomy based solely on comparative morphology. An ideal program
for studying speciation might use experimental hybridization to under-
stand pre- and postzygotic genetic compatibility, and morphometries and
biochemical genetics as independent measures of genetic similarity and
gene exchange between populations. No single example here employs all
these approaches. In particular no direct observation of mating behavior
or prezygotic isolation was made for any group except the Hyalophora ,
where the pheromone mating system is more easily manipulated than
the complex courtship of butterflies.

For the purpose of stimulating thought and discussion, I offer in some
cases alternative interpretation of likely modes of population differentia-
tion and speciation. Thus, for the tiger swallowtails, I give evidence for
a secondary contact between full species in place of an ecological model
of primary intergradation between glaucus and canadensis. Similarly,
for the Limenitis problem I propose for consideration a secondary
intergradation/hybridization model for the blend zone between arthemis
and astyanax , although in this example definitive allozyme and experi-
mental hybridization data are not yet available. The dines in mimetic
morphs in Heliconius illustrate how selection can maintain a dramatic
shift in phenotype over a short geographic distance. Yet the genetic
differences between these subspecies appear to be relatively minimal. In
the Hyalophora premating barriers are lacking and the degree of
postzygotic isolation in hybrids among the various taxa reveals a hierar-
chy of levels of speciation.

The Tiger Swallowtail Species Group.

These large, attractive butterflies are well known to collectors yet still
pose new and interesting questions concerning species relationships.
The eastern Papilio glaucus has a female polymorphism: a yellow form
differing slightly from the male, and a dark morph believed to be a mimic
of the distasteful Battus philenor. The three western species are
eurymedon , multicaudatus , and rutulus (which most closely resembles
glaucus ). No dark female morph occurs in the western species.

Clarke & Sheppard ( 1955, 1957, 1962) hybridized glaucus with rutulus
and eurymedon to study the genetics of the dark morph. Their results
indicated significant genetic differences between glaucus and the west-
ern forms. In glaucus the locus controlling the dark morph appeared to
reside on the Y chromosome, but the expression of this gene was blocked
in hybrids with the western species. Prolonged or “permanent” diapause
in female pupae may occur commonly in these hybrid crosses, although
this may be overcome by injection of eedysone (Clarke & Willig, 1977). A
similar diapause disruption occurs with glaucus x multicaudatus crosses
(West & Clarke, 1988). These crosses were made using the technique of
hand pairing; nothing is known concerning premating barriers in nature.

Brower (1959ab) investigated species relationships within this group

30(l-2):45-81, 1991

61

by means of morphology, and field observations. He attempted to estab-
lish the existence of hybridization between glaucus and rutulus using a
single wing character and slight differences in male genitalia. Areas of
suspected hybridization occurred in British Columbia and the Black
Hills of South Dakota. On the basis of genitalia alone certain specimens
intermediate for wing spot color in each locality would have been
classified as glaucus. Brower (1959a) noted that the most convincing
data for natural hybridization are for the Black Hills, where rutulus is
rare compared to glaucus.

To date these populations have not been reexamined using modern
methods of multivariate analysis or surveys of allozyme variation. No
series of putative hybrids have been illustrated in any publication. Yet,
recent literature treats the existence of hybrid zones as well established.
Scott (1986) reduces rutulus to a subspecies of glaucus, presumably on
the basis of putative hybrid zones, but without a discussion of justifying
criteria.

Scriber (1983,1984) has extensively documented geographic variation
in host plant adaptation in the glaucus group and the genetics of sexual
dimorphism in glaucus (Scriber et al., 1987). The experimental basis for
this work has been hybridization with a northern taxon, canadensis,
considered a subspecies of glaucus, which lacks the dark female morph,
is univoltine throughout its range, is smaller than glaucus and differs in
several wing pattern characters. Hybridization and controlled rearing
experiments have shown the following: 1) the gene for dark morph in
glaucus is carried on the Y (W) chromosome and is suppressed by an X-
(Z-) linked locus in canadensis ; 2) obligate diapause in canadensis is
controlled by a sex-linked locus, whil e glaucus has a facultative diapause
responsive to daylength and controlled by autosomal loci (Hagen &
Scriber, 1989; Rockey, et al., 1987ab); 3) hybrid female pupae typically
enter a prolonged or indefinite diapause, as in the interspecific crosses
described above. While both forms feed on Wild Black Cherry, they are
each unable to metabolize a common host of the other taxon - Tulip Tree
for glaucus and Quaking Aspen for canadensis. Hybrid Fx survive on all
three hosts.

The two taxa are parapatric in the Great Lakes region and in the
Northeast. Obviously, a clear understanding of the origin of ecological
and physiological differences between glaucus and canadensis, and
moreover the nature of species and speciation in this group, rests on a
correct interpretation of contact zones between the various taxa.

Scriber et al.(1987), Scriber & Evans (1988), and Luebke et al.(1988)
describe the interaction in Wisconsin as a “hybrid zone” between subspe-
cies. Scriber (1983) proposes that ecological factors determine the tran-
sition zone between subspecies. The northern limit of glaucus corre-
sponds to the demarcation of 1200 - 1300 degree-days, which lab studies
show to be the northern limit for completion of a second brood. This region
fairly closely parallels the southern extent of Quaking Aspen and thus

62

J. Res. Lepid.

the southern range limit of the canadensis phenotype. Importantly, all
diagnostic ecological, morphological, and physiological characters change
concordantly and abruptly in a north-south transect. There is no evi-
dence for dines in polymorphism in host plant adaptation, dark morph
frequency, or voltinism as one would expect for overlapping, interbreed-
ing subspecies.

The only evidence for interbreeding in the Great Lakes region comes
from morphometric analysis using lab hybrids as a third reference group.
This analysis was made difficult by variability in key characters in the
lab hybrids (Collins & Luebke, unpubl.; Luebke, 1985; Luebke et al., 1988).
Only two wing pattern variables plus wing length are statistically
significant, resulting in relatively large misclassification errors: 19.3%
for reference lab hybrids and 8.3% for glaucus reference specimens
(incorrectly classified as hybrids). The results for the transition zone in
Dane Co. were 2.7 % canadensis , 12.2% hybrids, and 85.1 % glaucus for
a sample of 74. It is unclear if these phenotype ratios can be found in a
single local population, but if so the presence of parental phenotypes
indicates either assortative mating or selection against hybrids.

It is possible that both the hybrids and canadensis scores are at least
partly due to misclassification of glaucus , as noted above. Another
possible source of error is confusion arising from the resemblance
between the “spring form” of glaucus with canadensis. Ideally, the
glaucus reference group should be composed only of these spring adults.

The Wisconsin transition zone thus appears to be a case of parapatry
between morphologically, ecologically, and physiologically distinct spe-
cies. They are partly isolated by postzygotic incompatibility in diapause
physiology, and the morphometric evidence for interbreeding is inconclu-
sive. The concordant and abrupt discontinuity for key distinguishing
traits is highly suggestive of either very limited hybridization with
selection against hybrids, or for a narrow zone of sympatry without
interbreeding.

Hagen (1990) reports on the contact zone between glaucus and
canadensis in the Northeast, using allozyme variation, host plant suit-
ability, voltinism, and a diagnostic wing character. The two taxa differ in
only two enzymes of 13 polymorphic loci tested. Populations were
polymorphic at these two loci in three sites in northern Pennsylvania and
south-central New York, whereas populations north and south of this
region were fixed for alternate alleles. The loci are sex-linked. This region
also corresponds with a within-brood intermediate ability to survive on
some hosts of both subspecies, and with intermediacy in a wing band
trait. Like canadensis , “hybrid zone” populations lack the dark female
morph and are univoltine. Hagen interprets the contact zone as a very
narrow hybrid zone maintained by hybrid unfitness, perhaps due to
disrupted diapause in hybrids and/or metabolic “costs” associated with
maintaining detoxification systems for the hosts of both taxa. Since the
allozyme loci are sex-linked, their sharp frequency dines in the hybrid

30(l-2):45-81, 1991

63

zone may be due to close linkage with the locus controlling diapause,
rather than to selection directly on the enzyme loci. Hagen ascribes
variability in host plant utilization and allozyme frequencies to gene
introgression. However, the “spring form” of glaucus in New York is
known to feed on Tulip Tree yet may resemble canadensis in wing
markings, lacks the dark morph, and is univoltine at its northern limit
(Hagen, 1990; Luebke et al.; 1988; Shapiro, 1974). In two of the three
“hybrid zone” locations the allozyme frequencies for one locus closely
approximate those of glaucus to the south. The best evidence for inter-
breeding is the very sharp cline in allozyme frequencies for two loci.

The data at present do not allow an estimate of the relative frequency
of primary hybrids or the extent of backcrossing and recombination, if
any. It is possible that a past period of interbreeding produced the
pattern of allozyme variation through introgression into a predominately
glaucus genome. Climatic variation during the Pleistocene could have
resulted in cycles of allopatry between what we now call canadensis and
glaucus followed by range expansion and parapatry. If interbreeding is
ongoing, the narrow width of the apparent hybrid zone is more likely due
to strong selection on some locus, probably sex-linked, than to the
differentiating effects of broader ecological selection gradients.

Very little is known about prezygotic isolating mechanisms in nature.
The small female canadensis is reportedly difficult to hand mate to
larger male glaucus , and the flight characteristics of the two taxa are
apparently different in terms of plant community association (Scriber et
al., 1987; Scriber, pers. comm.) although R. Lederhouse minimizes these
differences (pers. comm.). In Wisconsin recent agricultural disruption
may have created a patchy plant community association which could
promote either limited hybridization or an increase in sympatry between
the two taxa in terms of islands of Aspen interdigitating with open fields
and fence rows with Wild Black Cherry supporting gZazzcws populations.
More extensive morphometric analysis using the spring form of glaucus
needs to be done in both localities and on a finer demographic scale.
Further analysis of these zones using mitochondrial genetics (Hagen,
1990) and multivariate phenetics will help determine the true extent of
gene exchange.

Scriber has attributed the occasional occurrence of gynandromorphs
and color mosaics in female glaucus to a disruption of development
resulting from long distance introgression from the putative northern
“hybrid zone” into more southern glaucus populations (Scriber et al.,
1987). This interpretation has been cited and accepted by West & Clarke
(1988) in their investigation of the inheritance of the black morph.
However, long distance introgression of a deleterious allele is not credible
in terms of our current knowledge of hybrid zones. Gynandromorphs and
color mosaics have been collected far from the range of canadensis and
occur in both pure and interspecific lab broods.

64

J. Res. Lepid.

The true taxonomic status of canadensis must also be considered in
future investigation of the western contacts between rutulus and the
eastern glaucus forms. It is interesting to note that canadensis resembles
rutulus in terms of sharing Quaking Aspen as a host, and in the form of
the yellow submarginal fore wing band and other adult wing characters.
Recently, Hagen and Scriber (1991) published an allozyme survey of
taxonomic relationships among the tiger swallowtail and P. troilus
species groups. Genetic relationships inferred from these enzyme data
confirm the full species status of canadensis , and show a very close
relationship between rutulus and eurymedon. Apparently, in the latter
two species rates of divergence have been greater for morphology and
ecological characters than for metabolic enzymes; there is no evidence for
intergradation in nature.

The Western Papilio machaon group of Black Swallowtails.

This relatively small group of butterflies illustrate many of the taxo-
nomic problems one encounters in applying species concepts to morpho-
logically distinct populations which lack effective reproductive barriers
(Sperling, 1990). The various taxa exhibit seasonal “forms” (polyphenism),
“dark” and “yellow” polymorphism, and discordant clinal variation. In
some areas where several taxa are sympatric the various taxa have
overlapping morphological variation and no one wing nor genitalic
character can reliably separate the various entities.

Sperling (1987) employed a multidisciplinary approach involving lab
broods reared from specific hosts, allozyme analysis, and morphometric
analysis of 11 key wing and body characters. Multivariate analysis of
allozymes, morphological characters, and the combined data set pro-
duced similar groupings of individuals and populations, but in some
cases adding host plant association as a variable improved species
separation. Host plant adaptations appear to play a central role in the
evolution of the machaon group.

Sperling synonymized with machaon the taxa oregonius and hairdii
and all other forms feeding on Artemisia dracunculus. Papilio zelicaon
and polyxenes remain distinct species. [Papilio indra has a distinct adult
morphology, genitalia, and larval phenotype and is not included in this
study. This species appears to be reproductively isolated from others in
the machaon group.] Hybrid zones occur between zelicaon and machaon
(primarily in Alberta), and between machaon and polyxenes (primarily in
Manitoba). Interestingly, these pairs of taxa remain distinct in sympatry
in many areas but hybridize to varying extent in others. Hybridization
suggested by morphometric analysis was confirmed by comparing allozyme
frequencies and calculating departure from Hardy-Weinburg expecta-
tions based on free gene exchange. No corroborative laboratory hybrid-
ization was done to measure pre- and postzygotic isolation among
populations or taxa.

30(l-2):45-81, 1991

65

Degree of hybridization seems to depend on the topography and host
plant association in the area in question, probably in terms of the effect
on mating behavior, especially “hilltopping”. In the Riding Mts. of
Manitoba habitat disturbance by man may have promoted recent hybrid-
ization, but other hybrid zones probably arose during the Pleistocene. In
nature it appears that machaon, zelicaon, and polyxenes are isolated by
ecological characteristics more than by mating barriers or postzygotic
incompatibilities. Experimental hybridization (Clarke et al., 1977;
Remington, 1968a) indicates minimal postzygotic genetic incompatibil-
ity in terms of fertility, embryo viability, disruption of diapause, or adult
sex ratios. These findings are in contrast to the Tiger Swallowtail group
discussed above. Nevertheless, more extensive hybridization should be
done, especially comparing populations in hybrid zones with areas of
successful sympatry.

Sperling (1987, 1990) believes that speciation and differentiation
within species in the machaon group occurred in allopatry in various
refugia during the Pleistocene. The Umbellifer feeding ancestors of
machaon apparently shifted to Artemisia dracunculus and became
subdivided into various subspecies. The present distribution of dines,
hybrid zones, and regions of successful sympatry are probably the result
of range expansion within the last 10,000 or so years.

The studies of Thompson (1988ab) on the genetics of oviposition
preference and specialization complement the work of Sperling. Female
zelicaon and machaon oregonius , bred from wild-collected iso-female
strains, were allowed to oviposit in cages on Cymopterus terehinthinus ,
Lomatium grayi (both native Umbellifer hosts of zelicaon ), Foeniculum
(an introduced Umbellifer host of zelicaon ), and Artemisia dracunculus
(the normal machaon host). Both species laid preferentially on their
native hosts and the ranking order was consistent overall within each
species. Yet, within each species, the iso-female strains did differ in
degree to which females laid some ova on hosts of the other species.
Within strains of m. oregonius some females laid ova on all four hosts,
while others laid only on the natural Artemisia host. Within strains of
zelicaon females differed in the ranking of their normal hosts, but
females within all strains laid at least some ova on A. dracunculus and
significantly more on Foeniculum.

Thompson interprets the variation in oviposition preference among
stains and within strains as evidence of genetic variation controlling this
behavior. Thus, these butterflies seem to have speciated by means of the
genetic-based flexibility to undergo a host shift, based on oviposition
behavior. Under novel selection favoring a host shift, as during range
expansion into new plant communities, the genetic potential exists in
latent form rather than requiring mutation. Future research may reveal
more detail about the underlying genetic basis and variation among
populations with respect to host plant adaptations.

66

J. Res. Lepid.

Intergradation between Limenitis arthemis and L. astyanax :
hybrid zone or selection along a cline?

The zone of phenotypic intergradation between the white-banded
Limenitis arthemis and the unhanded form L. astyanax was the subject
of a now-classic study by Platt and Brower (1968). They interpret the
zone of intergradation as an area of relaxed selection between the
northern banded form, which benefits from a wing pattern disrupting the
outline of the adult, and the dark blue unhanded form, which is seen as
a Batesian mimic of the unpalatable Battus philenor. In their model no
other restrictions on gene exchange exist, the two forms are considered
conspecific, and the distribution of phenotypes represents an example of
adaptive response to a gradient of selection, in other words primary
intergradation in the sense of Endler (1977). The opposing view is that
the two forms are actually well- differentiated and are hybridizing as a
result of secondary contact (Remington, 1968b, 1985). The present data
do not allow us to resolve this controversy, but the problem is worth
reviewing in some detail because it embodies so many key questions in
speciation theory.

The southern limit of the disruptive pattern phenotype is approxi-
mately concordant with the southern limit of Ice Age glaciation and so
the present distribution must represent recolonization. Old World and
western North American Limenitis are banded and this is assumed to be
the ancestral phenotype pattern. The zone of intergradation is about 160
km wide. Correspondingly, the northern extent of the unhanded pheno-
type is roughly the same as the northern limit of the mimicry model,
Battus philenor.

A selective basis for the mimetic astyanax phenotype has been estab-
lished experimentally by demonstrating that B. philenor is unpalatable
and that birds avoid the astyanax phenotype after exposure to philenor
(Brower, J,V,Z., 1958; Brower & Brower, 1962; Platt et al., 1971). The
supposedly disruptive value of the arthemis phenotype has not been
tested, which would probably require difficult field experiments, since a
flight behavior component may complement the wing patterns. No
experiments have been performed on the probability of predation on
phenotypically intermediate phenotypes.

Platt and Brower (1968) and Platt (1983) reason that since the two
forms are conspecific then the zone of intergradation must be of primary
origin, and not a hybrid zone of secondary contact. They cite three
categories of evidence for conspecific status: 1) wild-caught and reared
material show no deviation from a 1:1 sex ratio nor departure from
expected phenotype frequencies which might occur due to hybrid
postzygotic incompatibility (but see below); 2) the two phenotypes share
identical genitalic structures; 3) hybrid broods showed no gross evidence
for developmental incompatibility.

Several points of criticism can be made, especially of the 1968 study, in
light of recent advances in concepts and methods of investigation. Most

30(l-2):45-81, 1991

67

of the data on sex and phenotype ratios in Platt & Brower (op. cit.) are
based on crosses made with the same male individual and most females
are from one mating. No replicate crosses have since been published.
Quantitative studies of fertility, embryo viability, diapause disruption,
emergence schedules, and hybrid female fecundity should be done,
comparing different phenotypic classes and geographically separated
populations, as in the cited work by Oliver. Tests of the genetic model of
wing pattern inheritance proposed by Platt (1983) might provide an
estimate of the number of loci controlling the expression of the mimetic
phenotype. Platt ( 1975) and Platt et al. ( 1978) have investigated the wing
pattern genetics of the monarch mimic, L. archippus , as well as the
genetic compatibility of this species with arthemis/astyanax. Finally, as
noted in the review of concepts above, genitalic structure is not necessar-
ily a reliable index of relationships near the species boundary.

The Hardy- Weinberg analysis of phenotype ratios (Platt & Brower, op.
cit.) is flawed because it rests on pooling data from many subpopulations
or demes rather than one large interbreeding population. This results in
a tendency to record a lower number of heterozygotes than exist in the
population. This distortion can also result from observation of the same
population over long time periods. As noted by Platt (1983) this type of
criticism is inherent in the limitations of the Hardy- Weinberg test, and
is referred to as the Wahlund Effect (Hedrick, 1983, p. 284; Lewontin &
Cockerham, 1959). If one could determine the boundaries of an inter-
breeding population, the Hardy- Weinberg test for selection would be
more robust. However, the relative deficiency of intermediate pheno-
types (presumed heterozygotes) attributed to the Wahlund Effect by
Platt could also be ascribed to selection. The Hardy- Weinberg analysis as
a test for selection is compromised further by likely violation of one or
more of its assumptions: no gene flow, no effect from genetic drift, no
mutation, and the requirement for random mating. With regard to
natural selection, there is a self-contradiction in arguing for conspecificity
from a Hardy- Weinberg analysis in this case, since agreement with
predicted frequencies rests on the assumption of no selection, yet differ-
ential selection on wing morphs is what the authors are attempting to
prove. But to prove conspecificity Platt and Brower (op. cit.) try to
establish that the frequency of heterozygotes (inferred indirectly from
phenotype) does not differ from an expected value based on absence of
selection against hybrids, i.e. no postzygotic isolation. The contention
that the 160km wide overlap zone is an area with “no selection” on wing
phenotype is unlikely on general principles (Endler, 1986, espec. ch.4)
and unproven from the data. To postulate a condition of no selection in
this case requires that each phenotype class has the same survival
potential in the face of predation and all other agents of selection.

The Hardy-Weinberg test suffers from an insensitivity to certain types
of selection including reduced fecundity and sexual selection (Endler, op.
cit. p.65), either of which might be operating in nature in this case. A

68

J. Res. Lepid.

further complication arises from estimating heterozygote frequencies
from intermediate phenotypes, since a range of intermediate morphs
occur and modifier loci are postulated to operate in the expression of the
mimetic astyanax phenotype. Hardy- Weinberg in this application is
based on a simple one-locus Mendelian inheritance.

In contrast to Platt and Brower’s taxonomic arguments, the two forms
can be conspecific and the intergradation could also have arisen by
secondary contact. The two conditions are not mutually exclusive. There
seems to be no supporting evidence that an allopatric model is “unneces-
sarily complex” (Platt & Brower, op. cit.) compared to a model of primary
differentiation along a gradient of variable selection. Furthermore, the
selection regime (agent of selection, environment, population structure,
etc.) present during the evolution of wing pattern differences almost
certainly differs from the present situation. Pleistocene range changes
could have separated the Limenitis into allopatric populations. The
mimetic form could have evolved as an isolated population under condi-
tions of more intense selection than are experienced now on the average.

Indeed, a wide zone of intergradation would in itself suggest very weak
selection on mimetic versus disruptive wing patterns, which makes a
primary intergradation model less convincing. Although we know little
about dispersal rates in these butterflies, selection during the evolution
of the mimetic pattern in the face of gene flow would necessarily have to
be relatively strong to produce the genetic changes regulating the
coordinated expression of various pattern elements. It is also possible
that habitat alteration by man has increased the width of the blend zone.

In fact, the width of the intergrade zone in Limenitis is considerably
greater than average compared to those hybrid zones listed by Barton &
Hewitt (1985), which would tend to support the concept of minimal
genetic incompatibility in hybrids between arthemis and astyanax , in
agreement with the Platt and Brower model and in contradiction to the
two species - hybrid zone model of Remington (op. cit.). However, the
presence of intermediate and both “parental” phenotypes in local popu-
lations presents a difficulty. Random mating with no differential selec-
tion on adult phenotypes should break down the genetic basis for
parental phenotypes through recombination, especially if modifier loci
are involved in the expression of the astyanax phenotype. The persis-
tence of both the astyanax and arthemis phenotypes suggests a highly
“canalized” development such that expression of the wing pattern is
stable over a range of genotypes, and/or that selection has an effect on
phenotype frequencies.

Historical range changes must also be considered in understanding the
present blend zone. It is important to note that the narrow-banded
arthemis-\ike form “albofasciata” occurs quite far south well into the
range of B. philenor, and even south of the zone of hybridization ( Clark
& Clark, 1951; Platt, 1983; Shapiro, 1966; Shapiro, pers. comm.). The
persistence of the northern arthemis phenotype in these southern loca-

30(l-2):45-81, 1991

69

tions may represent the present effect of relict populations left behind as
changing climate in the Holocene allowed the Limenitis, and their zone
of intergradation, to move northward. Interestingly, a more abrupt cline
in arthemis/astyanax phenotype frequencies was found by Waldbauer et
al. (1988) in the Upper Peninsula of Michigan, but lake barrier effect
complicates interpretation of the relative importance of gene flow and
selection.

An alternative interpretation is that the blend zone represents range
overlap where genotypes producing both the arthemis and astyanax
phenotypes are adaptively superior to intermediate forms, but weak
mating barriers result in continuing hybridization between parental
forms. Some degree of genetic incompatibility appears to limit the genetic
recombination resulting from backcrossing and therefore hybrid-like
phenotypes are present in a lower frequency than in hybrid zones such
as the Hyalophora in the Sierra Nevada (discussed below). The unusu-
ally wide zone in this model is due to range overlap with limited
hybridization, in contrast to more typical narrow hybrid zones composed
of intermediate and recombinant phenotypes bounded by dines into
parental populations. As noted above, the unusual width of the zone may
also be due to historical range changes. This interpretation is similar to
that of Remington ( 1968b, 1985) in that it hypothesizes secondary contact
producing hybridization in a zone of range overlap. Unlike the Remington
model, I do not postulate the subsequent origin of reproductive isolation.
These Limenitis appear to be more genetically distinct than alternate
morphs in a polymorphism, but less divergent than full species.

Allozyme analysis would provide an independent measure of genetic
similarity between arthemis and astyanax and could detect the true
extent of overall gene exchange across the intergrade zone. Electrophore-
sis can provide unambiguous identification of heterozygotes, and if
significant or fixed allele frequency differences occur between pure
arthemis and astyanax reference populations, then the population
genetics of the zone of intergradation might be better understood.
Finally, allozyme analysis could produce an truer index of hybridity for
phenotypically intermediate specimens. As noted in the hybrid zone
discussion in Part I, hybrid zone phenotypes often do not accurately
reflect underlying genotypes.

Evolution in these Limenitis stands in contrast to the Papilio glaucus
/ canadensis situation in which glaucus seems to have evolved a mimetic
form, also based on the Battus philenor model, yet appears to be
reproductively isolated from the very similar, non-mimetic canadensis to
the north. The respective ranges of the two pairs are similar as arthemis
is nearly concordant with canadensis and astyanax is approximately
sympatric withglaucus. These shared distribution patterns in unrelated
groups are evidence of a common response to changing climates during
and after the Ice Age. Remington (1968b) hypothesizes that such
condordant “suture zones” formed as biota rejoined following a period of

70

J. Res. Lepid.

allopatric divergence as relict populations, in this case in refugia in the
southeastern United States.

Selection along a cline between alternate mimetic morphs in
Heliconius.

Heliconius erato and H. melpomene are brightly colored, unpalatable
tropical butterflies which occur sympatrically as Mullerian mimics. By
closely resembling each other, each species benefits because their com-
bined numbers in a population increase the rate at which predators learn
to avoid attacking any butterfly with the distinctive warning color
phenotype. By contrast, a Batesian mimic, which by definition is palat-
able, cannot theoretically exist in a population at a frequency higher than
its distasteful model. These Heliconius are especially interesting because
they exhibit parallel polymorphisms. Distinctive color morphs in differ-
ent populations of the more common H. erato are often accompanied by
similar geographic variation in the wing patterns of H. melpomene.
Mallet and Barton (1989ab) and Mallet et al. (1990) have extensively
studied the nature of selection on these shared polymorphisms in regions
of Peru where wing phenotype dines connect populations of distinctive,
alternate warning color patterns.

The population genetics of these dines represent a special case of
“frequency dependent selection” wherein for a given population the most
common phenotype (of either species) has a selective advantage over the
alternate, less frequent morph. This is true because predators are more
likely to encounter the more common morph. Since a certain minimum
number of encounters is needed for a “naive” predator to learn to avoid
a warning color pattern, the less common of two alternative morphs will
suffer disproportionate predation. Thus, given regions support popula-
tions of the two Heliconius which have very similar morphs at high
frequencies. Adjacent regions may have populations where the alternate
morph is present at a high frequency in both species. The two areas are
connected by very abrupt dines in wing morph type.

Mallet and Barton (1989a) estimated the strength of selection main-
taining these dines using marked foreign (experimental) and native
(control) butterflies which were released and subsequently recaptured at
intervals along dines between regions supporting alternate wing morph
populations. Their analysis showed that losses among odd morph
experimental individuals occurred soon after release, most probably due
to predation, since released butterflies did not disperse and tended to join
communal roosts irrespective of wing phenotype. Within each species,
populations with differing wing patterns showed no significant differ-
ences in allozymes, nor any important ecological distinctions. Alternate
wing morphs are controlled by three loci in H. erato and by four loci in H.
melpomene. The inference is that cline structure in each species is a
function of selection on these loci, rather than due to more complex hybrid
incompatibility or differing ecological adaptations. Estimates of the

30(l-2):45-81, 1991

71

coefficient of selection i n IL erato was quite high, about .17 per locus, or
about a 52% selection against experimental individuals released into
populations of the opposite warning color phenotype.

Because the genetics of this polymorphism are well understood, Mallet
and Barton (1989b) and Mallet et al (1990) were able to compare their
experimental data with simulation models of selection along dines. F rom
experimental studies they obtained allele frequencies for wing pattern
loci along dines between populations with differing wing morphs. In
their models, dine width can be expressed as a function of gene flow and
selection. Gene flow also causes a linkage correlation (linkage disequilibria)
between loci that differ along a dine. Thus, they used observed dine
width and linkage relationships to estimate gene flow and selection.
Estimates for selection generally agreed with those obtained from mark
and recapture studies. The usefulness of various computer simulations
of dines was confirmed. Moreover, these analyses tended to confirm
theoretical models of the evolution of Mullerian mimicry based on a few
loci with large effect.

Mallet and Barton investigated the maintenance of a dine between
genetically differentiated populations without directly addressing the
origin of these genetic differences. The approach of Platt and Brower to
the Limenitis astyanax/arthemis problem explicitly equated origin and
maintenance in terms of primary intergradation due to differential
predation on wing pattern morphs. However, the occurrence of parental
phenotypes at high frequency in the broad Limenitis blend zone suggests
that these forms are differentiated beyond the level of polymorphic
conspecifics.

The natural experimental system represented by the mimetic morph
dines in Heliconius was more amenable to study than the situation in
the Limenitis because 1) the genetics of the polymorphism were under-
stood; 2) the inference of selection based on predation was not subject to
uncertainty concerning the possible added effect of hybrid unfitness; 3)
allozyme data established that the taxa are conspecific; 4) selection was
measured directly by field experiment rather than by statistical infer-
ence from Hardy- Weinberg calculations; and 5) the width of the zone was
narrow enough that mark-recapture techniques could be used simulta-
neously across the entire zone. The narrow width was also partly due to
the strength of selection, which in turn aided the interpretation of the
mark-recapture data.

Natural Hybridization in the genus Hyalophora .

The members of this genus of large, attractive saturniid moths readily
hybridize in captivity, generally producing fully viable adults. Female
hybrids are usually barren, but males can be backcrossed and even three
or four nominate taxa have been combined in the genomes of lab hybrids.
Natural hybridization occurs to varying degrees in areas of congener
sympatry. By the Recognition Species Concept this genus would contain

72

J. Res. Lepid.

only a single species, although this viewpoint would ignore important
morphological, ecological, and genetic differences.

Sweadner (1937) was the first to study natural hybridization in
Hyalophora and his work was an early and important recognition that
traditional morphological criteria are inadequate in describing species
boundaries for groups such as the Hyalophora. We now know that this
genus contains a hierarchy of taxa as judged by the degree of reproductive
isolation among the various taxa. Tuttle (1985) showed through a series
of careful field tests in Michigan that the large, eastern H. cecropia is
partially isolated by seasonal and diurnal separation in flight activity
from interbreeding with the smaller, dark H. Columbia. Occasional
hybrids occur in nature (Collins, 1973; Ferge, 1983; Sweadner, op. cit.)
but no true hybrid zone occurs between these taxa.

By contrast, a zone of intergradation in Manitoba and Ontario connects
populations of Columbia with the larger, brighter colored H. gloveri which
occurs in the Canadian Prairie Provinces and south through the Rocky
Mts. and Great Basin. Both adult and larval phenotypes intergrade and
blend zone females oviposit on hosts of gloveri in addition to the conifer
Larix (tamarack), on which the eastern Columbia is a specialist (Collins,
op. cit.; Kohalmi & Moens, 1975, 1988). Laboratory hybrid females
between these taxa are typically fecund, in contrast to crosses between
cecropia and other congeners (Collins, op. cit.; unpub. data). Lemaire
(1978) synonymizes gloveri as a subspecies under Columbia.

Hyalophora euryalus on the west coast is quite distinct from gloveri in
all stages, yet the two species form a large hybrid zone on the east slope
of the Sierra Nevada south of Lake Tahoe in California (Collins, 1984).
Multivariate analysis shows that adult phenotypic variability in mid
hybrid zone greatly exceeds that seen in lab reference Fj hybrids;
extensive backcrossing and recombination appear to be responsible for
this variation, not merely the production of primary hybrids each season.
While crosses between widely separated population of euryalus and
gloveri produce barren female hybrids, females with intermediate or
recombinant phenotypes from the hybrid zone are fully fecund. More-
over, in test crosses using females derived from various sites along a
transect across the hybrid zone, genetic compatibility was optimal with
males from the source population, but decreased with males from more
distant populations, even with as little 15km separation. Collins (1984,
ms in prep.) interprets this result as evidence for the regional elimination
through selection of incompatible genotypes. Local optimization may be
aided by restrictions on gene flow due to mountainous topography,
although these moths are known to be quite vagile. Genetic compatibility
data on the fine structure of hybrid zones are rare, since most subject
organisms are not as easily experimentally hybridized. In spite of
obvious morphological differences between euryalus and gloveri in all
stages, relatively few loci controlling gametogenesis may regulate the
genetic structure of this hybrid zone. It would be important to verify this

30(l-2):45-81, 1991

73

model by means of allozyme or other biochemical genetic test of genetic
differentiation.

Sweadner (1937) attempted to document by use of a hybrid index the
existence of an intergrade zone in northern Idaho and western Montana
between gloveri on the east and euryalus on the west. Ferguson (1972)
treats this population, referred to as “kasloensis” , as a melanic northern
subspecies of euryalus, based primarily on genitalic structure. A recent
reanalysis (Collins, ms in prep.) has revealed that “ kasloensis ” is inter-
mediate and hybrid-like for several wing pattern characters, resembles
gloveri in early larval stages, yet possesses a unique mature larva
phenotype. Morewood (1991) illustrates a similar phenotype for British
Columbia “kasloensis” . The unusual red dorsal scoli pigmentation of
“kasloensis” may be the expression in a hybrid genome of a gene in the
fifth instar which is normally “turned on” only in the penultimate instar
of gloveri. The dorsal scoli of euryalus are yellow in both the 4th and 5 th
instars. The cocoon resembles that of a lab gloveri x euryalus hybrid.
Females from the intergrade zone are fully fertile, and have a decreased
compatibility when crossed with gloveri compared to near-normal fertil-
ity and viability in hybrids with euryalus. Judged by several criteria, the
“kasloensis” intergrade population seems to be of hybrid origin, but
appears to be restricted in gene exchange with gloveri to the east,
probably due to decreased host plant availability as a result of a rain
shadow effect of the Bitterroot Mts. along the Idaho-Montana border.
There is an abrupt transition to the gloveri phenotype to the east and a
more gradual intergradation into euryalus in British Columbia. Never-
theless, the “kasloensis” population maintains genetic integrity from the
swamping effects of gene flow from euryalus. It is unknown at present if
this equilibrium is due to intrinsic genetic compatibility factors or the
effect of ecological selection.

Two other hybrid populations occur between euryalus and gloveri in
Idaho, each much different from the “kasloensis” population (Collins,
unpubl. data). A hybrid swarm of great phenotypic variability occurs
northeast of Boise in Clear Creek Canyon. The adults more resemble the
range of phenotypes seen in the Sierra Nevada hybrid zone, which may
be due to a more balanced gene input from the two parental populations,
euryalus from the northern, panhandle region of Idaho, and gloveri from
near the Sun Valley/Ketchum area.

An extension of Great Basin habitat occurs in southeast Idaho and is
occupied by nominate gloveri which extends north through the Salmon
area to Lost Trails Pass, where the Bitterroots and the Continental
Divide merge. This pass appears too high to support a continuous
population of Hyalophora, but an occasional hybrid-like individual can
be taken in the gloveri population just to the south and an occasional
gloveri-\ike moth occurs in the “kasloensis” population just north of the
pass. Either limited dispersal occurs over the pass or a period of more
extensive gene exchange occurred during a warmer interglacial period.

74

J. Res. Lepid.

The hybrid Hyalophora populations in the northwest illustrate the fact
that unique population genetic factors acting in each situation have
produced three very different hybrid zones, regardless of the fact that the
same two species are interbreeding in each case. Topography and plant
community distribution, and historical climatic changes no doubt played
a role in shaping the present structure of these hybrid zones. By
extension, all the Hyalophora hybrid populations discussed here verify
the general premise that individual populations are the true units of
ecological and evolutionary change.

Acknowledgments. I would like to thank Lincoln Brower, Sterling Mattoon, Rudi
Mattoni, Paul Tuskes and Arthur Shapiro for their careful reviews of this
manuscript and Michael Turelli for his thoughtful discussion of hybrid zones and
related problems in population genetics.

Literature Cited

Arnold, R.A. 1983. Speyeria callippe (LepidopteriaiNymphalidae): Application of
information-theoretical and graph clustering techniques to analyses of
geographic variation and evaluation of classifications. Ann. Entomol. Soc.
Amer. 76:929-941.

. 1985. Geographic variation in natural populations of Speyeria callippe

(Boisduval) (Lepidoptera:Nymphalidae). Pan-Pacific Entomologist 61:1-23.
Avise, J.C., 1976. Genetic differentiation during speciation. pp.106-122. InF.J.Ayala
(ed.). Molecular Evolution. Sinauer, Sunderland, Mass.

Barton, N.H., 1989. Founder effect speciation, pp. 229-256. In Otte, D. & J.A.

Endler (eds.). Speciation and its Consequences. Sinauer, Sunderland, Mass.
Barton, N.H. & Charlesworth, B., 1984. Genetic revolutions, founder effects, and
specication. Annu. Rev. Ecol. Syst. 15:133-164.

Barton, N.H. & G.M. Hewitt. 1981. Hybrid zones and speciation. pp. 109-145. In
W.R. Atchley & D.S. Woodruff (eds.). Evolution and Speciation. Cambridge
Univ. Press, Cambridge, Mass.

1985. Analysis of hybrid zones. Annu. Rev. Ecol. Syst. 16:113-148.

. 1989. Adaptation, speciation and hybrid zones. Nature 341:497-503.

Berven, K.A. & D.E. Gill. 1983. Interpreting geographic variation in life- history
traits. Amer. Zool. 23:85-97.

Bigelow, R.S. 1965. Hybrid zones and reproductive isolation. Evolution 19:449-
458.

Britten, R.J. & E.H. Davidson. 1969. Gene regulation for higher cells: A theory.
Science 165:349-357.

Brittnacher, J.G., S.R. Sims & F.J. Ayala. 1978. Genetic differentiation between
species of the genus Speyeria (Lepidoptera: Nymphalidae). Evolution 32:199-
210.

Bowers, M.D. 1986. Population differences in larval host plant use in the checkerspot
butterfly Euphydryas chalcedona . Entomol. Exp. App. 40:61-69.

Brower, J.V.Z. 1958. Experimental studies of mimicry in some North American
butterflies. II. Battus philenor and Papilio troilus, P. polyxenes and P. glaucus.
Evolution 12:123-136.

Brower, L.P. 1959a. Speciation in butterflies of the Papilio glaucus group. I.
morphological relationships and hybridization. Evolution 13:42-63.

30(l-2):45-81, 1991

75

. 1959b. Speciation in butterflies of the Papilio glaucus group. II. Ecological

relationships and interspecific sexual behavior.

Brower, L.P. & J.V.Z. Brower. 1962. The relative abundance of model and mimic
butterflies in natural populations of the Battus philenor complex. Ecology
43:154-158.

Bush, G.L., 1969. Sympatric host race formation and speciation in frugivorous flies
of the genus Rhagoletis (DipteraiTephritidae). Evolution 23:237-251.

. 1975. Modes of animal speciation. Annu. Rev. Ecol. Syst. 6:339-364.

. 1982. What do we really know about speciation? pp. 119-128. In R. Milkman

(ed.). Perspectives on Evolution. Sinauer, Sunderland, Mass.

Butlin, R., 1987. Anew approach to sympatric speciation. Trends Ecol. Evol. 2:310-
311.

. 1989. Reinforcement of premating isolation, pp. 158-179. In D. Otte & J.A.

Endler (eds.). Speciation and its Consequences. Sinauer, Sunderland, Mass.

Byers, J.R. & J.D. LaFontaine. 1982. Biosystematics of the genus Euxoa
( Lepidoptera :Noctuidae ) XVI . Comparative biology and experimental taxonomy
of four subspecies of Euxoa comosa. Can. Ent. 114:551-565.

Garde, R.T. 1987. The role of pheromones in reproductive isolation and speciation
in insects, pp. 303-317. In M. D. Huettel (ed.). Evolutionary Genetics of
Invertebrate Behavior. Plenum.

Garde, R.T. & T.C. Baker. 1984. Sexual communication with pheromones, pp.355-
383. In W.J. Bell & R.T. Garde (eds.). Chemical Ecology of Insects. Chapman &
Hall.

Carson, H.L. 1968. The population flush and its genetic consequences, pp. 123-137.
In R.C. Lewontin (ed. ). Population Biology and Evolution. Syracuse Univ.Press,
New York.

. 1982. Speciation as a major reorganization of polygenic balances, pp. 411-

433. In A.R. Liss (ed.). Mechanisms of Speciation. Liss. N.Y.

Carson, H.L. & K.Y. Kaneshiro. 1976. Drosophila of Hawaii: Systematics and
ecological genetics. Annu. Rev. Ecol. Syst. 7:311-346.

Carson, H.L. & A.R. Templeton. 1984. Genetic revolutions in relation to speciation
phenomena: the founding of new populations. Annu. Rev. Ecol. Syst. 15:97-131.

Charlesworth, B., J.A. Coyne & N.H. Barton. 1987. The relative rates of evolution
of sex chromosomes and autosomes. Am. Nat. 130:113-146.

Clark, A.H. &L.F. Clark. 1951. The butterflies of Virginia. Smithsonian Misc. Coll.
116:1-239.

Clarke, C.A. & P.M.Sheppard. 1955. The breeding in captivity of the hybrid Papilio
rutulus female x Papilio glaucus male. Lepid. News 9:46-48

. 1957. The breeding in captivity of the hybrid Papilio glaucus female X

Papilio eurymedon male. Lepid. News 11:201-205.

. 1962. The genetics of the mimetic butterfly Papilio glaucus. Ecology 43 : 159-

161.

Clarke, C.A., U. Mittwoch & W. Traut. 1977. Linkage and cytogenetic studies in
the swallowtail butterflies Papilio polyxenes Fab. and Papilio machaon L. and
their hybrids. Proc. R. Soc. Lond. 198:385-399.

Clarke, C.A. & A. Willig. 1977. The use of alpha ecdysone to break permanent
diapause of female hybrids between Papilio glaucus L. female and Papilio
rutulus Lucas male. J. Res. Lepid. 16:245-248.

Collins, M.M. 1973. Notes on the taxonomic status of Hyalophora Columbia
(Saturniidae). J. Lepid. Soc. 27:225-235.

76

J. Res. Lepid.

. 1984. Genetics and ecology of a hybrid zone in Hyalophora

(Lepidoptera:Saturniidae). Univ. Calif. Publ. Entomol. 104:1-93.

Coyne, J.A. & H.A. Orr. 1989a. Patterns of speciation in Drosophila. Evolution
43:362-381.

. 1989b. Two rules of speciation, pp. 180-207. In D. Otte & J.A. Endler (eds.),

Speciation and its Consequences. Sinauer, Sunderland, Mass.

Coyne, J.A., H.A. Orr & D.J. Futuyma. 1988. Do we need a new species concept?
Syst. Zool. 37:190-200.

Diehl, S.R. G.L. Bush. 1989. The role of habitat preference in adaptation and
speciation. pp. 345-365. In D. Otte & J.A. Endler (eds.), Speciation and its
Consequences. Sinauer, Sunderland, Mass.

Dobzhansky, T., 1937. Genetics and the Origin of Species. Columbia Univ. Press,
New York. 364 p.

, 1970. Genetics and the Evolutionary Process. Columbia Univ. Press. New

York. 505 p.

Ehrlich, P.R. &D.D. Murphy. 1981. The population biology of checkerspot butterflies
(. Euphydryas ) - a review. Biol. Zentral. 100:613-629.

Ehrlich, P.R. & P.H Raven. 1969. Differentiation of populations. Science 165:1228-
1232.

Ehrlich, P.R. & R.R. White. 1980. Colorado checkerspot butterflies: Isolation,
neutrality, and the biospecies. Am. Nat. 115:328-341.

Ehrman, L., 1962. Hybrid sterility as an isolating mechanism in the genus
Drosophila . Quart. Rev. Biol. 37:279-302.

Endler, J.A. 1977. Geographic variation, speciation, and dines. Princeton Univ.
Press, Princeton, New Jersey.

. 1986. Natural Selection in the Wild. Princeton Univ. Press. Princeton, N. J.

336.

. 1989. Conceptual and other problems in speciation. pp. 625-648. In D. Otte

& J.A. Endler (eds.). Speciation and its Consequences. Sinauer, Sunderland,
Mass.

Endler, J.A. & T. McClellan. 1988. The processes of evolution: toward a newer
synthesis. Annu. Rev. Ecol. Syst. 19:325-321.

Felsenstein, J. 1981. Skeptacism towards Santa Rosalia, or why are there so few
kinds of animals? Evolution 35:124-138.

Ferge, L.A. 1983. Distribution and hybridization of Hyalophora Columbia
(Lepidoptera: Saturniidae) in Wisconsin. Great Lakes Entomol. 16:67-71.

Ferguson, D.C. 1972. In R.B. Dominick, et al. The moths of America north of
Mexico, Fasc. 20. 2B, Bombycoidea: Saturniidae (in part):243-275.

Forbes, W.T.M. 1954. Presidential address to Los Angeles meeting of the
Lepidopterist’s society. Lepid. News 8:1-4.

Fox, L.R. &P. A. Morrow. 1981. Specialization: species property or local phenomenon?
Science 211:887-893.

Futuyma, D., 1986. Evolution, 2nd ed. Sinauer, Sunderland, Mass. 600 p.

Futuyma, D. & G.C. Mayer. 1980. Non-allopatric speciation in animals. Syst. Zool.
29:254-271.

Gall, L.F. & F.A.H. Sperling. 1980. A new high altitude species of Boloria from
southwestern Colorado (Nymphalidae), with a discussion of phenetics and
heirarchical decisions. J. Lepid. Soc. 34:230-252.

Geiger, H.J., 1988. Enzyme electrophoresis and interspecific hybridization in
Pieridae (Lepidoptera)-the case for enzyme electrophoresis. J. Res. Lepid.
26:64-72.

30(l-2):45-81, 1991

77

Geiger, H.J. & Scholl, A., 1985. Systematics and evolution of the holarctic
Pierinae: an enzyme electrophoretic approach. Experimentia 41:24-29.

Geiger, H.J. & A.M. Shapiro. 1986. Electrophoretic evidence for speciation within
the nominal species Anthocaris sara Lucas (Pieridae). J.Res.Lepid. 25:15-24.

Gilbert, L.E. 1983. Coevolution and mimicry, pp. 263-281. In D.J. Futuyma & M.
Slatkin (eds.). Coevolution. Sinauer, Sunderland, Mass. pp. 263-281.

Gillham, N. W. 1956. Geographic variation and the subspecies concept in butterflies.
Syst. Zool. 5:110-120.

Grula, J.W. & O.R. Taylor. 1980a. Some characteristics of hybrids derived from the
sulfur butterflies, Colis eury theme and C. philodice : phenotypic effects of the X
chromosome. Evolution 34:673-687.

. 1980b. The effect of X-chromosome inheritance on mate-selection behavior

in the sulfur butterflies, Colias eurytheme and C. philodice. Evolution 34:688-
695.

Hafernik, J. 1982. Phenetics and ecology of hybridization in buckeye butterflies
(Lepidoptera:Nymphalidae). Univ. Calif. Publ. Entom. 96:1-109, Univ. Calif.
Press, Berkeley.

Hagen, R.H. 1990. Population structure and host use in hybridizing subspecies of
Papilio glaucus (Lepidoptera:Papilionidae). Evolution 44:1914-1930.

Hagen, R.H. & J.M. Scriber. 1989. Sex-linked diapause, color, and allozyme loci in
Papilio glaucus: Linkage analysis and significance in a hybrid zone. J. Heredity
80:179-185.

. 1991. Systematics of Papilio glaucus and Papilio troilus species groups

(Lepidoptera:Papilionidae): inference from allozymes. Annals Entomol. Soc.
Am. 84:380-395.

Hammond, P.C. 1985(1986). Opinion: a rebuttal to the Arnold classification of
Speyeria callippe (Nymphalidae) and defense of the subspecies concept. J. Res.
Lepid. 24:197-208.

. 1990(91). Patterns of geographic variation and evolution in polytypic

butterflies. J. Res. Lepid. 29:54-76.

Harrison, R.G. & D.M. Rand. 1989. Mosaic hybrid zones and the nature of species
boundaries, pp. 111-133. In D. Otte & J.A. Endler (eds). Speciation and its
consequences. Sinauer, Sunderland, Mass.

Hedrick, P.W. 1983. Genetics of Populations. Science Books International. Boston.
629p.

Hedrick, P.W. & J.F. McDonald. 1980. Regulatory gene adaptations: an evolutionary
model. Heredity 45:83-97.

Hewitt, G.M. 1989. The subdivision of species by hybrid zones, pp. 85-110. In
D.Otte & J.A. Endler (eds.). Speciation and its Consequences, Sinauer, Mass.

. 1990. Divergence and speciation as viewed from a hybrid zone. Can. J. Zool.

68:1701-1715.

Janzen, D.H., 1984. Weather-related color polymorphism of Rothschildia lebeau
(Saturniidae). Bull. Entmol. Soc. Amer. 30:16-20.

Johnson, G.B. 1976. Enzyme polymorphism and adaptation in alpine butterflies.
Ann. Missouri Bot. Gard. 63:248-261.

Kingsolver, J.G. & D.C. Wiernasz. 1991. Development, function, and the quantitative
genetics of wing melanin pattern in Pieris butterflies. Evolution 45:1480-1492.

Kohalmi, L. & P. Moens. 1975. Evidence for the existence of an intergrade
population between Hyalophora gloveri nokomis and H. Columbia in
northwestern Ontario (Lepidoptera:Saturniidae). Can. Ent. 107:793-799.

78

J. Res. Lepid.

. 1988. Pattern of inheritance of a larval tubercle-colour polymorphism in

Hyalophora Columbia of northwestern Ontario (Lepidoptera: Saturniidae).
Genome 30:307-312.

Lemaire, C. 1978. The Attacidae of America. Edition C. Lemaire. Paris. 238p.
Lewontin, R.C., 1974. The Genetic Basis of Evolutionary Change. Columbia Univ.
Press, New York. 346p.

Lewontin, R.C. & C.C. Cockerham. 1959. The goodness-of-fit test for detecting
selection in random mating populations. Evolution 13:561-564.

Lorkovic, Z. 1985(86). Enzyme electrophoresis and interspecific hybridization in
Pieridae (Lepidoptera). J. Res. Lepid. 24:334-358.

Luebke, H. 1985. Hybridization in the Papilio glaucus group (Lepidoptera): a
morphometric study using multivariate techniques, unpubl. MS thesis. Univ.
Wisconsin, Dept. Entomol. Madison.

Luebke, H.J., J.M. Scriber & B.S. Yandell. 1988. Use of multivariate discriminant
analysis of male wing morphometries to delineate the Wisconsin hybrid zone for
Papilio glaucus glaucus and P.g. canadensis. Am. Midi. Nat. 119:366-379.
Mallet, J. & N. Barton. 1989a. Strong natural selection in a warning-color hybrid
zone. Evolution 43:421-431.

. 1989b. Inference from dines stabilized by frequency-dependent selection.

Genetics 122:967-976.

Mallet, J., N. Barton, G. Lamas M., J. Satisteban C., M. Muedas M., & H. Eeley.
1990. Estimates of selection and gene flow from measures of cline width and
linkage disequilibrium in Heliconius hybrid zones. Genetics 124:921-936.
Mayr, E., 1963. Animal Species and Evolution. Belknap Press, Cambridge.

, 1982. Processes of speciation in animals, pp. 1-19. In C. Barigozzi (ed.).

Mechanisms of Speciation. Liss, N.Y.

Mayr, E. & D. P. Ashlock. 1991. Principles of Systematic Zoology. 2nd. Ed.
McGraw-Hill. 475p.

McKechnie, S.W., P.R. Ehrlich & R.R. White. 1975. Population genetics of
Euphydryas butterflies. I. Genetic variation and the neutrality hypothesis.
Genetics 81:571-594.

Moore, J.A., 1957. An embryologist’s view of the species concept. In E. Mayr, The
species problem, pub. 50. Amer. Assoc. Adv. Sci. Washington D.C.

Morewood, W.D. 1991. Larvae of Hyalophora euryalus kasloensis (Lepidoptera:

Saturniidae). J. Entomol. Soc. Brit. Columbia 88:31-33.

Murphy, D.D. & P.R. Ehrlich. 1984. On butterfly taxonomy. J. Res. Lepid. 23:19-
34.

Oliver, C.G. 1972. Genetic and phenotypic differentiation and geographical distance
in four species of Lepidoptera. Evolution 26:221-241.

. 1978. Experimental hybridization between the nymphalid butterflies

Phyciodes tharos and P. montana. Evolution 32:594-601.

. 1979a. Genetic differentiation and hybrid viability within and between

some Lepidoptera species. Amer. Nat. 114:681-694.

. 1979b. Experimental hybridization between Phyciodes tharos and P.

batesii (Nymphalidae). J. Lepid. Soc. 33:6-20.

. 1980. Phenotypic differentiation and hybrid breakdown within Phyciodes

“tharos” (Lepidoptera: Nymphalidae) in the Northeastern United States. Ann.
Entomol. Soc. Am. 73:715-721.

Otte, D. & J.A. Endler (eds.).1989. Speciation and its Consequences. Sinauer,
Sunderland, Mass. 679 p.

30(l-2):45-81, 1991

79

Parker, H.R., D.P. Phillip & G.S. Whitt. 1985. Gene regulatory divergence among
species estimated by altered developmental patterns in interspecific hybrids.
Mol. Biol. Evol. 2:217-250.

Paterson, H.E.H., 1986. The recognition concept of species. In E.S. Vrba (ed.),
Species and Speciation. Transvaal Museum Monograph No. 4, Pretoria.

Platt, A.P. 1975. Monomorphic mimicry in Nearctic Limenitis butterflies:
experimental hybridization of the arthemis-astyanax complex withL. archippus.
Evolution 29:120-141.

. 1983. Evolution of North American admiral butterflies. Bull. Entomol. Soc.

Am. 29:10-22.

Platt, A.P. & L.P. Brower. 1968. Mimetic versus disruptive coloration in intergrading
populations of Limenitis arthemis and astyanax butterflies. Evolution 22:699-
718.

Platt, A.P., R.P. Coppinger & L.P. Brower. 1971. Demonstration of the selective
advantage of the mimetic Limenitis presented to caged avian predators.
Evolution 25:692-701.

Platt, A.P., G.W. Rawson & G. Balogh. 1978. Interspecific hybridization involving
Limenitis archippus and its congeneric species (Nymphalidae). J. Lepid. Soc.
32:239-303.

Porter, A.H. & Geiger, H., 1988. Genetic and phenotypic population structure of
the Coenonympha tullia complex (Lepidoptera: Nymphalidae: Satyrinae) in
California: no evidence for species boundaries. Can. J. Zool. 66:2751-2765.

Porter, A.H. & A.M. Shapiro. 1990. The lock-and-key hypothesis: evolutionary and
biosystematic interpretation of insect genitalia. Annu. Rev. Entomol. 34:231-
245.

Rabb, R.L. 1966. Diapause in Protoparce sexta (Lepidoptera: Sphingidae). Ann.
Entomol. Soc. Amer. 59:160-169.

Remington, C.L. 1968a. A new sibling Papilio from the Rocky Mountains, with
genetic and biological notes (Insecta, Lepidoptera). Postilla 119:1-40.

. 1968b. Suture-zones of hybrid interaction between recently joined biotas.

Evolutionary Biol. 2:321-428.

. 1985. Genetical differences in solutions to the crises of hybridization and

competition in early sympatry. Boll. Zool. 52:21-43.

Rice, W.R. 1985. Disruptive selection on habitat preference and the evolution of
reproductive isolation: an exploratory experiment. Evolution 39:645-656.

Rockey, S.J., J.H. Hainze & J.M. Scriber. 1987. A latitudinal and obligatory
diapause response in three subspecies of the eastern tiger swallowtail, Papilio
glaucus (Lepid. :Papilionidae). Am. Midi. Nat. 118:162-168.

Rutowski, R.L. 1984. Sexual selection and the evolution of butterfly mating
behavior. J. Res. Lepid. 23:125-142.

Sage, R.D. & R.K. Selander. 1979. Hybridization between species of the Rana
pipiens complex in central Texas. Evolution 33:1069-1088.

Schafer, U., 1978. Sterility in Drosophila hydei x neohydei hybrids. Genetica
49:205-214.

Scriber, J.M. 1983. Evolution of feeding specialization, physiological efficiency,
and host races in selected Papilioniidae and Saturniidae. pp. 373-412. In R.F.
Denno & M.S. McClure (eds.). Variable Plants and Herbivors in Natural and
Managed Systems. Academic Press, New York.

. 1984. Host plant suitability. In W.J. Bell & R.T. Garde (eds.) Chemical

Ecology of Insects. Chapman and Hall, p.159-202.

80

J. Res. Lepid.

Scriber, J.M. & M.H. Evans. 1988. Bilateral gynandromorphs, sexual and/or color
mosaics in the tiger swallowtail butterfly, Papilioglaucus. J. Res. Lepid. 26:39-
57.

Scriber, J.M. , M.H. Evans, & D. Ritland. 1987. Hybridization as a causal mechanism
of mixed color broods and unusual color morphs of female offspring in the
eastern tiger swallowtail butterfly, Papilioglaucus. In Evolutionary genetics of
invertebrate behavior. M. Huettel (ed.) Univ. Florida Press. Gainesville, p. 119-
134.

Scott, J.A. 1986. The butterflies of North America. Stanford Univ. Press. Stanford.
583p.

Shapiro, A.M. 1966. Butterflies of the Delaware Valley. Special Publication,
American Entomological Society. 79p.

1974. Butterflies and Skippers of New York State. Search. 4:1- 59.

. 1984a. Polyphenism, phyletic evolution, and the structure of the Pierid

genome. J. Res. Lepid. 23:177-195.

. 1984b. The genetics of seasonal polyphenism and the evolution of general-

purpose genotypes in butterflies. In K. Wohrman & K. Loschke (eds.), Population
Biology and Evolution. Springer-Verlag, New York, pp. 16-30.

Shapiro, A.M. & H. Geiger. 1986. Electrophoretic confirmation of the species status
of Pontia protodice and P. occidentalis (Peiridae). J. Res. Lepid. 25:39-47.

Slatkin, M., 1973. Gene flow and selection in a cline. Genetics 75:735-756.

, 1987. Gene flow and the geographic structure of natural populations.

Science 236:787-792.

Spencer, H.G., B.H. McArdle, & D.M. Lambert. 1986. A theoretical investigation
of speciation by reinforcement. Amer. Natur. 128:241-262.

Sperling, F.A.H. 1987. Evolution of the Papilio machaon species group in western
Canada (Lepidoptera:Papilionidae). Quaest. Entomol. 23: 198-315.

. 1990. Natural hybrids of Papilio (Insecta:Lepidoptera): poor taxonomy or

interesting evolutionary problem? Can. J. Zool. 68:1790-1799.

Stebbins, G.L. 1974. Flowering plants, evolution above the species level. Belknap
Press of Harvard Univ. Press, Cambridge, Mass.

Sved, J.A., 1979. The “hybrid dysgenesis” syndrome in Drosophila melangaster.
Biosci. 29:659-664.

Sweadner, W.R. 1937. Hybridization and the phylogeny of the genus Platysamia.
Ann. Carnegie Mus. 25:163-242.

Tauber, C.A. & M. J. Tauber. 1989. Sympatric speciation in insects: perception and
perspective, pp. 307-344. In Otte, D. & J.A. Endler (eds.). Speciation and its
Consequences. Sinauer, Sunderland, Mass.

Taylor, O.R. 1972. Random vs. non-random mating in the sulfur butterflies, Colias
eurytheme and Colias philodice (Lepidoptera: Pieridae). Evolution:26:357-365.

Templeton, A.R. 1980. The theory of speciation via the founder principle. Genetics
94:1011-1038.

, 1981. Mechanisms of speciation - a population genetic approach. Annu.

Rev. Ecol. Syst. 12:23-48.

, 1982. Genetic architectures of speciation. pp. 105-121. In A.R. Liss (ed.).

Mechanisms of Speciation. Liss, N.Y.

, 1989. The meaning of species and speciation. pp. 3-27. In D. Otte & J.A.

Endler (eds.). Speciation and its Consequences. Sinauer, Sunderland, Mass.

30(l-2):45-81, 1991

81

Thoday, J.M. & J.B. Gibson. 1962. Isolation by disruptive selection. Amer. Natur.
104:219-230.

Thompson, J.N. 1988a. Variation in preference and specificity in monophagous and
oligophagous swallowtail butterflies. Evolution 42:118-128.

. 1988b. Evolutionary genetics of oviposition preference in swallowtail

buttflies. Evolution 42:1223-1234.

Tuskes, P.T. & M.M. Collins. 1981. Hybridization of Saturnia mendocino and S.
waiter or urn, and phylogenetic notes on Saturnia and Agapema (Saturniidae).
J. Lepid. Soc. 35:1-21.

Tuttle, J.P. 1985. Maintaining species integrity between sympatric populations of
Hyalophora cecropia and Hyalophora Columbia (Saturniidae) in central
Michigan. J. Lepid. Soc. 39:65-84.

Vane-Wright, R.I. & P.R. Ackery. 1989. The Biology of Butterflies. Princeton. 429p.

Wake, D.B. 1981. The application of allozyme studies to problems in the evolution
of morphology, pp. 257-270. In G.G.E. Scudder and J.L. Reveal, eds. Evolution
Today. Proc. 2nd inti. Cong. Syst. & Evol. Biol.

Waldbauer, G.P. 1978. Phenological adaptation and the polymodal emergence
patterns of insects. In H. Dingle (ed.) Evolution of Insect Migration and
Diapause. Springer- Verlag, New York.

Waldbauer, G.P., J.G. Sternburg & A.W. Ghent. 1988. Lakes Michigan and Huron
limit gene flow between the subspecies of the butterfly Limenitis arthemis. Can.
J. Zool. 66:1790-1795.

Watt, W.B. 1968. Adaptive significance of pigment polymorphisms in Colias
butterflies. I. variation of melanin pigment in relation to thermoregulation.
Evolution 22:437-458.

West, D.A. & C.A. Clarke. 1988. Suppression of the black phenotype in females of
P. glaucus (Papilionidae). J. Res. Lepid. 26:187-200.

Wilson E.O. & W.L. Brown. 1953. The subspecies concept and its taxonomic
application. Systematic Zoology 2:97-111.

Woodruff, D.S. 1973. Natural hybridization and hybrid zones. Syst. Zool. 22:213-
218.

. 1981. Toward a genodynamics of hybrid zones: studies of Australian frogs

and West Indian land snails, pp. 171-197. In W.T. Atchley & D.S. Woodruff
(eds.). Evolution and Speciation. Cambridge Univ. Press, Cambridge, Mass.

Wright, S., 1931. Evolution in Mendelian populations. Genetics 16:97-159.

Zink, R.M. 1988. Evolution of brown towhees: allozymes, morphometries and
species limits. Condor 90:72-82.

NOTE ADDED IN PROOF: In a recent paper, Hagen et al. (1991)

corroborate the full species status of P. canadensis using physiological

and genetic data. Allozyme data indicate limited hybridization across a

very narrow zone in Michigan.

Hagen, R.H., R. C. Lederhouse, J.L. Bossart & J. M. Scriber. 1991. Papilio
canadensis and P. glaucus (Papilionidae) are distinct species. J. Lepid. Soc.
45:245-258.

Journal of Research on the Lepidoptera

30(l-2):82-94, 1991

Quantification of Lepidoptera wing patterns using
an image analyzer.

Jack J. Windig

Section of Evolutionary Biology, Department of Population Biology, University of Leiden,
Schelpenkade 14A, 2313 ZT Leiden, The Netherlands.

Abstract. The use of an image analyser for measuring butterfly wings
is discussed. As an example the measurement of the phenotypic
plasticity in the wing pattern of a tropical butterfly is described in
detail. A set of 7 characters was selected to optimize the complete
description of the plasticity with regard to both accuracy and speed of
measurement. Several other examples of the measurement of Lepi-
dopteran wing patterns with an image analyzer are given. The advan-
tages and disadvantages of an image analyser are discussed. An image
analyser can be an effective tool, especially for complex measurements,
but the development of software is critical.

Introduction

The wing patterns of butterflies and moths have been studied in many
biological disciplines (for example in genetics (Brakefield 1984, Robinson
1990), developmental biology (Nijhout, 1985, 1991), evolutionary ecology
(Endler 1984, Kingsolver 1987). The traditional technique used is to
quantify the patterns into classes, as estimated by eye. This may be
adequate for discrete morphs but is likely to introduce bias for a more
complex variation (see e.g. Brakefield & Dowdeswell 1985). A more
detailed quantification of particular pattern elements can sometimes be
made using a microscope fitted with a micrometer (e.g. Ehrlich & Mason
1966, Mason et al. 1968, Bowers et al. 1985).

Computers and electronic measuring instruments can provide a more
accurate analysis. Examples are a color analyser (Brakefield & Liebert
1985), and a digitizing pad (Strauss 1990, Kingsolver & Wiernasz 1991).
An image analyser combines the functions of most electronic measuring
instruments. Typically it consists of a TV-camera, a computer and an
image processor (often called a frame grabber) that converts the TV-
signal into digital form suitable for the computer (Gonzalez & Wintz
1987). Until recently image analysers existed only with mainframe or
minicomputers and its use was hardly accessible to entomologists. Daly
(1985) concluded that image analysis was not fully developed and of
limited use in entomology.

Image analysers which can be built into a personal computer are now
readily available to entomologists. Detailed measurements are possible
with such a system, not only of distances but also of areas. In addition it
also enables analysis of color intensities. Specific parts of an image can
be analyzed separately. Simple (e.g. subtractions) or complex (e.g. Fou-
rier analysis for shape) calculations can be performed on the image. Use

30(l-2):82-94, 1991

83

by entomologists is still very restricted (e.g. Hagerup et al. 1990) but
growing. This paper reports on the usefulness of an image analyser for
measuring Lepidoptera wings in general.

Anyone interested in using such a system will probably ask two
questions:

1) Does it enable the researcher to approach a particular problem in a
novel and enlightening way?

2) Does the gain of using an image analyser counter the costs?

We used three criteria to evaluate these questions (in order of impor-
tance):

• completeness: A procedure must adequately quantify the pattern
with respect to all characters of the wing. Colors or color intensities for
example often vary within and between species and in these cases
measurements cannot be restricted to sizes.

• repeatability: A procedure is needed which provides a highly repeat-
able result so that wings can be measured at different times and by
different people without influencing the conclusions.

• speed: A procedure is needed which is fast, so that sufficiently large
samples can be measured to obtain statistical reliable results.

This paper describes in detail the measurement of wings of Bicyclus
safitza (Hewitson 1851) with an image analyser, and the difficulties
encountered when developing the measurement procedure. With regard
to the last criterium (speed) an additional question is addressed: must all
characters be measured or can a subset of characters be used to describe
the whole pattern of variation? Some other uses of the image analyser in
our laboratory are outlined before discussing the general questions
asked in the previous paragraph.

MEASUREMENT OF B. SAFITZA WINGS WITH THE IMAGE
ANALYSER

B. safitza is a satyrine butterfly which occurs throughout tropical
Africa (Condamin 1973), and exhibits marked phenotypic plasticity in
the form of seasonal polyphenism. The form occurring in the wet season
has conspicuous submarginal eyespots and a median white band on its
ventral wing surface. The dry season form lacks these pattern elements
and is essentially a plain brown butterfly. Intermediates between these
extremes can sometimes be found in the field.

Before the image analyser was available to us, forms were evaluated by
ranking them into different classes as estimated by the eye, and by
measuring the wings with a microscope fitted with a micrometer
(Brakefield &Reitsma 1991). With the image analyser we tried to include
different measurements (e.g. of color) and more accurate measurements
to enable us to estimate the “dry/wetness” of the wings more rigorously.
Furthermore, we needed faster measurements for (especially in quanti-
tative genetical analysis) generally large or very large sample sizes are
needed to obtain statistically significant results (see e.g. Shaw 1987).

84

J. Res. Lepid.

Fig. 1 Set up of the image analyzing system

Material and methods.

Measurements. Our image analysis system (fig. 1) was supplied by DIFA
Measurement Inc. It consists of a black/white video camera (HTH MO), an IBM-
AT compatible Personal computer (Epson PC AX2) with a frame grabber (PC-
Vision plus) in one of its additional slots. The video signal is converted by the
frame grabber to a digital image. The image is made up of 512x512 pixels, each
pixel having a grey value between 0 (=black) and 255 (=white). An additional color
monitor displays images; the PC monitor shows the software instructions of an
interactive package called TIM (developed by TEA Inc.).

To evaluate the wing pattern of the Bicyclus butterflies, a measurement
procedure was developed with regard to the three criteria described in the
introduction. Software for the measurements was developed partly by DIFA Inc.,
but for the greater part by ourselves. Results of measurements were written to
ASCII files and could easily be used in other (e.g. statistical) programs. Each
criterium had its own requirements and the software needed continuous updat-
ing. Difficulties and their solutions will be described separately for each criterium.

Completeness. To obtain a range of butterflies from dry to wet season forms,
final instar larvae were raised at six different temperatures. The temperatures
used were 14°C, 15°C, 17°C, 20°C, 23°C and 28°C, similar to the range of
temperatures found in Malawi, the origin of our stocks. Five males and five
females were selected at random from the butterflies raised at each temperature.
“Dryness” of the butterflies was estimated by eye by ranking them into six
classes, from 0 (=dsf, no eyespots and no white band) to 5 (=wsf, large white band
and large eyespots). The ranking was done independently by two persons
resulting in a nearly identical classification (Spearman rank r= 0.942).

With the image analyser we were able to measure almost any wing pattern
character we could think of. A large set containing all potentially important
characters of the ventral side was measured for each butterfly (fig. 2). An ANOVA
was performed on each of these 67 characters to detect significant differences (at
a = 0.01) between the six seasonal classes and the sexes. All the areas covaried
with wing size and for these an ANCOVA instead of an ANOVA was used.
Principal Component Analysis (PCA) was used to provide a more comprehensive
picture. Partial correlations with respect to total wing size were used for

30(l-2):82-94, 1991

85

+

*

Nr.

Object

Measurements:

Nr.

Object

1,11

total wing 1

1

3-6

chevron

r area, contrast area &

7-18

eyespots

II -VIII

wing fields *

1 contrast level

19-22

white band1

23

white band

1 -2

fringe

area

24-32

white band

Measurements:

area & grey values
distance

grey value (for 32
range & st. dev. of
grey values)

Fig. 2 Characters measured for analysis of B. safitza wings.

correlations between two area measurements (other than wing area). A mixed
matrix of correlations and partial correlations was used for PCA.

Repeatability. A subset of five females was selected to examine repeatability.
This sample included two of the most extreme butterflies (1 wsf, 1 dsf) and three
intermediates. Characters measured were similar to those of the completeness
test. The characters were measured by six different persons, ranging from
inexperienced to experienced computer users, and from persons both unfamiliar
and familiar with the butterflies. They were all untrained in the specific software
program. Coefficients of variation for all six sets of measurements were calcu-
lated for each character measured in each wing. The means of these coefficients
over all sets of five wings were calculated as an index of repeatability.

Speed. Speed of measurement was evaluated when measuring the wings for
the completeness test and for the repeatability test.

Weights for PC 2

86

J. Res. Lepid.

Weights for PC 1

Fig. 3 Weightings for first two principal components in PCA of all characters.
Significance of differences in the characters in a multiple ANOVA for sexe
and seasonal form are indicated. Names of the groups as in Fig. 2. (E =
Eyespot field, M = Median field, B = Basal field).

Results

Completeness. Most characters contribute to the distinction between
the seasonal forms and/or the sexes (see AN(C)OVA results in fig. 3). The
areas of pattern elements in particular, show significant differences with
respect to the seasonal classes in the ANCOVA’s (fig. 3, squares and
upward pointing triangles). Differences between the sexes are mainly

30(l-2):82-94, 1991

87

with respect to grey values, or, as in the case of contrast, due to grey
values (fig 3, dark symbols).

The PCA ranked the seasonal classes adequately in the first principal
component (PCI), the x-axis in fig. 4a. It shows that within each class, as
estimated by eye, variation still exists; some butterflies within a class are
of a drier/wetter form than others. The sexes are ranked in the second
component (PC2), the y-axis in fig. 4a. There is substantial variation
within each sexe; some butterflies resemble the opposite sex more than
others, especially in the drier forms. To conclude, the measurements of
the 67 characters with the image analyser allowed us to make a finer
division than ranking them in classes (seasonal or sexual) by eye.

Repeatability. Initially the means of the coefficients of variation
ranged up to 28.8%. The high errors were caused by difficulties in
separating the different pattern elements. Manual drawing of a line
around the pattern element with the mouse was very unreliable. Not
everybody was able to draw the lines precisely and different persons often
interpreted the elements differently.

Using coloration thresholds (e.g only pixels lower (=darker) than a
certain threshold value can belong to the black ring) proved more reliable
since a certain threshold always gives the same result for different
persons. When each person was allowed to choose their own threshold the
different interpretation of pattern elements still caused high errors
(COV around 10%). Setting a fixed threshold for each pattern element
resulted in acceptable errors (most COV’s <2%).

Another source of error was the different exposure to light of the wings
when recorded. This influences not only measurements of the intensity
of colors, but also all other measurements when a fixed grey value
threshold is used for separating the different pattern elements. The
internal systems of the image analyser regulating the conversion of
different light intensities into different grey values proved very sensitive
and vulnerable. So great care always had to be taken to calibrate the
system with a standard grey scale, and check for differences with
previous measurements.

The most reliable method for separating pattern elements was the use
of contrast. Different pattern elements differ in coloration, and conse-
quently differences between neighbouring pixels (= contrast) is highest
where two pattern elements border. The software program indicated
pixels where contrast was high, and these enclosed elements completely,
or left small gaps which could be closed manually. In this way differences
between persons are reduced (contrast is always the same) and the
influence of exposure is also reduced (contrast is similar at different
exposure levels), resulting in COV’s of less than 1%.

The white band was very difficult to measure in all forms, regardless
of separation procedure, since its outer edge fades into the background
and is never clear. Use of a fixed threshold was problematic because grey
values within the band varied over the wing. The inner edge however is

PC 2

88

J. Res . Lepid.

Fig. 4 First two principal components for butterflies in PCA of all characters (a) and

for seven selected characters (b, see table 1).

always clear, and it proved to be in a different position in the different
forms, according to how well the band was developed. The distance of the
white band to the crossing point of two veins proved easy and reliably
measurable.

Speed. In theory a software program can be written that finds and
measures all pattern elements itself. The only time needed in such a
program is for recording and storing the wings; automatic image analysis

30(l-2):82-94, 1991

89

PC 1

can be performed overnight. Such a software program, will, however be
very complicated and take a very long time to develop. Instead we used
a combined approach, where the greater part of the analysis is automati-
cally performed overnight, and only the functions that are too compli-
cated to program are performed manually.

When contrast was used for separating pattern elements, most time
was needed for closing the small gaps in the pixels enclosing the pattern
elements. For measuring 7 characters this took together with recording
and storing about 7 minutes per wing. When a fixed threshold value was
used for separating the pattern elements the only extra time needed was

90

J. Res. Lepid.

Table 1. The character subset selected for measurement of B. safitza wings,
indicating the amount of variation between seasonal form and sex, and
coefficients of variation. Numbers in brackets refer to fig. 2. — = not significant,
* = p < 0.05, ** = p < 0.01 , *** p < 0.001 .

Character AN(C)OVA CV

for season for sex

Area of hind wing (II)
Area of contrast (II)
Area of white pupil
in 5th cell (18)

Area of black ring
in 5th cell (17)
Distance from crossing
point of Cubital &
Median veins to
edge of band (23)
Mean grey value of
midfield in 5th
cell (28)

St. dev. of grey values
between eyespot
and chevron (32)

0.56%

1.17%

0.25%

0.26%

0.34%

1 .42%

0.82%

for indicating where the different pattern elements were located. To-
gether with recording and storing this took less then 1 minute per wing
for 7 characters. Because the gain in accuracy of measurement is only
small when using contrast instead of fixed threshold values, but the loss
in speed is substantial, the fixed threshold method is now always used by
us.

Selection of a subset of characters. Characters were selected with
reference to the three demands, completeness, repeatability and speed,
to reduce the time required to measure one butterfly. Only characters in
the 5th cell of the hind wing were selected. This resulted in a small subset
of seven characters (Table 1). A PCA comparable to that for all 67
characters was performed on these seven characters. It gave a closely
similar result to the PCA for all characters (fig. 4b). The correlation
coefficient for both PCA’s was 0.92 (p«0.001), showing that there is little
loss of information. Each of the major character sets (band, eyespots,
color) is represented.

OTHER USES OF THE IMAGE ANALYSER

Crypsis in Melanitis leda. A pilot study has attempted to quantify
the crypsis of insects at rest by 2-D image analysis of the extent of
background matching (see Endler 1984). Six individuals of the polyphenic

Image analysis - PC 1 value

30(l-2):82-94, 1991

91

Fig. 5 Crypsis ranking of Melanitis butterflies in photos by the image analyzer vs.
the human eye.

satyrine Melanitis leda , two wsf and four dsf, were each photographed on
a range of six backgrounds; three green leaves and three brown. A
comparison was made of the mean grey value, size and shape of the
butterfly wings with those of background objects. The same photos were
ranked according to conspicuousness of the butterfly by sixteen indepen-
dent observers. The data were analyzed by PC A. Fig. 5 shows the values
of the PC 1 for each photo plotted against their ranks. It demonstrates the
potential of image analysis for measuring the crypsis of butterfly wings.

Ephestia kuhniella. The measurement of grey values is used to
quantify variation in the amount of melanization of the wings within and
between different melanic and non melanic genotypes of the flour moth
E. kuhniella (fig 6). It shows that with the image analyser a finer
distinction between genotypes is possible than ranking them as melanic
and non melanic.

Development of eyespots in Bicyclus anynana. Experiments
involving cauterization of cells within developing eyespots in the pupae
of B. anynana are being performed to investigate the developmental
biology of the seasonal polyphenism. Earlier experiments of Nijhout
( 1981, 1985) indicated that an information gradient is established during
pattern determination in the early pupal stage. It is established around

92

J. Res. Lepid.

Fig. 6 Mean grey value and 95% confidence intervals for different genotypes of
Ephestia moths, as measured by the image analyser. Melanie alleles: An
= Alanigra, Ch = Charcoal, b = black (recessive).

a focus of cells within a putative eyespot. Image analysis in combination
with the making of Camera Lucida drawings provided the means of
accurately quantifying the differences between cauterized and control
eyespots at a magnification of about 250x.

Discussion

Biological problems can be approached in a different way (question 1 in
the introduction) when analysing lepidopteran wings with an image
analyser. The main difference for the analysis of Bicyclus wings, is that
it is now possible for us to make a better distinction between wing
patterns according to seasonal differences. Furthermore, we can process
far more wings than is possible with only a microscope. Another advan-

30(l-2):82-94, 1991

93

tage is the flexibility of the system as is shown by the other uses of it in
our laboratory. It has potential to quantify crypsis, it can distinguish
melanic genotypes in Ephestia , and it can work at a high magnification.

In general there are several advantages (the gain of question 2,
introduction) in using an image analyser for lepidoptera wing measure-
ments. It can perform two types of measurements which most other
electronic equipment cannot. It can measure not only distances (like a
digitizing pad), but also areas. This can be very useful if for example the
total black area in a complicated pattern (e.g. Pierid butterflies, Kingsolver
& Wiernasz, 1991) must be evaluated. Furthermore it can measure color
intensities, especially in lepidoptera wings often an important character
distinguishing between forms, species etc. However continuous varia-
tion in color intensity is, though certainly present, rarely studied. This
is probably due to the difficulty of measuring it, but with an image
analyser this need no longer be the case.

Other advantages are the speed of measurement and the gain in
precision of measurements. Both criteria were, however, not easily met
and took a lot of software development in our laboratory. If gain in
precision of linear measurements is the only goal, it is doubtful whether
the purchase of an image analyser will be the best investment. Precision
in these measurements can also be increased in other ways (e.g. repeat-
ing and averaging). The gain in speed can, however, be considerable. The
measurements themselves are faster and the data are directly stored and
accessible for computer programs. If analyses are constrained by the time
available, purchase of an image analyser might be worth the investment.

There are two disadvantages (the costs of question 2, introduction) of
the image analyzer. The first is its costs. Our system, including basic
software, cost nearly $20,000. The second disadvantage is its complexity.
There are so many functions that can be, and often must be, performed
before a measurement takes place, that programs become complex.
There is always the possibility of purchasing commercially written
programs, but they increase the price and seldom fit the demands of the
user completely. It took us a year to modify the basic software, that we
purchased commercially, into a program which fully satisfied our de-
mands.

To conclude, the image analyzer can be a valuable tool in analyzing
butterfly wings, and probably in other organisms too. It is flexible; both
simple and complex analyses are possible. Measurements can be made
which are otherwise impossible, and the gain in speed of measurements
can be considerable. It is, however, not a cheap system, and one must bear
in mind that software development can be a considerable effort.

Acknowledgements'. I thank Paul Brakefield and Jose Oudenaarden for their
helpful comments on earlier drafts of the manuscript, Paul Brakefield and Martin
Witzenburg for ranking the butterflies into seasonal classes, Anne Schot for the
data on crypsis in Melanitis , IJsbrand Swart for the data on Ephestia genotypes,
and all the members of the department that measured butterflies for the repeat-
ability experiment.

94

J. Res. Lepid.

Literature Cited

Brakefield, P.M., 1984. The ecological genetics of quantitative characters in
Maniola jurtina and other butterflies. Symp. R. Ent. Soc. 11, The biology of
butterflies pp. 167-190.

Brakefield, P.M. & W.H. Dowdeswell, 1985. Comparison of two independent
scoring techniques for spot variation in Maniola jurtina (L. ) and the consequences
of some differences. Biol. J. Linn. Soc. 24: 329-325.

Brakefield, P.M. & T.G. Liebert, 1985. Studies of colour polymorphism in some
marginal populations of the aposematic Jersey tiger moth Callimorpha
quadripunctaria. Biol. J. Linn. Soc. 26: 225-241.

Brakefield, P.M. & N. Reitsma, 1991. Phenotypic plasticity, seasonal climate and
the population biology of Bicyclus butterflies in Malawi. Ecol. Entom. 10

Bowers, M.D., I.L. Brown & D. Wheye, 1985. Bird predation as a selective agent in
a butterfly population. Evolution 39: 93-103.

Condamin, M., 1973. Monographic du genre Bicyclus (Lepidoptera Satyridae) -
Mem. Inst. Fond. Afr. Noire 88: 1-324.

Daly, H.V., 1985. Insect morphometries. Ann. Rev. Entomol. 30: 415-438.

Ehrlich, P.R. & L.G. Mason, 1966. The population biology of the butterfly Euphydryas
edit ha III. Selection and the phenetics of the Jasper ridge colony. Evolution 20:
165-173.

Endler, J., 1984. Progressive background matching in moths, and a quantitative
measure of crypsis. Biol. J. Linn. Soc. 23: 187-231.

Gonzales R.C. & P. Wintz, 1987. Digital image processing. 2nd. ed. 503 pp. Addison-
Wesley publishing company, Reading Massachucetts.

Hagerup, M.I., I. Sondergaard & J.K. Nielsen, 1990. Measurements of areas
consumed from leaf discs consumed by chewing phytophagous insects: description
of a new method involving image processing. Entomol. exp. appl. 57: 105-113.

Kingsolver, J.G., 1987. Evolution and coadaptation of thermoregulatory behavior
and wing pigmentation pattern pierid butterflies. Evolution 41: 472-490.

Kingsolver, J.G. & D.C. Wiernasz, 1991. Development, function and the quantitative
genetics of wing melanin patterns in Pieris butterflies. Evolution 45: 1480-
1492.

Mason, L.G., P.R. Ehrlich & T.C. Emmel, 1968. The population biology of the
butterfly Euphydryas editha VI. Phenetics of the Jasper ridge colony, 1965-66.
Evolution 22: 46-54.

Nijhout, H.F., 1981. The color pattern of butterflies and moths. Sci. Am. 245: 145-
151.

1985. The developmental physiology of color patterns in Lepidoptera. Adv.

Ins. Phys. 13: 181-247.

1991. The development and evolution of butterfly wing patterns . Smithsonian

Institution Press, Washington.

Robinson, R., 1990. Genetics of European butterflies, pp. 234-306. In: Butterflies
of Europe Vol. 2, Introduction to Lepidopterology, O. Kundra (ed.) Aula Verlag
Wiesbaden.

Shaw, R.G. , 1987. Maximum likelihood approaches applied to quantitative genetics
of natural populations. Evolution 41: 812-826.

Strauss, R.E., 1990. Patterns of quantitative variation in Lepidopteran wing
morphology: the convergent groups Heliconiinae and Ithomiinae (Papilionoidea:
Nymphalidae). Evolution 44: 86-103.

Journal of Research on the Lepidoptera

30(l-2):95-112, 1991

Quantification of Ant Attendance (Myrmecophily)
of Lycaenid Larvae

Gregory R. Ballmer

Department of Entomology, University of California, Riverside, California, 92521
and

Gordon F. Pratt

Department of Entomology and Applied Ecology, University of Delaware, Newark, Delaware,
19716

Abstract. A methodology is presented for quantifying ant attendance
of lycaenid larvae. Attendance of larvae by Formica pilicornis Emery
(Formicidae) is compared for 58 lycaenid species (primarily from
California). The presence of larval myrmecophilous organs is com-
pared with ant attendance rankings to estimate their relative contribu-
tions to attendance. Evidence is presented that dendritic setae may be
as important in ant recruitment as the better known honey gland and
eversible tubercle organs, and may be a primitive precursor of the
latter.

Introduction

The family Lycaenidae is known for its larval myrmecophily and
symbiotic ant-larval associations ranging from simple co-existence to
parasitism have been described (see Hinton, 1951; Malicky, 1969a, b;
Cottrell, 1984; Pierce, 1987; Maschwitz et al., 1988, 1989). Perhaps the
most common ant-larval interaction is mutualism, whereby larvae re-
ward ants with a mixture of carbohydrates and amino acids (honey dew)
in return for protection from predators. Malicky (1970) rejected mutu-
alism in favor of ant appeasement as the dominant aspect of myrmecophily
among lycaenids, largely because of lack of convincing evidence that ant
attendance reduced larval predation. However, recent workers have
provided evidence both of the nutritive value of larval secretions and the
protective value of attending ants (Maschwitz, et al, 1975; Pierce & Mead,
1981; Pierce & Easteal, 1986; Pierce, et al, 1987; Fiedler & Maschwitz,
1988a; DeVries, 1988a, 1991a).

The diversity of ant-lycaenid relationships, combined with related
morphological specializations, should provide valuable clues to lycaenid
phylogeny. Indeed, Henning (1983a) updated Hinton’s (1951) biological
groups within the Lycaenidae (based on myrmecophily, carnivory, etc.)
and noted their similarity to Eliot’s (1973) proposed phylogeny, which is
based primarily on adult morphology. Cottrell (1984) reviewed the
complex diversity of lycaenid feeding strategies (e.g. aphytophagy has
arisen independently at least eight times), and noted that attempts to

96

J. Res . Lepid.

explain these strategies must be linked to a full appreciation of phylog-
eny. However, the apparent loss of myrmecophilous organs and
myrmecophily in members of many lycaenid lineages and absence of
convincing evidence of intermediate stages in the evolution of the various
organs responsible for myrmecophily complicates the use of such fea-
tures for determining phylogenetic relationships. For entire groups
which lack certain organs, one can only speculate whether the group
predates the evolutionary origin of those organs or if the organs were lost
in the group’s progenitor. Further, the anecdotal nature of most descrip-
tions of myrmecophily allows only subjective qualitative comparisons
among taxa.

In this paper a standardized testing procedure is presented to quanti-
tatively compare a major aspect of myrmecophily, the degree of ant
attendance of last instar lycaenid larvae. The degree of attendance is
correlated with the presence of myrmecophilous organs (ant-organs):
lenticles, eversible tubercles, honey gland(s) [respectively, perforated
cupola organs, tentacular organs, and dorsal nectary organ and tentacu-
lar nectary organs of Cottrell (1984)], and dendritic setae (Ballmer &
Pratt, 1989). The contribution of dendritic setae to myrmecophily,
previously a matter of speculation (Fiedler & Maschwitz, 1988b; Ballmer
& Pratt, 1989), is demonstated by comparison of their densities with ant
attendance rates among various Lycaena species, which lack all other
known ant-organs except lenticles. Finally, the results of this study
provide a basis for speculations on the sequence of origin of the various
ant-organs.

Materials and Methods

The degree of ant attendance was measured as the mean number of seconds
that larvae were attended per five minutes (300 s). For each five-minute
observation period, one last instar larva was placed in a clear plastic arena (12
X 8.5 X 6 cm) containing five workers of Formica pilicornis Emery. Each
observation period began ten seconds after initial ant-larval contact. For most
species, observations were replicated at least ten times; the number of larvae
tested per species ranged from 1 to 10. The number of seconds of attendance per
five minutes was converted to percentage and then transformed to arc sine values
for analysis.

Ant attendance primarily consisted of active antennation of the larval cuticle,
as described by Malicky (1970), coupled with walking back and forth over the
larva. Alarm behavior of ants in response to eversion of eversible tubercles
(Malicky, 1969a; Claassens & Dickson, 1977; Fiedler & Maschwitz, 1988b) was
also counted as attendance, although it usually resulted in briefly suspended ( 1-
5 s) contact. Other behaviors that were not considered attendance include
incidental contact (contact < 1 s), ants at rest or preening atop larvae (no
antennation), and aggression. Instances of multiple simultaneous attendance
were counted the same as single ant attendance.

Ants obtained from a wild Formica pilicornis colony near the community of
Mountain Home (el. ca 1300 m), San Bernardino Co., CA, were used because they
commonly tend lycaenid larvae in nature. Ants were collected as needed and

30( l-2):95-l 12, 1991

97

seldom kept for more than four days, since their reliability for larval attendance
decreased over time in captivity. Ants that failed to attend larvae (which other
ants did attend) usually died within a day. Therefore, control larvae of Icaricia
acmon (Westwood & Hewitson) and 7. lupini (Boisduval), which are commonly
attended by F. pilicornis in nature, were used to gauge the attendance capacity
of subject ants: Ants which demonstrated a reduced attendance of 7. acmon or 7.
lupini (< 75% of normal attendance) (Table 1) were discarded. Ants were housed
in 1 L plastic food containers and fed 10% honey water dispensed by a cotton wick.
Lycaenid larvae were field-collected or reared from ova; 55 species were from
Arizona and California; one species each came from Brazil, Thailand and the
eastern United States.

A second experiment measured differences in ant attendance of Plebulina
emigdionis (Grinnell) for two colonies of F. pilicornis. The ants were from
Mountain Home and Victorville, CA, 57 km NW of Mountain Home. Ants from
the latter site were found at the base of the host plant, Atriplex canescens (Pursh)
Nutt, where the P. emigdionis larvae had been collected three months earlier. In
order to remove bias due to possible differences in larval attractiveness, the same
larvae were alternately exposed to ants from each site. This experiment was
inspired by initial test observations that larvae were poorly attended by F.
pilicornis (Mountain Home colony) in the laboratory even though they were
strongly associated with the same ant species in the field.

Four species of Lycaena ( editha , heteronea, rubida, and xanthoides) known to
be myrmecophilous were compared with respect to ant attendance and the
abundance of both dendritic setae and lenticles. This comparison also included
one population ofL. heteronea (Tioga Pass) which is apparently not myrmecophi-
lous in nature.

Data were analyzed using AN OVA with Duncan’s new multiple range test
( statistical package adapted for personal computer by the UCR Statistics Depart-
ment).

The nomenclature of higher taxonomic groups follows Eliot ( 1973), except that
Riodinidae is treated as a subfamily of Lycaenidae.

Results

ANT-ATTENDANCE

Table 1 lists all species tested in decreasing order of their mean
measured ant attendance. These are grouped in 13 sets whose members’
attendance means are not significantly different (P < 0.01). Generally,
these sets broadly overlap; however, there is very little overlap between
sets of means greater than 50% (sets A-G) and those less than 50% (sets
I-M).

Table 2 lists the tested species according to natural groups and
compares them with respect to mean ant attendance, the presence of ant-
organs, and observed myrmecophily in nature. All species have cuticular
lenticles; additional ant-organs are absent in two Riodininae, one
Miletinae, 7 Lycaeninae, and two Theclini. A honey gland is present in
all 22 Polyommatinae and all 19 Theclinae. Tentacle nectary organs are
present in one riodinine. Eversible tubercles occur in one thecline (F.
fulgida ) and all polyommatines except P. speciosa. Dendritic setae occur
in one riodinine, 4 lycaenines, 15 theclines, and 20 polyommatines.

98

J, Res . Lepid.

Table 1 . Larval attendance by Formica pilicornis

Species

Lycaena editha (Mead)

Phaeostrymon alcestis (W. H. Edwards)
Satyrium auretorum (Boisduval)

Satyrium behrii (W. H. Edwards)

Satyrium californicum (W. H. Edwards)
Satyrium saepium (Boisduval)

Satyrium sylvinum (Boisduval)

Icaricia lupini (Boisduval)

Flos fulgida (Hewitson)

Everes comyntas (Godart)

Harkenclenus titus (Fabricius)

Philotes sonorensis (C. & R. Felder)
Glaucopsyche lygdamus (Doubleday)
Hemiargus isola (Reakirt)

Eurybia sp.

Celastrina argiolus (L.)

Glaucopsyche pi as us (Boisduval)
Lycaena rubida (Behr)

Brephidium exile (Boisduval)

Icaricia icarioides (Boisduval)

Icaricia acmon (Westwood & Hewitson)
Hemiargus ceraunus (Fabricius)

Everes amyntula (Godart)

Strymon melinus (Hubner)

Lycaena xanthoides (Boisduval)
Euphilotes pallescens elvirae (Mattoni)
Satyrium tetra (W. H. Edwards)

Lycaeides melissa (W. H. Edwards)
Incisalia mossii (Hy. Edwards)

Icaricia shasta (W. H. Edwards)

Incisalia augustinus (W. Kirby)

Icaricia neurona (Skinner)

Lycaeides id as (L.)

Mitoura spinetorum (Hewitson)

Fixsenia Ontario (W. H. Edwards)
Satyrium fuliginosum (W. H. Edwards)
Lycaena heteronea (Boisduval) (W)2
Lycaena mariposa (Reakirt)

Lycaena arota (Boisduval)

Mitoura loki (Skinner)

Callophrys perplexa (Barnes & Benjamin)
Plebulina emigdionis (Grinnell)

N

Mean1

Duncan’s Test

12

100

A

10

100

A

10

100

A

10

100

A

10

100

A

10

100

A

10

100

A

11

100

A

13

99

A

10

99

A

10

99

A

10

99

A

10

99

A

10

98

A

10

98

A

10

97

A

10

96

A

10

96

AB

10

92

ABC

10

91

ABCD

11

89

ABODE

10

87

ABODE

10

86

ABODE

10

86

ABODE

21

85

ABODE

10

82

ABCDEF

10

82

ABCDEFG

20

76

BCDEFG

10

72

BCDEFG

21

70

BCDEFG

12

69

CDEFG

11

68

DEFG

10

68

DEFG

10

66

DEFG

1-1

65

EFG

10

64

FGH

10

63

GH

10

41

HI

10

28

IJ

10

26

UK

7

24

IJK

20

16

IJKL

30(l-2):95-112, 1991

99

Calephelis wrighti Holland
Euphilotes mojave (Watson & Comstock)
Hahrodais grunus (Boisduval)

Lycaena heteronea (Boisduval) (T)3
Hypaurotis crysalus (W. H. Edwards)
Mitoura siva (W. H. Edwards)

Leptotes marina (Reakirt)

Lycaena phlaeas (L.)

Lycaena hermes (W. H. Edwards)
Lycaena gorgon (Boisduval)

Feniseca tarquinius (Fabricius)

Philotiella speciosa (Hy. Edwards)
Apodemia mormo (C & R. Felder)
Lycaena nivalis (Boisduval)

Euphilotes bernardino martini (Mattoni)
Lycaena cuprea (W. H. Edwards)

Atlides halesus (Cramer)

11

18

IJKL

10

18

IJKLM

10

13

IJKLM

10

12

IJKLM

10

11

IJKLM

10

11

JKLM

10

9

JKLM

10

7

JKLM

13

7

JKLM

10

5

KLM

10

5

KLM

10

4

LM

7

4

LM

5

3

LM

7

2

LM

10

0

M

12

0

M

1 Rounded mean percent time of ant attendance per 300 s; means followed by same letter
are not significantly different (P < 0.01) according to Duncan’s new multiple range test.

2 W = unnamed subspecies from Warren Canyon, Mono Co., Ca.

3 T = unnamed subspecies from Tioga Pass, Mono Co., Ca.

Table 2. Ant attendance and myrmecophilous organs of some lycaenid larvae

Subfamily1 Species AA2 HG3 TNO ET L DS M

Riodininae

Riodinini

A. Mormo 4

C. wrighti 18

Eurybia sp. 98

Miletinae

Miletini

F. tarquinius 5

Lycaeninae

Lycaenini

L. cuprea 0

L. phlaeas 7

L. arota 28

L. hermes 7

L. mariposa 41

L. nivalis 3

L. gorgon 5

?

?

?

?

L. heteronea (W)4
L. heteronea (T)5
L. edit ha
L. rubida
L. xanthoides

Theclinae

Theclini

H. grunus

H. crysalus

Eumaeini

S. melinus
C. perplexa

I. augustinus
I. mossii

M. spinetorum
M. loki
M. siva
A. hales us
P. alcestis
H. titus
F. Ontario
S. auretorum
S. behrii
S. californicum
S. fuliginosum
S. saepium
S. syivinum
S. tetra

63

12

100

96

85

13

11

86

24

69

72

66

26

11

0

100

99

65

100

100

100

64

100

100

82

Arhopalini

F. fulgida

99

Polyommatinae

Polyommatini

B. exile 92

L. marina 9

C. argiolus 97

P. sonorensis 99

P. speciosa 4

E. bernardino martini 2

E. enoptes mojave 1 8

E. pallescens elvirae 82

G. lygdamus 99

G. piasus 96

30(l-2):95-112, 1991

101

H. ceraunus

87

+

—

+

+

+

+

H. isola

98

+

—

+

+

—

?

1. acmon

89

+

—

+

+

+

+

1. icarioides

91

+

—

+

+

+

?

1. lupini

100

+

—

+

+

+

+

1. neuroma

68

+

—

+

+

+

?

1. shasta

70

+

—

+

+

+

+

P. emigdionis

16

+

—

+

+

+

+

L. idas

68

+

—

+

+

+

?

L. melissa

76

+

—

+

+

+

+

E. amyntula

86

+

—

+

+

+

+

E. comyntas

99

+

—

+

+

+

+

1 Taxa grouped according to subfamily, tribe, and lower category relationships.

2 Rounded mean percent ant attendance per 300 s by Formica pilicornis..

3 HG = honey gland, TNO = tentacle nectary organs, ET = eversible tubercles, L = lenticles,
DS = dendritic setae, M = myrmecophily in nature; + = present, <197> = absent,

? = undetermined.

4 W = unnamed subspecies from Warren Canyon, Mono Co., Ca.

5 T = unnamed subspecies from Tioga Pass, Mono Co., Ca.

Although greater mean ant attendance roughly coincides with pres-
ence of more ant-organs, there are notable exceptions. Of the 27 most
strongly attended species (Table 1, set A), 11 have four types of ant-
organs, but the remainder have only two or three. And while 9 of the 16
least attended species (Table 1, set M) lack any ant-organs other than
lenticles, 3 others have four types of ant-organs. Similar exceptions occur
within subfamilial and tribal lineages.

Myrmecophily has not been systematically investigated among the
lycaenid species addressed in this study. Table 4 tabulates the previ-
ously unreported ant-lycaenid relationships observed in the course of a
survey of the California lycaenid fauna (Ballmer & Pratt, 1989). Table
1 indicates the known status of myrmecophily for all test species com-
piled from field observations (Ballmer & Pratt, personal observations; D.
M. Wright, personal communication). Nearly all of the species known to
be myrmecophilous in nature are in the Polyommatinae; four lycaenines,
five theclines, and one riodinine are also known to be myrmecophilous.
This tabulation is biased by the nature of the larval collection techniques.
The larval hosts of most of the lycaenines and polyommatines are low
shrubs and herbs on which larvae were most easily located by searching;
in these species, ant associations were easily observed and aided in
locating larvae. Because many of the theclines were collected by beating
foliage or reared from ova in the lab, observations of natural ant
associations were precluded for most species. The high degree of
experimental ant attendance observed here suggests that some Eumaeini
may also be attended by ants in nature.

102

J. Res. Lepid.

ABUNDANCE OF DENDRITIC SETAE AND LENTICLES AMONG
LYCAENA SPECIES

Table 3 compares the abundance of dendritic setae and lenticles on
larvae of the four Lycaena species which were most ant-attended. All four
species have also been found attended by ants in the field (Ballmer &
Pratt, 1989, and personal observations). However, the population of L.
heteronea from Tioga Pass had the fewest dendritic setae and lenticles
and was also least ant-attended; coincidentally, when these larvae were
found in the field, no ants were in attendance. Because of variability in
abundance of lenticles and dendritic setae among individual larvae of
these populations, and because other larvae were used (in some cases) to
generate ant attendance values (Tables 1, 2), the former character data
cannot validly be correlated with the latter. Nevertheless, the data
suggest that abundance of both lenticles and dendritic setae may be
positively related to ant attendance. Of these, the abundance of dendritic
setae seems to be more strongly associated with attendance.

COLONY-RELATED DIFFERENCES IN ANT ATTENDANCE

Comparison of attendance of P. emigdionis by F. pilicornis from two
locations reveals significant differences. Larvae were tended very little
by ants from Mountain Home (Table 1), but significantly more (P = 5 X
10'15) by ants from Victorville (mean percent attendance = 75, N = 18).

NON-TENDING ANT BEHAVIOR

Non-tending ant responses to larvae during tests included indifference
and aggression. The majority of ants not attending larvae were indiffer-
ent to them. When these ants encountered larvae, they generally
investigated them briefly with minimal antennation and then moved on.
The infrequent instances of aggression were usually initiated immedi-
ately upon contact by a single ant and lasted for 2-10 s. Invariably, the
point of attack was just ventral to the lateral fold of the larva. Larvae of
A. mormo andL. arota were attacked for periods of 90 s, and one larva of

Table 3. Dendritic setae and lenticles on myrmecophilous Lycaena larvae

species

N

Dendritic Setae1

Lenticles1

L. rubidus

10

171 ± 41

A

993 ± 89 A

L. edit ha

4

140 ± 20

A

909 ± 67 A

L. xanthoides

4

84 ± 8

B

761 ± 83 B

L. heteronea (W)2

7

73 ± 44

B

679 ±123 B

L. heternoea (T)3

4

11 ± 8

C

646 ± 70 B

1 Mean totals ± standard deviation; means followed by the same letter are not significantly
different (P< 0.05) according to Duncan’s new multiple range test.

2 Population from Warren Canyon, Mono Co., California.

3 Population from Tioga Pass, Mono Co., California.

30(l-2):95-112, 1991

103

Table 4. Ants associated with lycaenid larvae in California

Attending ant species1 Lycaenid species

Aphaenogaster occidentalis Emery
Camponotus essigi M. Smith
Camponotus vicinus Mayr
Conomyrma bicolor (Wheeler)
Conomyrma sp.

Crematogaster californica Emery
Crematogaster mormon um Emery
Forelius pruinosus (Roger)

Formica altipetens Wheeler
Formica lasioides Emery
Formica moki Wheeler
Formica neoclara Emery
Formica neogagates Emery
Formica obscuripes Forel
Formica pilicornis Emery

Formica subsericea Say
Formica sp. (fusca group)

Formica sp. (microgyna group)
Formica sp. (rufa group)
Iridomyrmex humilis (Mayr)

Lasius niger (L.)

Lasius pallitarsus (Provancher)
Monomorium sp.

Myrmecocystus mimicus Wheeler
Myrmecocystus semirufus Emery
Tapinoma sessile (Say)

Icaricia acmon

Euphilotes pallescens elvirae
Satyrium fuliginosum
Everes amyntula
Glaucopsyche piasus
Euphilotes enoptes smithi
Philotes sonorensis
Euphilotes pallescens elvirae
Hemiargus ceraunus
Lycaena editha
Icaricia acmon

Euphilotes battoides comstocki
Glaucopsyche lygdamus
Lycaeides melissa
Everes amyntula
Icaricia acmon, Icaricia lupini,
Glaucopsyche piasus,

Lycaena heteronea,

Lycaena xanthoides,

Plebulina emigdionis
Euphilotes battoides battoides
Euphilotes battoides battoides,

E. b. comstocki,

Everes amyntula, Icaricia acmon
Satyrium fuliginosum
Satyrium fuliginosum
Euphilotes bernardino bernardino,
Icaricia acmon, Leptotes marina,
Strymon melinus
Everes amyntula
Euphilotes battoides battoides
Euphilotes battoides comstocki
Euphilotes pallescens elvirae
Glaucopsyche lygdamus
Euphilotes enotpes and I la,

E. e. smithi,

Glaucopsyche lygdamus

1 Records based on the authors’ field observations.

A. holesus (Cramer) was bitten continuously for 60 min, resulting in
perforation of the larval cuticle with loss of hemolymph.

When bitten, larvae usually remained motionless, sometimes after
curling into a ‘C’ shape. However, larvae ofA. mormo thrashed about and
regurgitated a fluid which, upon contact, caused ants to immediately
withdraw and preen themselves. This behavior may be similar to the
“beat reflex” observed in Hamearis lucina (L.) by Malicky ( 1969a, 1970).

104

J. Res. Lepid.

Another defense against ants occurs in the Neotropical riodinine Sarota
gyas (Cramer), whose larvae are densely covered with tufts of long,
slender, easily-broken setae; ants which contact the setae spend much
time cleaning their antennae and avoid further contact (DeVries, 1988b).
Larvae of C. wrighti, which are similarly covered with tufts of long,
slender, fragile setae (Ballmer & Pratt, 1989), did not induce such a
repellant response inF. pilicornis', as with other poorly attended species,
ants generally ignored C. wrighti after initial investigation.

Discussion

EXPERIMENTAL VS. NATURAL ANT ATTENDANCE

The ant attendance rankings reported here are not considered equiva-
lent to the relative degree of myrmecophily for the larvae in nature.
Many of the lycaenids tested have been found naturally in association
with two or more ant species (Table 4), as well as without any attending
ants (Ballmer & Pratt, personal observations). Even though myrmecophily
is facultative in most California lycaenids (Ballmer & Pratt, personal
observations) and most of the North American fauna (Pierce, 1987), this
condition does not preclude differences in relative mymecophily for
different ant-lycaenid species combinations. Such a situation does exist
with L. marina , which was well attended by 7. humilis in the field but was
poorly attended by F. pilicornis in the lab. Additionally, various environ-
mental factors could affect ant attendance under more natural conditions
(e.g. access to the nest and nest-mates, other nutrition sources, stress
levels, etc).

Intraspecific geographic differences can occur for ant-larval interac-
tion; this dictates caution in applying the test results to other popula-
tions. For example, two populations ofL. heteronea differed significantly
in attractiveness to the same ants, and, vice versa, larval attendance
differed significantly for two populations of the ant, F. pilicornis with
respect to P. emigdionis.

SOURCES OF EXPERIMENTAL ERROR

The design of this experiment could have contributed to erroneous
estimates of ant attendance values. Because single and multiple simul-
taneous attendance were counted equally, real differences among the
more highly attended species could not be distinguished; this problem
might be overcome by counting all attending ants separately. Also,
because many ants initially palpate and investigate any new object in
their environment, the relatively short exposure times may have re-
sulted in somewhat inflated attendance values for some species. How-
ever, the 10 s delay between initial ant contact and the onset of recording
attendance at least partially compensated for this behavior. It is also
possible that greater attendance values might have been obtained with
longer exposure times through nest-mate recruitment, as demonstrated
for the ant, Tetramorium caespitum (L.), attending larvae of Polyommatus
coridon (Poda) (Fiedler & Maschwitz, 1989).

30(l-2):95-112, 1991

105

In spite of the limitations described, this test procedure provides a
simple, repeatable means of measuring ant attendance of larvae under
uniform conditions.

Lenticles

Lenticles are low-relief cuticular features for which a putative chemical
communication function (Hinton, 1951; Malicky, 1969a, 1970; Henning
1983b) may require direct contact with ant antennal receptors. Indeed,
antennation of larval cuticle is an important feature in ant attendance in
this study and others (Malicky, 1969a, 1970; Fiedler & Maschwitz, 1989),
and is often most intense in regions where lenticles are concentrated. In
these areas ant-lenticle contact may be facilitated by the presence of
generally sparser and/or shorter setae than occur elsewhere (but see
dendritic setae, below).

In this study there is no evidence that lenticles contributed to ant
attendance. All species with greater than 50% attendance (Table 1, sets
A-G), possess additional ant-organs, whereas 12 of the 22 species having
less than 50% attendance (Table 1, sets I-M) do not. Nevertheless,
significant differences in attendance observed among the 12 species
having only lenticle ant-organs may be due to quantitative and/or
qualitative differences in semiochemicals associated with the lenticles.

Among populations of myrmecophilous Lycaena species, the ranked
order of ant attendance (Table 1 ) corresponds to the ranked order of mean
abundance of both lenticles and dendritic setae (Table 3). Of these, the
abundance of dendritic setae is a better indicator of attendance. While
the essentially non-myrmecophilous L. heteronea population from Tioga
Pass had significantly less attendance and significantly fewer dendritic
setae than the other populations, it did not have significantly fewer
lenticles than L. heteronea from Warren Canyon and L. x anthoides.

In spite of accumulated evidence of ant adoption substances in or on the
surface of lycaenid larvae (Malicky, 1969b, 1970; Henning, 1983b), the
contribution of lenticles to myrmecophily remains obscure. As noted by
Cottrell ( 1984), a number of different compounds may be involved and at
least some may arise from cuticular sources other than lenticles. Such
sources may include the cuticle itself, specialized setae, or the putative
dermal glands (indicated by some surface pores and deep pits) found on
lycaenid larvae (Wright, 1983; Kitching & Luke, 1985; DeVries et al,
1986; Kitching, 1987). Also, DeVries ( 1991b) noted that the evolutionary
origin of lenticles may be unrelated to myrmecophily, since they occur in
non-myrmecophilous hesperiids, as well as in both myrmecophilous and
non-myrmecophilous lycaenids.

Honey gland

Aside from lenticles, the honey gland is the most prevalent ant-organ
among the species tested. The function of this organ in ant recruitment
may be largely, if not solely, related to the secretion of liquid nutrition.

106

J. Res. Lepid.

Many ants which specialize in collecting the sugary fluids produced by
homopterans and plant nectaries also attend lycaenid larvae (DeVries,
1991a, b). The honey gland is therefore a broad-spectrum ant-attracter,
as opposed to other ant-organs which may produce more specific
semiochemical attractants. The attractiveness of the honey gland was
quantified by Fiedler & Maschwitz (1989) for Polyommatus coridon
(Poda) by comparing ant recruitment for larvae with functional and
blocked (resin-covered) honey glands. It maybe more difficult to quantify
the nutritional component of attractiveness of this organ among species
tested here due to possible additive attractiveness of concentrations of
lenticles and dendritic setae occurring along its margins. Although not
quantified here, the time spent in feeding from the honey gland was
usually relatively small compared to total attendance time.

The honey gland occurs widely in the Polyommatinae and Theclinae,
but is apparently absent in other lycaenid subfamilies (Cottrell, 1984).
Kitching and Luke (1985) refer to the condition in which a honey gland
is absent as myrmecoxeny, and note that the condition is imperfectly
correlated with lack of ant attendance. The strong ant attendance of
some Lycaena species reported here illustrates that myrmecophily and
myrmecoxeny are not mutually exclusive. Furthermore, although ants
were observed to imbibe fluid from the honey gland of polyommatine
larvae, such secretions were not observed in any thecline larva in the
tribe Eumaeini. The latter are apparently functional myrmecoxenes.

Tentacle nectary organs

Tentacle nectary organs occur only in Riodininae and combine the
general appearance, size, and location (posterior to A- 8 spiracles) of
eversible tubercles with the secretory function of the dorsal honey gland
of Polyommatinae and Theclinae. Although Cottrell (1984) noted that
tentacle nectary organs differ from eversible tubercles in producing
visible secretions, and in lacking terminal setae, he presumed these
organs to be homologous. DeVries (1991b), however, considered the ant-
organs of riodinines to be analogous, not homologous, to the ant-organs
of other lycaenid groups. Nevertheless, there are convergent exceptions
to the usual physical distinctions between tentacle nectary organs and
eversible tubercles. Thus, the eversible tubercles of some non-riodinine
lycaenids, e.g. the Australian Candalides xanthospilos (Hubner)
(Polyommatinae) and Ogyris genoueua Hewitson (Theclinae), lack termi-
nal setae (Ballmer & Pratt, 1989), while the tentacle nectary organs of at
least one myrmecophilous riodinine ( Setabis sp.) apparently do not
produce a liquid secretion (Ballmer, personal observations).

Eversible tubercles

The most notable effect of the eversible tubercles was the apparent
induction of heightened activity and aggressive posturing in attending
ants. When eversible tubercles (segment A-8) were briefly everted (< 1

30(l-2):95-lI2, 1991

107

s), attending ants usually dashed about with open mandibles, as though
alerted to danger. This phenomenon was typical in those species having
eversible tubercles. Such heightened ant activity (alarm behavior)
following eversion of the tubercles could drive away predators and is
probably commonplace in the Lycaenidae (Claassens & Dickson, 1977;
Henning, 1983b; Cottrell, 1984; DeVries, 1984; Kitching & Luke, 1985;
Fiedler & Maschwitz, 1988b), although Malicky (1969a, 1970) failed to
observe such behavior. Even non-excited ants walking on larvae may
inhibit attack by predators; K. Calloway {in litt ) has found that non-
attendant movement of ants on the hostplant near larvae of Brephidium
exile (Boisduval) can effectively 'scare’ off predators. Larvae of P .
emigdionis were exceptional in that they everted their tubercles more
frequently (especially while crawling) and for longer duration (often
several seconds) than did other species, yet alarm behavior by attending
ants was infrequent and generally less intense than with other species.

The mechanism by which the eversible tubercles induce heightened
activity in ants is generally believed to be through release of chemicals
which mimic ant alarm pheromone(s). Perhaps the most convincing
evidence for this is provided by Henning (1983b), who showed that an
extract of larval cuticle containing eversible tubercles of Aloeides dentatis
(Swierstra) elicited initial alarm reaction and subsequent attraction in
Acantholepis capensis Mayr. The anterior tentacle organs of some
riodinines appear to have a similar function (DeVries, 1988a).

Eversible tubercles are somewhat more wide spread than the dorsal
honey gland in the Lycaenidae. In addition to frequent occurrence among
theclines and polyommatines, they also occur in the Curetinae and
Liphyrinae (Cottrell, 1984).

Dendritic setae

The dendritic setae occurring on larvae and/or pupae of many members
of the Riodininae, Lycaeninae, Polyommatinae, and Theclinae (Ross,
1964; Lawrence & Downey, 1966; Schremmer, 1978; Kitching, 1983;
Kitching & Luke, 1985; Fiedler, 1988; Pratt, 1988; Ballmer & Pratt, 1989,
and unpublished observations) may release semiochemicals which affect
ant behavior (Fiedler & Maschwitz, 1988b). Although varying in density
and distribution among species, dendritic setae tend to be concentrated
in many of the same areas as lenticles, especially around the honey gland
and spiracles. They are usually longer, more erect, and have longer
lateral processes than surrounding setae (see Ballmer & Pratt, 1989);
their greater flexibility and reduced pigmentation also suggest a thinner
setal wall which could facilitate dissemination of volatile compounds.
The generally greater prominence of dendritic setae probably also facili-
tates their contact with the antennae of attending ants.

Among the species tested here, the presence of dendritic setae appears
to be a better indicator of ant attendance than is the presence of other ant-
organs. They were present in all theclines (except F. fulgida ) having

108

J. Res. Lepid.

greater than 50% attendance and absent in all theclines (except C.
perplexa) having less than 50% attendance. The relationship of dendritic
setae to ant attendance among polyommatines is less clear; the two
species lacking dendritic setae, B. exilis and H. isola, were strongly
attended.

The relationship of dendritic setae to ant attendance is most apparent
in the Lycaeninae, a group that lacks both honey gland and eversible
tubercles. The four species of Lycaena found to have dendritic setae were
also the most strongly attended (Table 2). Among these four species,
greater abundance of dendritic setae (Table 3) coincides with greater ant
attendance (Table 1). Two populations ofL. heteronea, representing two
subspecies (Emmel & Pratt, in press) from the eastern central Sierra
Nevada illustrate infraspecific variations in both abundance of dendritic
setae and ant attendance.

If dendritic setae are homologous with the spiculate setae of eversible
tubercles and anterior tentacle organs, then the latter organs may be
viewed as specializations of the former to accumulate (while retracted)
and disseminate (when everted) a higher concentration of semiochemicals.
Since ant pheromones which act as attractants at low concentrations
may incite alarm behavior at higher concentrations (Blum, 1974; Henning,
1983b), a constant release of semiochemicals by exposed dendritic setae
might yield concentrations sufficient to attract ants, while higher con-
centrations sufficient to alarm them might be disseminated when tu-
bercles are infrequently and momentarily everted.

The frequent and prolonged eversion of the tubercles observed for P.
emigdionis might result in a relatively constant low level of release of
semiochemicals similar to that hypothesized for exposed dendritic setae.
This might explain why little alarm behavior was elicited by the eversion
of tubercles by larvae of this species. The frequent and prolonged
eversion of the tubercles could be important in maintaining a retinue of
attending ants while larvae periodically move between subterranean
diurnal resting sites and above-ground nocturnal foraging sites (Ballmer
& Pratt, personal observations). In an apparently similar behavior,
larvae of Aloeides thyra (L.) and A. dentatis (Swierstra) are reported to
repeatedly and rapidly evert their tubercles when leaving their diurnal
ant nest shelters to feed nocturnally on vegetation (Claassens & Dickson,
1977; Henning 1983a, b).

COLONY-SPECIFIC ATTENDANCE RATES

An example of colony-specific differences in ant attendance is apparent
with P. emigdionis. Perhaps the more attendant group of ants had
‘learned’ to associate withP. emigdionis through long-term contact in the
field, or may also represent a cryptic sibling species or ecotype adapted
to P. emigdionis. The patchy distribution of this lycaenid (Emmel &
Emmel, 1973) and its occurrence in small dense colonies, often within
larger stands of host plant (Ballmer & Pratt, personal observations) may

30(l-2):95-112, 1991

109

be due to a special relationship between the butterfly and its attendant
ants. This is reinforced by observations that field-collected larvae were
nearly always found with F. pilicornis in diurnal shelters at the plant
base (below soil surface) or while feeding crepuscularly on above-ground
foliage.

NONSPECIFIC MYKMECOPHILY

In contrast to P. emigdionis, which was highly attended only by a
particular colony (or variety) of F. pilicornis from the larval collection
site, some lycaenids were highly attended in spite of their origin outside
the ant’s range of distribution. Flosfulgida from Southeast Asia (Chiang
Mai, Thailand) was highly attractive to F. pilicornis. Twenty larvae of
F. fulgida were found in nature individually accompanied by numerous
(often more than twenty) ants ( Hypoclinea sp.) in folded leaf shelters.
Although that situation may signal a species-specific relationship, the
presence of a more generalized myrmecophilic factor in F. fulgida is
indicated by its attractiveness to F. pilicornis.

THE ORIGINAL ANT ORGAN?

Knowledge of the sequence of origin of the various ant-organs could be
helpful in understanding the evolutionary history of the Lycaenidae.
Although it has been speculated that a nutritional secretory organ, such
as the dorsal honey gland, was a feature of the primitive lycaenid
ancestor (Hinton, 1951; Malicky, 1970; Pierce, 1987), cuticular lenticles,
which occur much more widely (Hinton, 1951; Malicky, 1969a; Ballmer
& Pratt, 1989), were probably derived earlier. Lenticles are the only ant-
organs that occur in all lycaenid subfamilies for which larvae have been
examined. Dendritic setae and their possible derivatives, eversible
tubercles and anterior tentacle organs, also may be more widespread and
potentially of earlier origin than nectary organs.

Whereas nutritive secretions from a honey gland or tentacle nectary
organs may attract and bind a broad spectrum of fluid-foraging ant
species, semiochemicals released by lenticles, dendritic setae, eversible
tubercles, and anterior tentacle organs might achieve a similar attrac-
tion or bonding with less energy expenditure. Of these organs, dendritic
setae may offer the best combination of structural simplicity, energy
efficiency and effectiveness in disseminating ant-attracting
semiochemicals. Such semiochemicals might also facilitate more specific
and more efficient symbiotic relationships than nectary gland secretions.

The similarity of dendritic setae to spiculate setae typically found on
eversible tubercles and anterior tentacle organs, in both structure and
apparent function, suggests that the latter organs may be specialized
derivatives of the former. Such a derivation would strengthen Fiedler’s
(1988) hypothesis that the Lycaeninae is an ancient group predating the
origins of the closely related Polyommatinae and Theclinae. Although
Malicky (1970) concluded that the lack of a honey gland and eversible

110

J. Res. Lepid.

tubercles in the Lycaeninae are derived features, he was apparently
unaware of strongly myrmecophilic members of this group and of the
relationship of dendritic setae to myrmecophily.

DeVries (1988, 1991b) concluded that riodinine ant-organs are analo-
gous but not homologous to the ant-organs of other lycaenids, and that
myrmecophily among riodinines was derived independently. DeVries
( 1991b) also questioned the importance of lenticles in myrmecophily, and
rejected the notion that myrmecophily is ancestral in Lycaenidae. How-
ever, he was unaware that dendritic setae, which occur widely among
myrmecophilous lycaenids (including riodinines), may represent not
only the most primitive ant-organ (if lenticles are excluded) but a
common precursor to eversible tubercles and anterior tentacle organs, as
well. Nevertheless, because the presence of dendritic setae is apparently
as labile as that of other ant-organs, and because the structure of
dendritic setae may be a simple derivation of ‘standard’ body setae, the
possibility of independent derivation in different lineages cannot be
ignored.

Resolution of questions regarding lycaenid phylogeny may ultimately
depend on an integrated analysis of many factors involving both immatures
and adults. Because of broad diversity and lability of feeding habits,
myrmecophily, and larval morphology within higher taxonomic groups,
evolutionary patterns based on such features should be corroborated by
comparison with ostensibly more stable characters, such as first instar
chaetotaxy. Also valuable may be studies of the sites of origin and
structures of semiochemicals related to myrmecophily and comparisons
of DNA among representatives of the various taxonomic groups.

Acknowledgments. We thank Roy Snelling and Phillip Ward for ant identifications.
Charles Huszar of the UCR Statistics Department provided the AN OVA program.
David M. Wright provided larvae of Feniseca tarquinius. We also thank Karen
Calloway, Phillip DeVries, Ian Kitching, and David M. Wright for their suggestions
in improving the manuscript.

Literature Cited

Ballmer, G. R. and G. F. Pratt, 1989. A survey of the last instar larvae of the
Lycaenidae (Lepidoptera) of California. J. of Res. Lepid. 27: 1-80.

Blum, M.S., 1974. Pheromonal sociality in Hymenoptera. In M. C. Birch (ed)
Pheromones. Amsterdam (North-Holland Publishing Co.), pp. 222-249.
Claassens, A. J. M. and C. G. C. Dickson, 1977. A study of the myrmecophilous
behaviour of the immature stages of Aloeides thyra (L.) (Lep: Lycaenidae) with
special reference to the function of the retractile tubercles and with additional
notes on the general biology of the species. Entomologist’s Record and Journal
of Variation 89: 225-231.

Cottrell, C.B., 1984. Aphytophagy in butterflies: its relationship to myrmecophily.

Zoological Journal of the Linnean Society 80: 1-57.

DeVries, P. J., 1984. Of crazy ants and the Curetinae: are Curetis butterflies tended
by ants? Zoological Journal of the Linnaean Society 20: 621-633.

30(l-2):95-112, 1991

111

, 1988a. The larval ant-organs of Thisbe irenea (Lepidoptera: Riodinidae) and

their effects upon attending ants. Zoological Journal of the Linnean Society 94:
379-393.

, 1988b. The use of epiphylls as larval hostplants by the neotropical riodinid

butterfly, Sarota gyas. Journal of Natural History 22: 1447-1450.

, 1991a. Mutualism between Thisbe irenea butterflies and ants, and the role

of ant ecology in the evolution of ant-larval associations. Biological Journal of
the Linnaean Society (in press).

, 1991b. Evolutionary and ecological patterns in myrmecophilous riodinid

butterflies. In C. Huxley & D. F. Cutler (eds.) Ant-Plant Interactions. Oxford
University Press, (in press).

DeVries, P. J., D. J. Harvey, & I. J. Hitching, 1986. The ant associated epidermal
organs on the larva of the lycaenid butterfly Curetis regula Evans. Journal of
Natural History 20: 621-633.

Eliot, J. N., 1973. The higher classification of the Lycaenidae (Lepidoptera): a
tentative arrangement. Bulletin of the British Museum (Natural History),
Entomology 28: 371-505.

Emmel, J. F. and G. F. Pratt, in press. New subspecies of Lycaeninae in California.
Natural History Museum of Los Angeles County Contributions in Science.

Emmel, T. C. and J. F. Emmel, 1973. The butterflies of southern California. Natural
History Museum of Los Angeles County Science Series 26: 1-148.

Fiedler, K., 1988. The preimaginal epidermal organs of Lycaena tityrus (Poda,
1761) and Polyommatus coridon (Poda, 1761) (Lepidoptera: Lycaenidae) - a
comparison. Nota Lepidopterologica 11: 100-116.

Fiedler, K. and U. Maschwitz, 1988a. Functional analysis of the myrmecophilous
relationships between ants (Hymenoptera: Formicidae) and lycaenids
(Lepidoptera: Lycaenidae). II. Lycaenid larvae as trophobiotic partners of ants
- a quantitative approach. Oecologia 75: 204-206.

, 1988b. Functional analysis of the myrmecophilous relationships between

ants (Hymenoptera: Formicidae) and lycaenids (Lepidoptera: Lycaenidae). III.
New aspects of the function of the retractile tentacular organs of lycaenid
larvae. Zoologisches Beitrage N. F. 31: 409-416.

, 1989. Functional analysis of the myrmecophilous relationships between ants

(Hymenoptera: F ormicidae) and lycaenids (Lepidoptera: Lycaenidae). I. Release
of food recruitment in ants by lycaenid larvae and pupae. Ethology 80: 71-80.

Henning, S. F., 1983a. Biological groups within the Lycaenidae (Lepidoptera).
Journal of the Entomological Society of Southern Africa 46: 65-85.

, 1983b. Chemical communication between lycaenid larvae (Lepidoptera:

Lycaenidae) and ants (Hymenoptera: Formicidae). Journal of the Entomological
Society of Southern Africa 46: 341-366.

Hinton, H. E., 1951. Myrmecophilous Lycaenidae and other Lepidoptera - a
summary. Transactions of the South London Entomological and Natural
History Society, 1949-50: 111-175.

Hitching, R. L., 1983. Myrmecophilous organs of the larvae and pupa of the
lycaenid butterfly Jalmenus evagoras (Donovan). Journal of Natural History
17: 471-481

, 1987. Aspects of the natural history of the lycaenid butterfly Allotinus major

in Sulawesi. Journal of Natural History 21: 535-544.

Hitching, R. L. and B. Luke, 1985. The myrmecophilous organs of the larvae of some

112

J. Res. Lepid.

British Lycaenidae (Lepidoptera); a comparative study. Journal of Natural
History 19: 259-276.

Lawrence, D. A. and J. C. Downey, 1966. Morphology of the immature stages of
Everes comyntas Godart (Lycaenidae). J. Res. Lepid 5: 61-96.

Malicky, H., 1969a. Versuch einer Analyse der Okologischen Beziehungen
zwischen Lycaeniden (Lepidoptera) und F ormiciden (Hymenoptera). Tijdschrift
voor Entomologie 112: 213-98.

, 1969b. Uebersicht ueber Praimaginal stadien, Bionomie und Okologie der

mitteleuropaischen Lycaenidae (Lepidoptera). Mitteilungen der
Entomologischen Gesellschaft, Basel N.F. 19: 25-91.

, 1970. New aspects on the association between lycaenid larvae (Lycaenidae)

and ants (Formicidae, Hymenoptera). J. Lep. Soc 24: 191-202.

Maschwitz, U., W. A. NAssig, K. Dumpert, and K. Fiedler, 1988. Larval carnivory
and myrmecoxeny, and imaginal myrmecophily in miletine lycaenids
(Lepidoptera, Lycaenidae) on the Malay Peninsula. Tyo to Ga 39: 167-181.

Maschwitz, U., M. Wust, and K. Schurian, 1975. Blaulingsraupen als
Zuckerlieferanten fur Ameisen. Oecologia 18: 17-21.

Pierce, N. E., 1987. The evolution and biogeography of associations between
lycaenid butterflies and ants. Oxford Surveys in Evolutionary Biology 4: 89-
116.

Pierce, N. E. and P. S. Mead, 1981. Parasitoids as selective agents in the symbiosis
between lycaenid butterfly caterpillars and ants. Science 211: 1185-1187

Pierce, N. E. and S. Easteal, 1986. The selective advantage of attendant ants for
the larvae of a lycaenid butterfly, Glaucopsyche lygdamus. Journal of Animal
Ecology 55: 451-462.

Pierce, N. E., R. L. Kitching, R. C. Buckley, M. F. J. Taylor, and K. F. Benbow, 1987.
The costs and benefits of cooperation between the Australian lycaenid butterfly,
Jalmenus evagoras, and its attendant ants. Behavioral Ecology and Sociobiology
21: 237-248.

Pratt, G. F., 1988. The evolution and biology of Euphilotes biotypes. Ph. D.
dissertation, University of California, Riverside. 653 pp.

Ross, G. M., 1964. Life history studies on Mexican butterflies. II. Early stages of
Anatole rossi a new myrmecophilous metalmark. J. Res. Lepid 3: 81-94.

Schremmer, F., 1978. Zur Bionomie und Morphologie der myrmekophilen Raupe
und Puppe der neotropischen Tagfalter-Art Ha mearis erostratus (Lepidoptera:
Riodinidae. Entomologica Germanica 4: 113-121.

Wright, D. M., 1983. Life history and morphology of the immature stages of the bog
copper butterfly Lycaena epixanthe (Bsd. & Le C.). J. Res. Lepid 22: 47-100.

Journal of Research on the Lepidoptera

30(1-2):113-120, 1991

Appearance of the “heathii” Aberration and
Genitalic Variation in a Mitoura Population from
Oregon (Lycaenidae: Theclinae)1

Clifford D. Ferris2

Bioengineering Program, University of Wyoming, Laramie, WY 82071-3295

Abstract. A juniper-association Mitoura species (ascribable to barryi
barryi Johnson) experienced a population explosion in eastern Oregon
in June, 1990. Substantial variation in the female genitalia was
observed. Two of the female specimens collected manifested the
ventral surface “heathii” aberration occasionally found in genus
Satyrium Scudder.

Introduction

The “heathii” aberration in the Theclinae, which is a broadening of the
white maculation on the ventral surfaces of the wings, was described in
1904 by Fletcher. In contemporary times, additional discussion has been
presented by Fisher (1976), and Ferris (1981, 1982). This aberration is
normally associated with the genus Satyrium Scudder, and its occur-
rence is considered to be rare in other theclines. In England, this
aberration has been observed in two species related to North American
Satyrium. These species are Strymondia w-album 9 ab. “albovirgata”
Tutt and Quercusia quercus 9 ab. “latefasciata” Courvoisier (see
Russwurm, 1978). Johnson et at. (1990) have also alluded to the
occasional manifestation of “white postmedial suffusions” in the Strymon
eremica (Hayward) group found in South America; they did not, however,
invoke the epithet “heathii.” The common aberration (if it really is such)
in Mitoura is a reduction in ventral maculation as described for M. n.
nelsoni (Boisduval) and given the varietal [form] name “exoleta” by Hy.
Edwards ( 1881). Based upon current and limited evidence, the “heathii”
variant may be the expression of a homologous allele that occurs widely
in the tribe Theclini. This aberration appears to occur more frequently
in females than in males, but it is not sex restricted.

In June, 1990, I collected approximately 375 specimens of a Mitoura
species in Crook, Grant, Harney, and Wheeler Cos., Oregon. The
population which I sampled is ascribable to Callophrys (Mitoura) barryi
barryi Johnson, based upon the discussion in Johnson (1976), and this
assignment will be discussed subsequently. This species experienced a
population explosion, and could be found virtually everywhere that

Published with the approval of the Director, Wyoming Agricultural Experiment Station as
Journal Article No. JA 1635.

2Research Associate: Allyn Museum of Entomology/Florida Museum of Natural History,
Sarasota, FL; Florida State Collection of Arthropods, DPI, FDACS, Gainesville, FL.
Research Associate in Entomology, Natural History Museum of Los Angeles County, Los
Angeles, CA.

114

J. Res. Lepid.

Figs. 1-4. Males of Mitoura b. barryi from Grant Co., Oregon, Malheur Nat. For., N.
of Mt. Vernon, 24 vi 90. Dorsal surfaces (1 -2); ventral (3). Tip of aedeagus
(4).

Figs. 5-6. Dorsal (5) and ventral (6) surfaces of typical females of Mitoura b. barryi
from Harney Co., Oregon, Malheur Nat. For. along For. Rd. 2820, 28.vi.90.
Figs. 7-8. Ventral surfaces of two aberrant females of Mitoura b. barryi from Oregon.
(7) Harney Co., Malheur Nat. For. along For. Rd. 2820, 28 vi 90. (8) Grant
Co., Malheur Nat. For., N. of Mt. Vernon, 24 vi 90.

30(1-2):113-120, 1991

115

Juniperus o. occidentalis Hook, occurs within the altitudinal range from
1250-1680 m. Adults were observed perching on juniper shrubs and
sipping nectar on a variety of flowers, although there was a definite
preference for yellow composites. The number of specimens collected
provided an ample study series. Other than the ventral form “exoleta”,
no unusual male specimens were taken. As is discussed below, there are
two normal dorsal-surface phenotypes in the adult males. There is little
variation in the dorsal wing surfaces of the females, and the ventral
“exoleta” form occurs regularly, as in the males. Two females were found
which exhibit the “heathii” aberration.

John Lane ( pers . comm.) has indicated that occasional specimens from
the San Bernardino Mts. population of juniper-feeding Mitoura exhibit
a tendency toward the “heathii” phenotype. These specimens, however,
are uncommon.

The genitalic discussion which follows below should not be construed
as having any bearing upon the “heathii” aberration. It is a general
discussion relative to the Oregon population sampled.

Photographic Records

Figs. 1-3 illustrate typical males from the Oregon population sampled.
In the form shown in Fig. 1, the dorsal surface is pale grayish-brown and
the FW “thecla” patch shows prominently. There is also a prominent
rusty marginal band extending from the anal angle of the HW (arrows in
photograph). The form shown in Fig. 2 manifests an almost uniform dark
ruddy-brown color, and the “thecla” patch is obscure. Fig. 3 illustrates
the typical ventral pattern found in the males.

Figs. 5-6 show, respectively, the dorsal and ventral maculation found
in normal females. The dorsal color is a warm ruddy-brown. Figs. 7-8
illustrate the ventral surfaces of the two aberrant specimens which
manifest the “heathii” aberration. Their dorsal surfaces (not illustrated)
are normal. The specimen shown in Fig. 7 is a partial expression of the
aberration, while it is fully developed in the specimen shown in Fig. 8.
Note the FW broad pale areas, and the extensive widening of the
normally thin post-discal band such that it merges with the submarginal
band. In round numbers, these two aberrant specimens constitute
approximately 1% of the females collected.

Host Plant Association

By way of summary, Johnson (1976) segregated the western North
America populations of Mitoura whose larvae feed upon the Cupressaceae
as follows: nelsoni muiri (Hy. Edwards) on Cupressus sargentii Jeps;;
nelsoni nelsoni on Libocedrus decurrens Torr.; loki (Skinner), siva ssp.
on J uniperus sp. ; rosneri Johnson on Thuja plicata Wats. Twoofhistaxa
were referred to “suspected” hosts; byrnei to Thuja plicata , and harry i to
T. plicata or Juniperus occidentalis. M. barryi barryi was described from
specimens collected in Grant, Union, and Wallowa Cos., Oregon, and

116

J. Res. Lepid.

Moscow Mt., Latah Co., Idaho. Regarding the host plant of barryi,
Johnson stated: “The larval foodplant is unknown.”

If one ignores Johnson’s genitalic descriptions (see discussion below),
and accepts his descriptions of the maculation of the imagines, then M.
b. barryi is the species that I collected and it is associated with J. o.
occidentalis. This butterfly is more widely distributed in Oregon than
Johnson indicated. Its range extends south into Harney Co., and west
into the Ochoco Mts. of Wheeler and Crook Cos. In the mountains within
this geographic area, there are dry habitats that apparently lie in the
rain shadow of surrounding peaks. These sites support stands of
Juniper, and this is where barryi occurs.

I can not verify Johnson’s records for barryi on Moscow Mt., Latah Co.,
Idaho. I have collected in June what is referable to byrnei on T. plicata
on Moscow Mt., where this butterfly is relatively common. I have
collected a similar population, also in June, on T. plicata in Idaho Co.,
Idaho. Again the butterflies were widely distributed and common. The
Thuja -association Mitoura occur in a much more mesic environment
than do the Juniper us -feeding Mitoura.

Based upon my collections, and extensive discussions with John Lane
who has been studying this group for some years, it seems to make more
sense to classify the Cupressaceae-feeding Mitoura on a biological basis
as opposed to questionable genitalic characters. These butterflies parti-
tion very nicely with regard to host plant. Upon further study, it is very
likely that chemical races will be identified, that is local populations of
Mitoura on hosts with chemical components (essential oils, etc.) specific
to geographic locality. How one handles the “classical” taxonomy of such
entities is another matter, while the biological basis of classification
(butterfly to host) is relatively simple. General discussions of butterfly
and plant coevolution may be found in Spencer (1988).

Genitalic Variation

In his 1976 discussion of Mitoura , Johnson primarily based his taxo-
nomic assignments upon genitalic characters. In the interim since this
paper appeared, various specialists have questioned the validity of these
characters for making species assignments. The dissections that I have
made of males of this Oregon population vary from the findings for barryi
reported by Johnson, but agree with anecdotal reports and information
provided to me by C. S. Guppy and J. Lane. Typical structures are
illustrated in Figs. 4, 9-10. Fig. 4 is a drawing of the tip of the aedeagus.
Because of the very small size of the structures involved, I was not able
to evert the vesica. The two “cornuti” that Johnson described are actually
two surfaces with spines or processes along their perimeters. The spines
are shaped much like rose thorns. In the preparations that I made, the
ventrad perimeter typically supported six major spines, and the dorsad
perimeter five. Lesser spines of variable number were noted basad of the
major processes.

30(1-2):113-120, 1991

117

Figs. 9-10. Male genitalia of Mitoura b. barryi from Grant Co., Oregon, Malheur Nat. For., N. of Mt. Vernon, 24 vi 90.

Figs. 11-12. Female genitalia with signafrom two specimens of Mitoura b. barryi from Oregon. (1 1 ) Harney Co., Malheur Nat. For. along For.

Rd. 2820, 28 vi 90. (12) Grant Co., Malheur Nat. For., N. of Mt. Vernon, 24 vi 90.

Figs. 1 3-1 4. Female genitalia without signa from two specimens of Mitoura b. barryi from Oregon. Harney Co., Malheur Nat. For. along For.
Rd. 2820, 28 vi 90.

118

J. Res. Lepid.

t

e

0.1mm

Fig. 15. Five signs forms found in female genitalia of Mitoura b. barryi from E.

Oregon.

The photomicrograph shown in Fig. 9 is a male preparation minus the
aedeagus. The uncus has been split to provide a dorsal-view mount. The
saccus is funnel shaped, as was described by Johnson for barryi , but the
juxta is narrow-rimmed and parabolic as Johnson described for byrnei.
Note that Johnson defined these structures as the “saccus valvae.” The
photographs shown represent only two of the dissections that I made. I
did not find the form of the “saccus valvae” that he described for barryi.
The valvae shown in Fig. 10 have been flattened to illustrate their full
surfaces. Generally their form agrees with that described by Johnson for
barryi.

Among females, I found substantial variation from that described by
Johnson for barryi regarding the presence of signa in the corpus bursae
of the genitalia. Figs. 11-14 are photomicrographs of the corpus bursae
of four different specimens. The genital plate with extraneous tissue
removed is also shown in Fig. 12. Of 18 preparations, only 12 (67%)
contained signa (arrows in Figs. 11-12). Figs. 13-14 illustrate bursae
lacking signa. The forms of the signa found are shown in Fig. 15. The
signa are very small, as the 0. 1 mm scale indicates. The triangular form
shown in Fig. 15 (a) corresponds to Fig. 11. The actual structures in the
bursae of Fig. 11 are very much smaller than the darkened areas that
appear in the photograph. The preparations have been stained with
Chlorazole Black E. Figs. 15(c) and 12 correspond. Based upon my study
of the female genitalia in the Oregon population of barryi sampled, there
are at least five forms of the signa when they are present; however, the
corpus bursae is frequently devoid of signa.

The signa in the female genitalia of TYmya-association Mitoura from
Idaho and Latah Cos., Idaho were also examined. In both size and form,
they were found to be similar to those of barryi from the Oregon localities
cited above.

30( 1-2): 113-120, 1991

119

Concluding Discussion

There are several schools of thought regarding the taxonomy of genus
Mitoura [or Callophrys (Mitoura) ]. At one extreme, is Scott’s (1986)
lumping of all but two of the western Mitoura ( sensu stricto ) into the
single species gryneus (Hiibner). At the other extreme, is Johnson’s
separation (1976) of the genus into a variety of species based upon
phenotype, host plant association, and presumed uniformity of genitalic
characters within given populations. Various studies are currently in
progress regarding the genus Mitoura in western North America. In
addition to my own field studies, John Lane and C. S. Guppy (pers.
comm. ), among others, have been rearing material sampled from various
geographic localities. Johnson’s studies are apparently continuing as
well (; in litt. ). Guppy has reared two species from the same locality on
Vancouver Island; the larvae of one feed on Thuja sp., while those of the
other feed on Juniperus sp. According to Guppy, these two Mitoura
species are respectively ascribable (tentatively) to rosneri Johnson and
harryi. Before publishing his results, however, he is searching for solid
morphological characters that will provide positive species separation.
The occurrence of juniper-association and red-cedar-association Mitoura
in close proximity is apparently not uncommon in British Columbia. I
have series of each taken a decade ago in the upper Okanagan Valley, and
additional records exist.

Based upon contemporary studies, and the data that are emerging from
them, it is evident that much additional work remains to elucidate the
relationships among the various western Mitoura. This is especially true
of the populations found along the eastern perimeter of the Great Basin,
and in the Intermountain Region. It is clear that there is considerably
more variability in the genitalia than was initially reported. Because of
this variability within given colonies, I suggest that genitalic characters
are of little use in making species assignments within the Cupressaceae-
feeding group of Mitoura. Based upon current data, biological classifica-
tion of this group, as opposed to genitalic classification, appears to be the
sensible approach.

In addition to reporting the occurrence of the “heathii” aberration in a
Mitoura species, the object of this paper is to document the wing patterns
and genitalic characters of Mitoura b. barryi indigenous to four Oregon
counties. It is further suggested that species partitioning of Mitoura
should be based upon biological and not genitalic characters.

Acknowledgments. . I wish to thank Drs. R. J. Lavigne and R. E. Pfadt of the Dept,
of Entomology, Univ. of Wyoming for critically reading a preliminary draft of this
paper. My thanks also to John Lane (Santa Cruz, CA) and to C. S. Guppy
(Vancouver, BC) for sharing their information and opinions with me. I wish also
to acknowledge suggestions from Kurt Johnson, who was an external referee for
this paper. The genitalic studies reported were conducted using an Olympus
SZ3060 zoom stereo microscope fitted with a 100AL 2X auxiliary lens to provide
magnification to 250X. Dimensions were scaled using Olympus H 10/10 and H10/
100 eyepiece micrometers

120

J. Res. Lepid.

Literature Cited

Dornfeld, E. J. 1980. The Butterflies of Oregon. Timber Press, Forest Grove, OR.

Edwards, H. 1881. On some apparently new forms of diurnal Lepidoptera. Papilio
1:50-55.

Ferris, C. D. 1981(1982). Field notes on four western hairstreaks (Lycaenidae:
Theclinae). J. Lepid. Soc. 35(4):325-330.

1982(83). Polymorphism in Satyrium calanus (Huebner) from Wyoming and

Colorado (Lepidoptera: Lycaenidae: Theclinae). J. Res. Lepid. 21(3):188-193.

Fisher, M. S. 1976. The heathii- white banding aberration in the Strymoninae
(Lycaenidae). J. Res. Lepid. 15(3):177-181.

Fletcher, J. 1904. Descriptions of some new species and varieties of Canadian
butterflies. Can. Ent. 36(5):121-130.

Johnson, K. 1976. Three new Nearctic species of Callophrys ( Mitoura ), with a
diagnosis of all Nearctic consubgeners (Lepidoptera: Lycaenidae). Bull. Allyn
Mus. (38):l-30.

Johnson, K., Eisele, R. C., and B. MacPherson 1990. The “hairstreak butterflies”
(Lycaenidae, Theclinae) of northwestern Argentina II. Strymon, sensu stricto.
Bull. Allyn Mus. (130): 1-77.

Russwurm, A. D. A. 1978. Aberrations of British Butterflies. E. W. Classey, Ltd.,
Oxon, England.

Scott, J. A. 1986. The Butterflies of North America. Stanford University Press,
Stanford, CA.

Spencer, K. C., ed. 1988. Chemical Mediation of Coevolution. Academic Press, Inc.,
San Diego, CA.

Journal of Research on the Lepidoptera

30(1-2):121-128, 1991

An Artificial Diet for Butterflies, Including Bicyclus
Species, and its Effect on Development Period,
Weight and Wing Pattern

G.J. Holloway, P.M. Brakefield, S. Kofman and J. J. Windig.

Section of Evolutionary Biology, Department of Population Biology, University of Leiden,
Schelpenkade 14a, 2313 ZT Leiden, The Netherlands.

Abstract. A semi-artificial diet based on bean flour suitable for the
laboratory rearing of butterflies is described. The effect of the diet on
development period, pupal weight and wing colour pattern was exam-
ined. Butterflies of the species Bicyclus anynana reared on living maize
plants developed faster and were heavier, but there was no significant
effect of diet on colour pattern.

Introduction

Bergomaz and Boppre (1986) described a simple semi-artificial diet
that could be used to rear moths in the laboratory or, in a pre-prepared
form, in the field. Although some species refused the diet, 38 species were
reared successfully on the diet from egg to adult with no apparent effects
on size, colour or other characteristics. Hence the diet provided them
with a cost effective means of rearing moths in the field. A further
advantage to using a pre-prepared diet is that rearing conditions can be
standardised. These circumstances are attractive where a balanced data
set is preferred in order to make statistical tests more robust. Also, the
use of diet may reduce environmental variation which could interfere
with certain types of study, such as in quantitative genetics.

Butterflies are frequently used as experimental material, usually with
the use of living plants for larval food. This is expensive both in terms of
space and time. We experimented with the use of the diet described by
Bergomaz and Boppre (1986) using particularly the tropical African,
brown butterfly, Bicyclus anynana (Butler), but also B. safitza and B.
ena. A number of pilot experiments showed that the larvae could develop
on the diet with over 90% pupating, but the majority of the emerging
butterflies were crippled, i.e. the wings failed to expand properly, if at all.

The purpose of the present study was to modify the diet to make it more
suitable for use with B. anynana by reducing the number of crippled
butterflies developing. B. anynana has a striking plastic response to
certain environmental cues which results in the production of dry season
or wet season butterflies (Brakefield & Reitsma, 1991). These seasonal
forms differ in their appearance, development time, physiology and
behaviour; the most conspicuous difference is the much larger, sub-
marginal eyespots present in wet season butterflies. Therefore, an
examination of the diet reared adults was performed to assess the
influence of the artificial conditions on size and wing pattern.

122

J. Res . Lepid.

Materials and Methods

The recipe of the diet used in the current study is shown in Table 1. The diet
differed from that devised by Bergomaz and Boppre (1986) in that p-formalde-
hyde and formaldehyde were omitted and linseed oil was added. Linseed oil
provided extra essential fatty acids, the lack of which in the original diet
presumably caused the observed crippling. The agar was dissolved in hot water.
The other ingredients (thoroughly mixed together) were then added to the agar
and stirred well to disperse the material evenly throughout. The gelling mixture
was then poured into a suitable tray to a depth of about 1.5cm and allowed to set
at room temperature for about one hour. The diet was then covered and
transferred to the fridge for storage.

Two alternative quantities of sorbic acid, methyl-p-hydroxybenzoate and
streptomycin were used (Table 1). Two types of diet were prepared: one
containing the higher amounts of the preservatives and antibiotics (normal diet)
and a second containing the lower quantities (low diet). This was done since
Bergomaz and Boppre (1986) suggested that these materials may have an
adverse effect on the development rate.

The experiment was carried out at 28°C and 80% relative humidity (12 hours
D:L) which normally produces only extreme wet season form butterflies (Brakefield
& Reitsma, 1991). For each type of diet, 100 newly hatched, unfed larvae from
eggs of a large mixed population stock laid on maize plants were placed
individually onto cubes of diet (approximately 1.5cm wide and 1.0cm high) in
9.5cm diameter petri dishes. 50 of the larvae were presented with fresh diet two
times per week and the other 50 only once a week. The old diet was always
removed from the petri dishes and discarded. Clean petri dishes were only
provided if fungus developed on the diet or, more often, on the frass. In addition,
to examine density effects, for each diet type 48 larvae (two per petri dish) and
120 larvae (five per petri dish) were also set up. Cannibalism is not a problem in
Bicyclus species. As before, half of the dishes had the diet renewed two times a
week and the other half only once a week.

Larvae pupated on the underside of the lids of the petri dishes. The resulting
pupae could easily be removed after a day or two by scraping the silken holdfast
from the petri dish lid. The day of pupation and the weight (to the nearest 0. Img)
of the pupae two days after pupation were recorded. Pupae were transferred after
weighing to small cardboard containers (height 4cm, diameter 5cm), covered with
a piece of gauze and returned to the 28°C cabinet. Butterflies were able to develop
and emerge normally from pupae lying down rather than hanging as is invariably
the case. Emergent butterflies were sexed and the widths of the 2nd forewing
eyespot and the 5th hindwing eyespot were measured under a Wild microscope
at 25x magnification using a graduated eyepiece. The length of the forewing was
also measured using the eyepiece at 6x magnification. Butterflies successfully
developing on the diet were placed together in a large breeding cage at 28°C and
90% relative humidity to assess whether they were capable of laying fertile eggs.

50 larvae were raised in the same 28°C cabinet on maize to enable comparisons
to be made with butterflies reared on a more natural diet. The larvae were reared
in groups of 10. The growing plants were coiled up into 12.5cm diameter plastic
cages. Availability of space limited us to rearing larvae in groups, which is also
the technique used generally in our experimental work. The same morphometric
measurements were made on the maize-reared butterflies.

30(1-2): 121-128, 1991

123

Table 1. Diet recipe (modified from Bergomaz and Boppre, 1986)

Above added to

Ingredients

Amounts

White bean flour

75g

Brewers yeast

17.5g

Ascorbic acid

3.5g

Cholesterol

0.5g

Sorbic acid

0.5g or

0.2g

Methyhp-hydroxybenzoate

0.5g or

0.2g

Streptomycin

0.4g or

0.1g

10% tocopherol in germ oil

3ml

linseed oil

4ml

solution:

High strength agar

15g

Deionised water

400ml

Results

The larvae on the low diet performed very poorly with no butterflies
emerging from diet changed once per week. This was probably due to the
development of mould on the diet, which may not have been toxic to the
larvae per se but rather formed a barrier between the larva and the food.
From the diet that was changed two times per week about 50% developed
from the petri dishes containing single and two larvae. Survival dropped
to 15% in petri dishes containing 5 larvae. The purpose of the study was
to develop an acceptable diet and to determine a suitable way of
presenting the diet to the larvae, therefore only data from the normal diet
will be considered from here on.

Eighty percent of the solitary larvae on the normal diet changed two
times per week developed successfully and 74% of the larvae on normal
diet changed once per week. The percentage of larvae developing
normally (i.e. into adults with fully expanded wings) also remained high
with two larvae per dish when the diet was changed two times per week
(79%), but dropped to 58% when the diet was changed only once per week.
The percentage of larvae successfully developing from five larvae per
dish was 55% when the diet was changed two times per week, but only
42% when the diet was changed once per week. Therefore, overall
survival was slightly improved by keeping the diet fresh (i.e. by renewing
the diet two times per week) and, under these conditions, survival was
not adversely affected by increasing the density from one to two larvae
per petri dish. In total, 77 males and 91 females developed, which is a sex
ratio not significantly different from unity (%2 = 1.17).

Table 2 shows the means for the development period (egg hatching to
pupation) for the various treatments and by sex. The anova in Table 2
shows a significant density effect with individuals held singly in petri
dishes developing slightly faster than higher densities. The frequency
with which the diet was changed did not have a significant effect on

124

J. Res. Lepici.

Table 2. Means and analysis of variance of development periods (emergence
from egg to emergence of adult in days) of butterflies reared from artificial diet in
petri dishes at a variety of densities and where the diet was renewed either one or

two times per week.

Diet renewal

Sex

Density

rate per week

1

2

5

1

6

28.3

30.6

31.2

9

33.0

34.3

34.2

o

6

28.6

32.3

28.5

9

31.3

31.3

34.8

Analysis of variance of development periods

Source

DF

MS

F

P

Density (d)

2

57.90

3.79

0.025

Renewal (r)

1

29.97

1.90

0.170

sex (s)

1

337.78

22.11

<0.001

d * r

2

0.17

0.01

0.989

d * s

2

37.69

2.47

0.088

r * s

1

7.09

0.47

0.497

Error

153

15.28

Total

162

development period. Males develop on average faster, are lighter and
have a different pattern than females, thus it is not surprising to find
significant sex effects throughout.

Table 3 shows the mean fresh pupal weights at two days old. There

were significant effects of density and the frequency with which the diet
was changed; the heaviest animals coming from the petri dishes contain-
ing one larva in which the diet was renewed two times per week.

Tables 4 and 5 compare the wing patterns by considering the ratios of
the width of the 2nd forewing spot (2spot) (25x magnification) over the
length of the forewing (6x magnification) and the ratio of 2spot over the
width of the 5th hindwing spot (5spot), respectively. Absolute spot size
was not considered as wing size and spot size are allometrically related,
but the pattern may be the same irrespective of the size of the butterfly.
There is a clear sex effect for both measures, as expected. Density had
no effect on pattern, but there was a marginally statistically significant
effect of frequency of diet renewal on the size of the 2spot/5spot ratio.
However, with a significant interaction between frequency of diet re-
newal and sex there was not a single method of diet presentation
producing the largest ratio.

Overall, neither density nor frequency of diet renewal had a clear effect
on wing pattern. However, if the quality of the diet is determined by fast
development and large butterflies, from Tables 2 and 3 it appears that

30(1-2):121-128, 1991

125

Table 3. Means and analysis of variance of pupal weights (grams) at two days
old from larvae reared from artificial diet in petri dishes at a variety of densities
and where the diet was renewed either one or two times per week.

Diet renewal

Sex

Density

rate per week

1

2

5

1

6

0.117

0.113

0.109

9

0.145

0.118

0.135

S

0.125

0.113

0.119

2

9

0.153

0.150

0.137

Analysis of variance of pupal weights

Source

DF

MS

F

P

Density (d)

2

0.0020

6.84

0.001

Renewal (r)

1

0.0036

12.43

0.001

sex (s)

1

0.0206

71.73

<0.001

d * r

2

0.0004

1.23

0.294

d * s

2

0.0002

0.79

0.458

r * s

1

0.0001

0.47

0.497

Error

153

0.0003

Total

162

one larva per petri dish (at least on the size of diet blocks used in the
current study) and diet renewed two times per week is best.

The results from the best diet treatment were compared with the
performance of a sample of larvae reared on maize at 28°C. 52% of the
larvae on maize reached the 4th instar and of the 26 larvae assigned to
28°C, 20 produced butterflies. Thus about 40% emergence from egg to
adult survival (however, if sufficient space was available to rear larvae
individually on living maize plants, much higher survival could certainly
be achieved). For none of the comparisons made between the diet results
and the results from maize was there a significant interaction between
food (maize or diet) and sex. Therefore, single F values are presented.
The mean development periods (emergence from egg to adult) for males
(20.4 days) and females (22.2 days) on maize were significantly shorter
than on diet (F=86.4, p<0.001). The mean pupal weight was also higher
from maize (males: 0.142g, females: 0.173g) than from diet (F=8.67,
p<0.05), although, curiously, there was not a significant difference
between the lengths of the forewings from the two foods (F=1.81, p=n.s.).
The ratios of 2spot/wing were 0.372 (males) and 0.370 (females) from
maize and were not significantly different from the diet butterflies
(F=0.21, p=n.s.). The 2spot/5spot ratios for maize butterflies were 0.76
(males) and 0.71 (females) and again there was no difference between the
two groups of butterflies (F<0.01, p=n.s.).

The butterflies developing on diet were placed together to crudely

126

J. Res. Lepid.

Table 4. Means and analysis of variance of ratio of the width of the 2nd forewing
spot (25x magnification) over length forewing (6x magnification) in butterflies
reared from artificial diet in petri dishes at a variety of densities and where the diet
was renewed either one or two times per week.

Diet renewal

Sex

Density

rate per week

1

2

5

1

7

0.38

0.37

0.38

/

0.34

0.32

0.33

7

0.37

0.39

0.35

2

/

0.37

0.35

0.37

Analysis of variance of ratio of 2nd spot over wing length

Source

DF

MS

F

P

Density (d)

2

0.0015

0.55

0.577

Renewal (r)

1

0.0092

3.30

0.071

sex (s)

1

0.0292

10.44

0.002

d * r

2

0.0007

0.24

0.787

d * s

2

0.0031

1.11

0.331

r * s

1

0.0209

7.46

0.007

Error

153

0.0028

Total

162

assess the effect of the diet on breeding potential. Although eggs were
laid, there were very few considering the number of females which
suggested that either only a small number of butterflies laid eggs or that
fecundity was generally much reduced. Most of the eggs hatched
normally.

Discussion

A large number of artificial diets suitable for the rearing of many
species of moths can be found in the literature (e.g. Vanderzant, 1967;
1974; Singh and Moore, 1985; Bergomaz and Boppre, 1986). However, it
is generally not as easy to rear butterflies on artificial diets (Wielgus,
1974; Morton, 1979). There could be several reasons for this. For
example, many moth species are generalists with respect to host plant
whilst butterflies tend to be more specialized. Also, some moths are
capable of boring into materials; a behaviour that butterflies, on the
whole, do not possess. However, we do know of a few instances where
butterflies are being reared on artificial diets. Among these workers it is
generally believed that butterflies may be unable to find certain essential
fatty acids in diets designed for moths. Indeed, this seems to have been
the cause of the extensive crippling of wings noted during all of our earlier
trials. The addition of linseed oil to the present diet (Table 1) removed

30(1-2):121-128, 1991

127

Table 5. Means and analysis of variance of ratio of the width of the 2nd forewing
spot over 5th hindwing spot in butterflies reared from artificial diet in petri dishes
at a variety of densities and where the diet was renewed either one or two times

per week.

Diet renewal Sex Density

rate per week 1 2 5

1

M

0.73

0.76

0.76

F

0.65

0.66

0.65

M

0.76

0.80

0.74

2

F

0.71

0.72

0.70

Analysis of variance of 2nd spot forewing over 5th spot hindwing

Source

DF

MS

F

P

Density (d)

2

0.0073

0.86

0.426

Renewal (r)

1

0.0406

4.76

0.031

sex (s)

1

0.2086

24.44

<0.001

d * r

2

0.0052

0.61

0.542

d * s

2

0.0018

0.21

0.812

r * s

1

0.0169

1.98

0.161

Error

153

0.0085

Total

162

this problem and we successfully reared B. anynana, B. safitza and B.
ena. However, it is of note that the earlier trials often resulted in
survivals of over 90%, sometimes even over 95%. The reason for this
higher survival may have had little to do with the addition of linseed oil,
but rather due to the handling and transfer of larvae onto the new blocks
of diet. During prior experiments the larvae were left to find the new diet
for themselves and the old diet was left in the petri dish. Therefore, it may
be better to avoid disturbing the larvae as much as possible, particularly
during the first instar when the larvae is most prone to damage.

Another factor shown to be of importance, to Bicyclus butterflies at
least, was the freshness of the diet. Butterfly larvae usually feed on fresh
material with a high water content. It is possible that the diet dried out
over a week to the extent that it was detrimental to the insect. Here again
is a factor to which many moth species are less susceptible. We
successfully reared the warehouse moth, Ephestia kuhniella, on the diet
(minus linseed oil) without renewing the diet throughout the entire
development period. Of course, E. kuhniella is a stored product insect
and, as such, is extremely resistant to low moisture conditions.

The diet described in Table 1 opens up the possibility of certain types
of experiments, particularly involving the study of colour patterns, that
would be difficult to perform using living plants. Examples include the

128

J. Res. Lepid.

manipulation of development time independent of temperature to exam-
ine the subsequent effect on wing pattern, and the study of pigments
sequestered from the host plant in larvae, pupae and adults. However,
the diet does not appear to be suitable for the maintenance of stock
animals. Diet reared butterflies had a low fecundity and living plants are
also required for the deposition of eggs.

A second important reason for developing artificial diets is to enable
amateur entomologists to rear a few mint condition specimens through
to adult to be added to collections without the need to use living plants.
Morton (1979) experimented with rearing butterflies on artificial diets
and presented a recipe with which he successfully reared 50 species.
However, Morton’s recipe is considerably more complicated than ours
(see Table 1) and the need to develop a simple diet without expensive and
difficult to obtain chemicals is of paramount importance if it is to be
widely used. Furthermore, Morton found that phagostimulants were
usually required in the diet in the form of dried natural host plant
material. Only Bicyclus species are considered in the present study, but
no dried maize leaves were needed in the diet to induce feeding. With
edge feeding species, Morton shredded his diet through a cheese grater
to facilitate feeding. Bicyclus species are also edge feeders (grasses), but
did not require the presentation of the diet to be modified in any way
beforehand.

The simplicity of the diet presented here and the ease with which it can
be used warrents further investigation into its applicability with other
species of butterflies. The colour pattern of the species used here did not
appear to be effected by the diet, but the size of the resulting animal did.
This may be an important consideration if collectors’ specimens are to be
reared, although it may be possible to overcome the problem through
slight modifications of the basic diet.

Acknowledgements. We are grateful to Prof. M. Boppre for supplying a sample of
his diet mixture which was used to initiate the present study and to Prof. H. F.
Nijhout for suggesting the use of linseed oil.

Literature Cited

Bergomaz, R. & M. Boppre, 1986. A simple instant diet for rearing Arctiidae and
other moths. J. Lepidopt. Soc. 40: 131-137.

Brakefield, P.M. & N. Reitsma, 1991. Phenotypic plasticity, seasonal climate and
the population biology of Bicyclus butterflies (Satyridae) in Malawi. Ecol.
Entomol. (in press).

Morton, R.S. 1979. Rearing butterflies on artificial diets. J. Res. Lepid. 18: 221-
227.

Singh, P. & R.F Moore, 1985. Handbook of insect rearing ( Vol. II ). Elsevier Science
Publishers, Amsterdam.

Vanderzant, E.S. 1967. Wheat germ diets for insects: rearing the boll weevil and
the saltmarsh caterpillar. Ann. Entomol. Soc. Am. 60: 1062-1066.

. 1974. Development., significance, and application of artificial diets for

insects. Ann. Rev. Entomol. 19: 139-160.

Wielgus R.S. 1974. Artificial diet: the key to the mass rearing of Megathymus
larvae. J. Res. Lepid. 13: 271-277.

Journal of Research on the Lepidoptera

30(1-2):129-139, 1991

Behavior of Male Desert Hackberry Butterflies,
Asterocampa leilia (Nymphalidae) at Perching Sites
used in Mate Location

Ronald L. Rutowski, Janis L. Dickinson1, and Barbara Terkanian

Department of Zoology, Arizona State University, Tempe, AZ 85287-1501

Abstract. Males of the desert hackberry butterfly (. Asterocampa leilia)
occupy and defend perches at mate encounter sites adjacent to the
larval foodplant. When returning from an interaction with a conspe-
cific, a male selects a perch within about 2 m2, surrounding his original
perch. These perching sites are surrounded by areas of significantly
lower vegetation than other areas around the larval foodplant.

Over the course of the morning activity period the perching behavior
of the males changes in ways that are quantitatively documented.
Early in the morning males perch on the ground with the wings open
facing away from the sun. Later they perch facing the same direction
but with the wings closed. Later still they perch on vegetation a little
less than a meter above the ground with wings closed, facing out of the
plant on which they are perched. The perch preferences and orienta-
tion of males when perched are discussed in light of the hypothesis that
males maximize their ability to detect males and females flying in their
vicinity, yet maintain tolerable body temperatures.

Introduction

In many insect species, males occupy and defend encounter sites as
part of their mate-locating effort (Thornhill and Alcock, 1983). The
selection of these sites and the behavior of males at them is thought to be
structured in ways that increase their contacts with receptive females.
Some studies have shown that males that defend encounter sites have a
higher rate of contacting receptive females than males that do not defend,
and that the most attractive areas, whether defended or not, are those
with the highest arrival rate of receptive females (Borgia, 1982; Courtney
and Parker, 1985; Forsyth and Montgomerie, 1987; Lederhouse, 1982;
Severinghaus et al., 1981; Shelley, 1987; Wickman, 1985). However,
there have been few studies that have quantified in detail the behavior
of males within an encounter site or the physical characteristics of these
sites to see if males select sites and orientations with characteristics that
might enhance their ability to detect females that fly by.

In the desert hackberry butterfly (. Asterocampa leilia Edwards), males
actively defend perching sites that are on or adjacent to the larval
foodplant, the desert hackberry tree ( Celtis pallida Torrey; Austin, 1977;
Rutowski and Gilchrist, 1988). From their perches, which can be on the
ground or the hackberry, males fly out at conspeciflcs; intruding males
are chased, and females are courted. Males use these sites to detect
newly-emerged virgin females as they first fly from their pupation sites

Hastings Natural History Reservation, Star Route Box 80, Carmel Valley, California 93924

130

J. Res. Lepid.

on or near the larval foodplant (Rutowski and Gilchrist, 1988). Perch
sites are most frequently occupied and most intensely defended in the
midmorning hours in central Arizona. In addition, some sites used by
males are highly attractive and are occupied on a daily basis during the
flight season. These attributes make A. leilia a good species for a study
of perching behavior and perching site characteristics.

Males visually detect females and other males. Hence, in this study we
have focused on those aspects of male perch selection and perching
behavior within a defended encounter site that might affect a male’s
visual field. To begin, we present measures of the size of perching sites
based on the distribution of perches adjacent to several hackberry trees.
These data give some indication of the constancy of perch selection both
within and between males. Next, we describe the distribution of vegeta-
tion around perching sites to see if views from the perches that males
select are more expansive than those from other spots around the
hackberry tree. Finally, we quantify how perched males are oriented
with respect to the sun, the larval foodplant, and other variables that
might affect a male’s view. The results are discussed in light of the
expectation that males behave in ways that enhance their ability to
detect females flying near their perching sites.

For clarity we define a perch as a specific point on the ground or on a
hackberry tree where a male has landed and a perching site as an area,
typically within an encounter site at the perimeter of a hackberry tree,
that encompasses the perches selected by one or more males. Lastly, a
sortie is when a male leaves a perch to chase a flying conspecific or other
animal, or a thrown object, and then returns to the perch site.

Methods

Study sites. The two study sites in central Arizona have been previously
described (Rutowski and Gilchrist, 1988; Rutowski et al., 1988). One was off of
the Bush Highway near the Salt River northeast of Mesa, Arizona. The second
was east of State Route 87 (Beeline Highway) where the highway crosses
Sycamore Creek south of Sunflower, Arizona. In both areas the predominant
trees were paloverde ( Cercidium spp.), mesquite ( Prosopsis spp.), and hackberry.
The study was conducted between April and June in 1988 and 1990.

Location of perches within a perching site. We documented the distribu-
tion of perches selected within a defended site to define the size and location of
the perching site. We identified and mapped 6 sites (SC1-SC6) where males were
regularly seen perched in the Sycamore Creek area. Sites SCI, SC2, and SC3
were located in a dry stream bed, while the other three were located along sandy
corridors in vegetation in the broad floodplain of Sycamore Creek. A 10 m
baseline was established along the stream bed or corridor at each site and was
used to map the locations of perches, prominent vegetation, rocks, etc.

At each of the six sites we obtained the location of five perches for each of 10
males. These observations were made between 0815 and 1045 MST between 17
and 26 May 1988. For each male we noted where on the site he was perched when
first seen and where he perched when he returned from his next four sorties,
which were either stimulated by an animal flying by (usually a conspecific) or by

30(1-2):129-139, 1991

131

a rock thrown over the male, or occurred for no apparent reason. If we lost sight
of a male during a sortie the observation on that male was terminated.

This protocol was developed from our knowledge of male behavior which
suggested that males typically spend 30 min or less at a site before moving on to
another, especially early in the activity period (Rutowski and Gilchrist, 1988).
Hence, 5 sorties were taken as a reasonable representation of perch selection by
a male at a site. Moreover, the rate of interactions is greatest early in the activity
period and falls off quickly as the morning progresses. Tossing rocks helped us
stimulate sorties at times when passing conspecifics were relatively rare. There
was no evidence that males were traumatized by the rocks; they chased the rocks
briefly before returning to perch.

For the set of 5 perches for each male within a site we calculated the distance
between sequential perches and the distance between the place the male was first
seen perched and each of the 4 subsequent perches. Then, using a digitizing pad
and maps of each male’s perches we measured the area of the polygon that
included all perches for males that used at least three different perches. If a male
used only one or two perches, area measurements were not possible.

Male perch sites relative to the hackberry tree. Occupied perching sites
were identified by first finding a perched male and noting the location of his
perch. We then threw rocks over him and elicited sorties. If after each of three
sorties the male selected a perch within one meter of his original perch we
identified the male as the resident of an occupied perching site. We then
examined various aspects of the perch at which he was first seen. These
observations were only made on males initially observed perched in full sun.

We first determined if males preferentially perched on a particular side of a
hackberry tree, such as the south side. To do this we characterized the position
of the perch at which a male was first seen with respect to the compass bearing
from the center of the nearest hackberry to the perch.

Next, we tested the prediction that males select perches that are in sites with
a relatively large view by describing the distribution of vegetation around a perch
in an occupied site.

From the original perch we measured the distance to the nearest vegetation
over one meter high at 0, 45, 90, 135, 180, 225, 270, and 315 degrees relative to
the bearing from the perch into the center of the nearest hackberry plant. We
then made the same measurements for a spot the same distance from the
perimeter of the hackberry tree as the perch, but at a point around the perimeter
randomly selected by coin toss. The point was selected from 8 possible points at
45 degree increments starting at magnetic north. A paired t-test was used to
evaluate the magnitude and direction of the differences in area between male-
selected and randomly-selected sites.

Male orientation when perching. Male perch sites were identified by the
three-perches-within-l-m criterion as described above. For the first perch
observed we noted where the male was perched (ground or vegetation), the
direction the male faced, the distance of the male from the nearest hackberry
plant, the direction from the male to the sun’s azimuth, the time of day, and the
direction to the center of the nearest hackberry tree.

Directional measurements and statistical evaluation. Directional mea-
surements in this study were made with a hand-held compass relative to
magnetic north or to the nearest hackberry tree depending on the purpose of the
measurement. Summaries and analyses of directional measurements were
made using techniques described in Batschelet (1981).

132

J. Res. Lepid.

All parametric summary statistics are given as mean ± standard deviation.
The results of all statistical tests are evaluated at the 0.05 level of significance.

Results

PERCHING SITE CHARACTERISTICS: SIZE
After a sortie, a male did not usually return to exactly the same perch.
In the 240 sorties that we observed, males returned to the same spot only
39 times. Thirty-six (60%) of the 60 males whose five perches were
recorded perched in a new location after each sortie. The remaining 24
males when returning from a sortie alit on a perch they had previously
used at least once. Of the males that used less than 5 different perches,
7 used 4, 7 used 3, 7 used 2, and 3 males used only a single perch.

Still, over a series of five perches males tended to return to the same
general area. The average distance between sequential perches observed
at each site was about 1 meter (Table 1). The perches selected after each
of four sorties were within 2 m of the original perch on a site (Table 1). Fig.
1 shows the distance subsequent perches fell from the first perch a male
selected. Although there was a slight trend for males to increase their
distance from the first perch with each sortie this trend was usually
reversed with the fifth perch.

Across the 6 sites at Sycamore Creek, the average area that included
all perches for males that used three or more different perches was 1.58
± 1.57 m2 (n = 50). The summary statistics for the area measurements

Table 1. The distances moved and areas including all perches for males
observed at the 6 Sycamore Creek study sites. Sample size for each mean is 10.

Site

SCI SC2 SC3 SC4 SC5 SC6

Average distance from 1st perch (m)

Mean

2.16

1.49

2.25

1.18

1.58

1.26

STD

1.34

0.98

1.39

1.4

1.31

1.0

Min

0.66

0.42

0.29

0

0.31

0.112

Max

5.51

3.44

4.75

3.81

4.15

2.98

Average <
Mean

distance moved (m)

1.6 1.09

1.74

0.87

1.24

0.95

STD

0.84

0.68

0.78

0.91

0.87

0.88

Min

0.31

0.32

0.43

0

0.1

0.05

Max

3.34

2.66

3.29

2.95

2.76

3.05

Area including all perches for an

individual male (m2)

Mean

2.22

1.52

1.95

1.55

0.85

0.97

STD

1.49

2.10

1.17

1.72

0.8

1.74

Min

0.3

0.07

0.02

0.39

0.13

0.05

Max

4.43

7.23

4.09

5.06

2.15

4.66

N

10

10

10

6

8

6

30(1-2):129-139, 1991

133

3.0

O 2.5

LLJ

a,

3? 2.0

o

DC 1 R

LL 1-0

LU

O

z

< 1.0

CO

Q

5 0.5

<

LU

0.0

Figure 1 . The distance subsequent perches were from the first perch. This is plotted
for the 1 0 males at each site that were observed to perch five times. Each
point represents the mean for the 10 males at that site. Legend: SCI
(square), SC2 (+), SC3 (diamond), SC4 (triangle), SC5 (x), SC6 (circle).

2 3 4 5

PERCH NUMBER

from each site are given in Table 1. These statistics exclude the 10 males
that used less than 3 different perches and so constitutes an overesti-
mate. The size of the perching areas did not differ between sites for males
that used 3 or more perches (ANOVA, F = 0.97, 5,44 df, p = 0.47). The
mean location for each of the 10 males observed to perch 5 times is shown
for each site in Fig. 2.

PERCHING SITE CHARACTERISTICS: PLACEMENT

How were perches placed with respect to the compass bearing from the
middle of the nearest hackberry tree? The distributions of perches
around the nearest larval foodplant relative to magnetic north are shown
for perches on the ground and on the plant in Fig. 3. Males were randomly
distributed with respect to the compass bearing out of the nearest
hackberry tree regardless of perch substrate (ground: Rayleigh test, r =
0.093, n = 26, p > 0.78; hackberry: Rayleigh test, r = 0.316, n = 23, p > 0. 1);
hence, perching areas are not located in any particular direction, north
east, south, or west, from the nearest hackberry.

Males perch in areas relatively free of vegetation. The size of the open
area surrounding a perched male (57.6 ± 6.04 m2, range 23 - 85.9 m2) was
significantly larger than that surrounding randomly-selected points
adjacent to the hackberry tree (12.6 ± 16.19 m2, range 0 - 41.7 m2; paired
t-test = 9.647, 9 df, p = 0.0005).

134

J. Res . Lepid.

. SC 3

.Ml1111

llll||||i1|.d[j

♦

♦ III'

♦

rrrt|

♦♦ ♦xt!/

fffrrrr

A '

<
I —
CO

DISTANCE IN METERS

Figure 2. Maps of sites SCI through SC6 showing the mean location for each of 1 0
males that were observed to perch five times on the site. Solid lines indicate
the borders of hackberry trees, dashed lines indicate the edge of other
vegetation, except in SCI where the dashed line indicates the edge of a
small cliff, and the crosshatching indicates the direction into the vegetation.

MALE ORIENTATION WHEN PERCHED
What factors influence the direction a male faces when he lands at a
perch? We examined three possibilities: (1) the type of perch (ground or
vegetation), (2) the direction to the nearest hackberry, and (3) the
direction to the sun’s azimuth. Early in the day virtually all males
perched on the ground or on low rocks, sticks or other objects (e.g. dried
cow excrement) adjacent to a hackberry tree. However, as the morning
progressed an increasing proportion of males perched on the tree about
1 m above the ground (Fig. 4; x2 = 51.5, 7 df, p < 10'5). Perches on the
ground were 1.37 ± 0.81 m (range, 0.1 - 3.4 m; n = 26) from the perimeter
of the nearest hackberry. Perches on hackberry were 0.87 ± 0.23 m
(range, 0.1 - 1.4 m; n = 52) from the ground.

% MALES ONGROUND

30(1-2):129-139, 1991

135

ON GROUND ON HACKBERRY

Figure 3. The compass bearing from the center of the nearest hackberry tree to the
perch for perches on the ground (n = 26) and on the tree (n = 23).

08:00 08:30 09:00 09:30 10:00 10:30 11:00 11:30

TIME OF DAY (MST)

Figure 4. The change in the proportion of individuals perched on the ground as a
function of time of day. The number above each bar is the number of males
observed at that time.

Fig. 5 shows the direction perched males faced relative to the sun’s
azimuth and the nearest hackberry tree for males perched on the ground
(n = 26) and on the hackberry (n = 23). The direction males faced when
perched on the ground was significantly correlated with the direction to
the sun (Circular rank correlation, r2 = 0.23, p < 0.01) but was not

136

J. Res. Lepid.

Figure 5. Directions faced by males perched on the ground (n = 26) and on the
hackberry tree (n = 23) relative to the sun’s azimuth and relative to the
center of the hackberry plant.

correlated with the direction to the nearest hackberry tree (Circular rank
correlation, r2 = 0.003, p > 0.99). Hence, on the ground males faced away
from the sun (mean difference between direction male faced and bearing
to the sun =177 degrees) and showed no special orientation to the plant.
Because of this negative orientation to the sun and the restricted range
of values for bearing to the sun’s azimuth the distribution of directions
males faced was significantly non-random (Rayleigh test, r = .64, p <
0.001).

In contrast, when perched on the tree, the distribution of compass
bearings males faced was random (Rayleigh test, r = 0.17, p > 0.05).
However, the direction they faced was significantly correlated with the
direction into the hackberry (Circular rank correlation, r2 = 0.34, p <

30(1-2):129-139, 1991

137

0.002) and was not correlated with the direction to the sun (Circular rank
correlation, r2 = 0.16, p > 0.7). Males perched on the hackberry, then,
faced out of the plant (mean difference in male bearing and bearing from
male to plant center = 190.4 degrees) and displayed no special orientation
to the sun.

DISCUSSION

The data reveal several features of the perches at encounter sites used
by A. leilia males in mate location that have been suggested by previous
authors (Austin, 1977; Rutowski and Gilchrist, 1988), but have not been
previously documented. First, males select perching areas in open areas
adjacent to the larval foodplant. Second, the perches selected are low,
either on the ground or within one meter of the ground. Third, the
perches selected at a site by a male tend to be within a meter or two of one
another. Fourth, male body orientation changes in a correlated fashion
with changes in perch substrate over the course of a morning. Early on,
males perch on the ground facing away from the sun, at first with the
wings spread and later with the wings closed. Later in the morning, they
switch to perches on the larval foodplant and face out of the plant.

These changes in perch and body orientation preferences at encounter
sites suggest that two primary selective factors have shaped the evolu-
tion of male perching behavior within an encounter site. The first is
selection favoring perch and body orientation preferences that produce
a visible field with characteristics that enhance the detection of passing
conspecifics (potential mates and intruding males) and predators. We
assume for the time being that being perched on or next to a desert
hackberry tree will produce the highest rate of encounters with receptive
females and focus this discussion on whether or not males behave in ways
that maximize the likelihood of detecting those passing animals.

The problem faced by perched males is that of detecting small, rapidly-
moving objects passing nearby which is certain to be affected by the
features of the background against which the objects are viewed. Studies
of visual system operation in insects and other animals suggests that a
bright, uniform background such as the sky or distant vegetation is best
for detection of small moving objects (Hailman, 1977; Horridge, 1977).
The apparent preferences of males for perches in open areas would avoid
obscuring vegetation and maximize the part of the visual field occupied
by sky. In addition, facing way from the sun may avoid excessively bright
backgrounds against which detection of small moving objects might be
difficult.

However, selection favoring preference for open areas should be af-
fected by selection favoring the avoidance of extreme body temperatures.
Butterflies, like most insects, are ectotherms which means their body
temperatures are greatly affected by solar radiation incident on their
body and convective heat gain from the environment (for review, see:
Clench, 1966; Kingsolver, 1985). Nonetheless, in a broad range of

138

J. Res, Lepid,

ambient air temperatures butterflies maintain their body temperature
within a restricted range through behavioral adjustments affecting heat
gain. In A. leilia, male behavior at perch sites may reflect thermoregu-
latory activities.

Early in the morning, solar heat gain can be important for maintaining
body temperatures that exceed the air temperature and permit the
butterflies to be active (Kingsolver, 1983a, b). Hence, at this time of day,
exposure to solar radiation can be advantageous and may have favored
males that perch in locations and orientations (facing away from the sun
with the wings open (“dorsal basking”)) that enhance solar heat gain.

Later in the morning, incident solar radiation as well as high air
temperatures near the substrate may favor postural adjustments that
reduce heat gain like those seen in other butterflies (Rawlins, 1980)to
avoid intolerable body temperatures. Facing away from the sun with the
wings closed minimizes the surface area exposed to solar radiation and
so reduces heat gain by this avenue.

As the morning passes and air and substrate temperatures rise, the
ground must often become too hot to use as a perch location as has been
senn in the males of a territorial digger wasp (O’Neill and O’Neill, 1988).
In early summer we have measured substrate surface temperatures late
in the activity period as high as 54 degrees C. At this time males move
into the shade (Austin, 1977) and up onto vegetation where it is presum-
ably cooler. Their body orientation is no longer relative to the sun when
perched on the plant, but instead is relative to the plant. They may face
out of the plant to maintain a view of the sky in as much of their visual
field as possible.

The sequence of perch site preferences displayed suggests that males
prefer ground perches out a meter or so from the edge of the larval
foodplant but that thermoregulatory concerns may drive males to less
satisfactory perch locations. Currently we are attempting to test this
scenario by gathering three types of information. One is data on the
flight paths of females and males as they pass near perching sites.
Perhaps the preference changes over the morning reflect changes in the
flight paths and altitudes at which conspecifics are likely to enter a
male’s perching site. The second is data on the thermal biology of A. leilia.
We are measuring male body temperatures under different conditions
and characterizing the thermal environment. We expect substrate
preferences and body orientations to be structured in a way that keeps
body temperature within the thermal preferences of A. leilia. The third
type of data are on the sensitivities of the visual system in A. leilia and
the visual field characteristics that permit detection of conspecifics from
the greatest possible distance from a male’s perch. These data should
permit a more complete understanding of the variables that influence the
selection of perches in species such as A. leilia that use perches to make
contact with airborne resources such a mates or prey.

Acknowledgements. We thank Tim Lukascko, Steve Schoech, Ruth Stanford,
Mike Demlong, and Thad Leffingwell for assistance in the field, Dr. P.-O. Wickman

30(1-2):129-139, 1991

139

for helpful comments on a preliminary draft of the manuscript, the National
Science Foundation (BNS 83-00317) and the Arizona State University College of
Liberal Arts and Sciences for financial support, and Prof. Rhondda Jones and the
Department of Zoology of Jaine Cook University of North Queensland (Townsville,
Australia) for use of the digitizing pad and other facilities used in preparation of the
article.

Literature Cited

Austin, G. T. 1977. Notes on the behavior of Asterocampa leilia (Nymphalidae) in
southern Arizona. J. Lep. Soc. 31: 111-118.

Batschelet, E. 1981. Circular Statistics in Biology (Mathematics in Biology).
Academic Press, London.

Borgia, G. 1982. Experimental changes in resource structure and male density:
size-related differences in mating success among male Scatophaga stercoraria.
Evolution 36: 307-315.

Clench, H. K. 1966. Behavioral thermoregulation in butterflies. Ecology 47: 1021-
1034.

Courtney, S. P. , & G. A. Parker. 1985. Mating behaviour of the tiger blue butterfly
( Tarucus theosphratus ): competitive mate searching when not all females are
captured. Behav. Ecol. Sociobiol. 17: 213-221.

Forsyth, A., & R. D. Montgomerie. 1987. Alternative reproductive tactics in the
territorial damselfly Calopteryx maculata : sneaking by older males. Behav.
Ecol. Sociobiol. 21: 73-81.

Hailman, J. 1977. Optical Signals. Indiana University Press. Bloomington.
Horridge, G. A. 1977. The compound eye of insects. Scientific American 237: 108-
120.

Kingsolver, J. 1983a. Thermoregulation and flight in Colias butterflies: elevational
patterns and mechanistic limitations. Ecology 64: 534-545.

1983b. Ecological significance of flight activity in Colias butterflies: implications

for reproductive strategy and population structure. Ecology 64: 546-551.

1985. Butterfly thermoregulation: organismic mechanisms and population

consequences. J. Research Lepid. 24: 1-20.

O’Neill, K. M., & R. P. O’Neill. 1988. Thermal stress and microhabitat selection
in territorial males of the digger wasp Philanthus psyche (Hymenoptera:
Sphecidae). J. Therm. Biol. 13: 15-20.

Rutowski, R. L., & G. W. Gilchrist. 1988. Mate-locating behavior of the desert
hackberry butterfly, Asterocampa leilia (Nymphalidae). J. Research Lepid. 26:
1-12.

Rutowski, R. L., G. W. Gilchrist, & B. Terkanian. 1988. Male mate-locating
behavior in Euphydryas chalcedona (Lepidoptera: Nymphalidae) related to
pupation site preferences. J. Insect Behav. 1: 277-289.

Severinghaus, L. L., B. H. Kurtak, & G. C. Eickwort. 1981. The reproductive
behavior of Anthidium manicatum (Hymenoptera: Megachilidae) and the
significance of size for territorial males. Behav. Ecol. Sociobiol. 9: 51-58.
Shelley, T. E. 1987. Lek behaviour of a Hawaiian Drosophila : male spacing,
aggression, and female visitation. Animal Behav. 35: 1394-1404.

Thornhill, R., & J. Alcock. 1983. The Evolution of Insect Mating Systems.

Harvard University Press, Cambridge, Mass.

Wickman, P. O. 1985. Territorial defense and mating success in males of the small
heath butterfly Coenonympha pamphilus ( L. ) (Lepidoptera: Satyridae ). Behav.
Ecol. Sociobiol. 16: 233-238.

30(1-2):140-141, 1991

Note

Hostplant record for Eunica bechina magnipunctata
(Nymphalidae) and observations on oviposition sites and
immature biology

The genus Eunica Hiibner, 1819 has 45 species distributed throughout the
Neotropical region. The vast majority are found in the Andean Region and in the
Amazon Valley (Jenkins 1990). The current knowledge on hostplant utilization
and immature biology of Eunica is restricted to seven species and the available
information is incomplete. Plants in the families Euphorbiaceae, Burseraceae
and Rutaceae are the most frequently recorded foodplants for Eunica (Barcant
1970; DeVries 1986, 1987; Ackery 1988; Jenkins 1990; and citations therein).
Eunica bechina magnipunctata Talbot, 1928 occurs in Southeast Brazil (Jenkins

1990) , where it is common in the cerrados (savanna-like vegetation) of the State
of Sao Paulo (Oliveira 1988). This study was carried out in a cerrado area in the
county of Itirapina (21°15’S, 47°49’W), Sao Paulo. The vegetation consists of a
scrub of shrubs and trees, which is the cerrado sensu stricto of Goodland (1971).

Eggs and larvae of E. bechina magnipunctata were observed on shrubs and
trees of Caryocar brasiliense Camb. (Caryocaraceae), one of the most common
plant species in the cerrado of Itirapina (Oliveira 1988). The eggs are yellowish,
conical, and flattened at the top; bear 12-14 longitudinal ridges and 10-12
transversal ones; average 0.76 mm high (sd=0.031 mm; n=15) and 0.72 mm
diameter (sd=0.057; n=15). The preference for oviposition site within the
hostplant was estimated through a census of 27 shrubs of Caryocar (35-150 cm
tall). The eggs (n=141) are laid singly on young leaves (87%), shoot tips (10%),
petiole (1%), stem (1%) and mature leaves (1%) of C. brasiliense. Although
Eunica was seen on Caryocar from September to January (rainy season), the
highest infestation level occurs between September and October when the
majority of the leaves are still young, soft, and red in color. The vertical
distribution of the eggs within the hostplant varied from 3 to 150 cm above ground
(x=60.5 ± 44.8 cm; n=141). Caterpillars were observed feeding preferentially on
young leaves of Caryocar. Early instar larvae of E. bechina construct frass
chains, as already described for other Nymphalidae (Casagrande & Mielke 1985;
Muyshondt 1976; DeVries 1987; Aiello 1991). Caryocar brasiliense bears
extrafloral nectaries on the outer surface of the sepals and ants are the most
frequent visitors to the plant in the cerrado (Oliveira 1988; Oliveira & Brandao

1991) . Foraging ants may encounter Eunica caterpillars when these are feeding
on Caryocar leaves, occasionally resulting in the removal of the larvae from the
hostplant. On the other hand ant visitors were never observed climbing on the
frass chains constructed by the larvae, a fact suggesting that this structure may
serve as a protective refuge against ant predation (Freitas & Oliveira, in
preparation).

Acknowledgements. We are grateful to K. S. Brown, Jr. for identifying the butterfly,
and to R. B. Francini and C. F. Klitzke for reviewing the manuscript. The study was
supported by research grants from the Conselho Nacional de Desenvolvimento
Cientffico e Tecnologico, the Funda^ao de Amparo a Pesquisa do Estado de Sao
Paulo, and the Fundo de Apoio ao Ensino e Pesquisa da UNICAMP.

30(1-2):140-141, 1991

141

Literature Cited

Ackery, P. R. 1988. Hostplants and classification: a review of nymphalid butterflies.
Biol. J. Linn. Soc. 33: 95-203.

Aiello, A. 1991. Pudelpha ixia heucas : immature stages and position within Adelpha
(Nymphalidae). J. Lepid. Soc. 45: 181-187.

B arc ant, M. 1970. Butterflies of Trinidad and Tobago. Collins, London.

Casagrande, M. M. & O. H. H. Mielke. 1985. Estagios imaturos d eAgrias claudina
claudianus Staudinger (Lepidoptera, Nymphalidae, Charaxinae). Revta. bras.
Ent. 29: 139-142.

DeVries, P. J. 1986. Hostplant records and natural history notes on Costa Rican
butterflies (Papilionidae, Pieridae & Nymphalidae). J. Res. Lep. 24: 290-333.

. 1987. The butterflies of Costa Rica and their natural history. Princeton

University Press, Princeton, New Jersey.

Goodland, R. 1971. A physiognomic analysis of the cerrado vegetation of central
Brazil. J. Ecol. 59: 411-419.

Jenkins, D. W. 1990. Neotropical Nymphalidae VIII. Revision of E unica. Bull.
AllynMus. 131: 1-177.

Muyshondt, A. 1976. Notes on the life cycle and natural history of butterflies of El
Salvador. VII. Pcrchaeoprepona demophon centralis (Nymphalidae). J. Lepid.
Soc. 30: 23-32.

Oliveira, P. S. 1988. Sobre a interagao de formigas com o pequf do cerrado, Caryocar
brasiliense Camb. (Caryocaraceae): o significado ecologico de nectarios
extraflorais. Doctor in Science Thesis, Universidade Estadual de Campinas,
Sao Paulo.

Oliveira, P. S. & C. R. F. BrandAo. 1991. The ant community associated with
extrafloral nectaries in the Brazilian cerrados. In: C. R. Huxley & D. F. Cutler
(eds.), Interactions between ants and plants. Oxford University Press, Oxford,
pp. 198-212.

Paulo S. Oliveira and Andre V. L. Freitas, Departamento de Zoologia,

C.P. 6109, Universidade Estadual de Campinas, 13081 Campinas SP, Brasil

30(1-2): 142-144, 1991

Book Reviews

BUTTERFLIES OF THE HOLARCTIC REGION. Part I. Papilionidae, Pieridae,
Danaidae, & Satyridae ( Partim ). Bernard d’Abrera. 1990. 185 pp. Hill House,
Victoria, Australia. $240.

Es ist nichts schrecklicher als eine tdtige Unwissenheit.

— Goethe, Maxims and Reflections

It is highly irregular to review only the first part of a projected three-volume
work, but this is a terrible book and this is not an ordinary review. The definitive
review of d’Abrera should come from someone (or a group of someones) competent
in systematics at a global level. I was asked by O. Kudrna to look at this book and
give him my reactions. They were so strong that I thought they should by shared
with a broader public. One tends to turn to the groups one knows best as a
benchmark of quality in a synoptic treatment like this, so I turned to the Pierini.
Under Pieris napi I found a long Jeremiad, probably too long to reproduce here
under copyright “fair use” guidelines, but I can pick out a few gems:

Pieris napi is to many a great enigma. Some workers consider napi to comprise
more than a single species and indeed have erected a “napi complex” to try to fathom
apparent anomalies in morphology and biology ... after a little while the student is
faced with the simple choice of Pieris napi contra Mundi! [=against the world —
AMS1 And so, indeed has the entire Pieris napi cultus arisen. ... My own theory is
that what we are facing is the simple phenomenon of an extremely plastic
polyphagous species, which has spread successfully over a great range, some
populations becoming isolated geographically and forced to become monophagous
because the range of host plants has become limited in the new environments.
Attempts to cross breed any two extreme populations end in frustration because of
the host plant preferences of the parent stock — not necessarily because of any
genetic considerations.

... I have seen the researches of others in this matter — and I am not convinced
by any of them. (In a sense, I too have fallen victim to this insignificant butterfly
— by writing the longest piece I have ever done on a single species!) (p. 74)
Simplicity is in the eye of the beholder, and anyone who beholds the napi
problem and finds it “simple” has a remarkable eye indeed. It is of course possible
that d’Abrera, who may never have reared a single member of the “ napi complex”
in his life, has shot insightfully to the heart of the problem, unlike all the previous
workers, including not only the naming-and-classifying enthusiasts like Ulf
Eitschberger, whose work d’Abrera dismisses with scorn, but the sophisticated
geneticists like Sydney Bowden in Britain and Zdravko Lorkovic in Croatia, who
between them worked on the napi problem for nearly a century, as well. D Abrera
does not allude to them explicitly, not does he explain his disdain for their work
as he does for Eitschberger’s. The underlying basis for his attitude, however, is
no mystery. Bowden and Lorkovic worked in a Darwinian framework, as do
Hansjiirg Geiger and myself. D’Abrera is a self-proclaimed “simple (Creationist)
Anglo-Hispanic,” and is uninterested in evolutionary explanations of anything.
The most complex of problems reduces to simplicity itself if the only acceptable
explanation is that “God did it.”

30(1-2):142-144, 1991

143

God, however, is still presumably the Designer of the genetic systems under-
lying the traits of organisms. It used to be fashionable to divide traits into
“genetic” (i.e., under the direct control of the genes) and “environmental”
(influenced or triggered by events or conditions external to the organism, such as
food, daylength, or temperature). Such a distinction was already conceptually
untenable in the latter decades of the 19th Century. By the middle of the 20th
Century it had been made amply clear (by Schmalhausen, Waddington, and their
successors) that responsiveness to the environment was itself a genetically-
determined trait and subject to natural selection. That is why, several years ago,
I had a bumper-sticker printed up to read “Epigenetics is an epiphenomenon of
the genome.” We are systematically filling in the links in the middle which
separate our fairly detailed knowledge at the two ends of the causal chain
(transmission genetics and environmental physiology; what we need more of is
the intervening molecular mechanisms). A creationist can still accept all of this
and hold it up proudly; it certainly is a tribute to the Creator’s skill and ingenuity.
But d’Abrera prefers to ignore it.

What does it mean to say that the host plant “preferences” of the “parent stocks”
need not have any “genetic considerations?” If the host plant “preferences”
(oviposition? feeding? detoxifying ability? growth suitability? etc., etc.) have no
genetic basis, they should be susceptible to rapid if not immediate change, and
there should be no problem with hybridization. On the other hand, if they have
no genetic basis at all, how are they maintained in nature? And how is it possible
for any trait in a non-human species (for safety’s sake, let us leave human culture
out of the argument) to have no genetic basis?

For the record, the “napi complex” is not polyphagous at all. All of its known
hosts belong to the small cluster of mustard-oil producing families, and nearly all
are in the single family Cruciferae. That is not to argue that they are not
chemically diverse; they are. European napi love Garlic Mustard ( Alliaria
officinalis) while North American ones do not accept it. This interesting observa-
tion suggests a variety of things to an evolutionarily-oriented geneticist or
ecologist. It correlates extremely poorly with the ability of European napi to
hybridize experimentally with different North American populations. Many of
the incompatible combinations manifest themselves in unbreakable, permanent
female diapause, a phenomenon discovered and widely commented upon by
Bowden. What this has to do with “the host plant preferences of parent stock” is
beyond me, though it must be crystal clear to d’Abrera. I wish he would enlighten
us. He may even be able to explain what the Creator had in mind when He
fashioned Haldane’s Rule (the generalization that hybrid incompatibility is more
strongly manifested in the heterogametic sex — males in most animals, but
females in birds and Lepidoptera). Maybe that’s an epiphenomenon of “host plant
preference” too, with no genetic basis...

Perhaps there is no need to continue to excoriate d’Abrera. His attitude is
profoundly anti-scientific — not unscientific, but hostile to science. For a
d’Abrera it is merely foolish and pedantic to investigate interesting-looking
problems in the living world. There are no interesting problems because there is
no hope of solution, because scientific approaches are so fundamentally wrong-
headed. Everything is blissfully simple, and Pieris napi is “insignificant” insofar
as it has nothing special to teach us. But d’Abrera is as bad a natural theologian
as he is a natural scientist. For a God who knows when every sparrow falls, there
are no insignificant creatures. As an inner-city kid once proclaimed: “God made
me, and He didn’t make no junk.”

144

J. Res. Lepid.

D’Abrera expresses his dislike for “that heaviest of all languages — scientific
German!” It is a dislike I share to some degree, though I read a lot of it. But
literary German is less ponderous and even elegant, and it is fitting that Goethe
should so precisely identify d’Abrera’s problem. The quote at the beginning of this
review translates: “Nothing is so terrible as energetic ignorance.”

Arthur M. Shapiro, Department of Zoology and Center for Population
Biology, University of California, Davis, CA 95616 USA

LE PEUPLEMENT DES LEPIDOPTERES DE LA BOURGOGNE. Claude
Dutreix. 1988. Natural History Society of Au tun. 15 rue Saint Antoine, 71400
Autun, France. 278 pp. 150 FF postpaid. (French, English summary)

Although a regional study of butterfly distribution, the area included in this
work is thoroughly and exhaustively covered. This includes the four depart-
ments that comprise the Bourgogne, located in the Northeast of France and
comprising some 30,000 km2. The basic data consists of classifying the contents
of 10 km2 UTM units for the presence of 127 species of butterflies and skippers
which inhabited the region. A background of the geology and climate is
presented, but unfortunately no information on plant communities is incorpo-
rated. Several analytic approaches are given which evaluate both species
diversity and correlation among the squares. Cluster analysis is thorough, and
all the data are shown. Maps of all species are given on a large scale for all of
Europe, a medium scale of France, and in detail for the Bourgogne. This is an
exceptionally data rich work accompanied by protocols and their rationale which
will be of interest to anyone working in biogeography.

R. H. T. Mattoni, 9620 Heather Road, Beverly Hills, CA 90210, USA.

INSTRUCTIONS TO AUTHORS

Manuscript format: Two copies must be submitted, double-spaced, typed, with wide
margins. Number all pages consecutively. If possible italicize rather than underline scientific
names and emphasized words. Footnotes are discouraged. Do not hyphenate words at the right
margin. All measurements must be metric. Time must be cited on a 24-hour basis, standard
time. Abbreviations must follow common usage. Dates should be cited as: day -Arabic numeral;
month-Roman numeral; year-Arabic numeral (ex. 6. IV. 1992). Numerals must be used for ten
and greater e.g. nine butterflies, 12 moths.

Electronic submission: The Journal is now being produced via desktop publishing, allowing
much shorter publication times. Although typewritten manuscripts are acceptable, those
submitted on computer disk are highly preferred. After being notified of your paper's
acceptance, submit either a Macintosh or IBM disk version. Include on your disk both the fully
formatted copy from your word processing program and a text-only (ASCII) copy. The two most
preferred formats are Microsoft Word for the Macintosh and either Microsoft Word or Word
Perfect for the IBM, although translation utilities will allow conversion from most formats. Put
returns only at the ends of paragraphs, not at the end of each line. Use one tab to indent each
paragraph. Even if your printer is incapable of outputting italics, please specify italics rather
than underline in your disk copy. Please note any special characters that are used in either the
body of the text or the tables (e.g. e, u, °, §, p, ?, /).

Title material: All papers must include the title, author's name, author's address, and any
titular reference and institutional approval reference. A family citation must be given in
parenthesis (Lepidoptera: Hesperiidae) for referencing.

Abstracts and Short Papers: All papers exceeding three typed pages must be accompanied
by an abstract of no more than 300 words. Neither an additional summary nor key words are
required.

Name citations and Systematic Works: The first mention of any organism should include
the full scientific name with unabbreviated author and year of description. There must be
conformity to the current International Code of Zoological Nomenclature. We strongly urge
depositing of types in major museums, all type depositories must be cited.

References: All citations in the text must be alphabetically listed under Literature Cited in
the format given in recent issues. Abbreviations must conform to the World List of Scientific
Periodicals. Do not underline or italicize periodicals. If four or less references are cited, please
cite in body of text not in Literature Cited. For multiple citations by the same author(s), use
six hyphens rather than repeating the author’s name.

Tables: When formulating tables, keep in mind that the final table will fill a maximum space
of 11.5 by 19 cm either horizontally or vertically oriented. Number tables with Arabic
numerals. When submitting tables on disk, use tabs between columns rather than multiple
spaces.

Illustrations: Color can be submitted as either a transparency or print, the quality of which
is critical. Black and white photographs should be submitted on glossy paper, and, as with line
drawings, must be mounted on stiff white cardboard. Authors must plan on illustrations for
reduction to page size. Allowance should be made for legends beneath, unless many consecutive
pages are used. Drawings should be in India ink. Include a metric scale. Each figure should
be cited and explained as such. Each illustration must be identified by author and title on the
back. Indicate whether you want the illustration returned at your expense. Retain original
illustrations until paper is accepted. Legends should be separately typed on pages entitled
“Explanation of Figures.” Number legends consecutively with separate paragraph for each
page of illustration.

Review: All papers will be read by the editor(s) & submitted for formal review to two referees.

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

Volume 30 Number 1-2 Spring 1991

IN THIS ISSUE

Date of Publication: December 1, 1992

The distribution of radiolabeled pigment precursors in the 1

wing patterns of nymphalid butterflies
H. F. Nijhout and P. B. Koch

Atepa, a new Sonoran Euliini genus (Lepidoptera: Tortricidae) 14

Jozef Razowski

Larval Foodplant and Other Effects on Troidine Guild Composition 19

(Papilionidae) in Southeastern Brazil

Ana Beatriz B. de Morais and Keith S. Brown Jr.

An Annotated List of Moths Recorded at Florissant Fossil Beds 38

National Monument, Colorado
Charles V. Covell Jr.

Speciation: A Review of Concepts and Studies with Special Reference 45
to the Lepidoptera

Michael M. Collins

Quantification of Lepidoptera wing patterns using an image analyzer 82
Jack J. Windig

Quantification of Ant Attendance (Myrmecophily) of Lycaenid Larvae 95
Gregory R. Ballmer and Gordon F. Pratt

Appearance of the “heathii” Aberration and Genitalic Variation in a 113

Mitoura Population from Oregon (Lycaenidae: Theclinae)

Clifford D. Ferris

An Artificial Diet for Butterflies, Including Bicyclus Species, and its 121
Effect on Development Period, Weight and Wing Pattern

G. J. Holloway, P. M. Brakefield, S. Kofman and J. J. Windig

Behavior of Male Desert Hackberry Butterflies, Asterocampa leilia 129

(Nymphalidae) at Perching Sites used in Mate Location
Ronald L. Rutowski, Janis L. Dickinson,
and Barbara Terkanian

Note 140

Book Reviews 142

Cover Illustration: Radiolabel incorporation pattern in the wings of Inachis
io by H. F. Nijhout and P. B. Koch.

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

ISSN 0022 4324
Published By:

Founder:
Editorial Staff:

Associate Editors:

The Lepidoptera Research Foundation, Inc.

9620 Heather Road

Beverly Hills, California 90210-1757

TEL (310)274-1052

FAX (310) 275-3290

William Hovanitz

Rudolf H. T. Mattoni, Editor
Scott E. Miller, Assistant Editor

Emilio Balletto, Italy
Henri Descimon, France
Philip DeVries, U.S.A.

Thomas Emmel, U.S.A.

Lawrence Gall, U.S.A.

Hansjuerg Geiger, Switzerland
Otakar Kudrna, Germany
Arthur Shapiro, U.S.A.

Atuhiro Sibatani, Japan
Karel Spitzer, Czechoslovakia

Manuscripts and Notices Material may be sent to the Editor at the above address.
The JOURNAL is sent to all members of the FOUNDATION.

Classes of Membership:

Regular (Individual)
Contributing

Student/Retired-Worldwide
Subscription Rate/Institutions
Life

$

20.00

year (vol.)

$

30.00

or more, year (vol.)

$

18.00

year (vol.)

$

28.00

year (vol.)

$

250.00

STATEMENT OF OWNERSHIP AND MANAGEMENT

THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA is published four times a year by THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The office of publication and the general business
office are located at 9620 Heather Road, Beverly Hills, California 90210. The publisher is THE
LEPIDOPTERA RESEARCH FOUNDATION, INC. The Editor is R. H. T. Mattoni at the above address.
The Secretary-Treasurer is Leona Mattoni at the general business office. All matters pertaining to
membership, dues, and subscriptions should be addressed to her, including inquiry concerning mailing,
missing issues, and change of address. The owner is THE LEPIDOPTERA RESEARCH FOUNDA-
TION, INC. , a non-profit organization incorporated under the laws of the State of California in 1965. The
President is R. H. T. Mattoni, the Vice President is John Emmel, the Secretary-Treasurer is Leona
Mattoni. The Board of Directors is comprised of Barbara Jean Hovanitz, Leona Mattoni, and R. H. T.
Mattoni. There are no bond holders, mortgages, or other security holders.

Journal of Research on the Lepidoptera

30(3-4):145-161, 1991

The Butterflies of Mauritius *

P. M. H. Davis
and

M. J. C. Barnes

4 Leigh Road, Clifton, Bristol, U.K.
and

"Avalon", Upottery, Honiton, Devon EX14 9PQ, ENGLAND

Abstract. All species of butterfly recorded from the Indian Ocean
island of Mauritius are listed, and known bionomic details are pre-
sented. Previous checklists, made in 1866, 1908 and 1938 are dis-
cussed, and taxonomic corrections to these publications are offered.

The status of some species is updated and we add two recently
introduced species, Erionota thrax thrax L., and Virachola antalus
Hopffer, to the list to Mauritian species. In addition to documenting the
turnover of species on this island over the past 120 years, we also hope
to stimulate further investigation into the present status of several
endangered or possibly extinct endemic species.

Introduction

The island of Mauritius is situated at 57°30' east, 20°10' south, in the
southern Indian Ocean. Its total land area is only 1858 km2. The slightly
larger (2502 km2) French island of Reunion — formerly known as
Bourbon — lies some 400 km to the south-west, and the Mauritian
dependency of Rodrigues (103 km2) lies some 1100 km to the east. The
three islands and their lesser dependencies are sometimes referred to as
the Mascarene archipelago. The nearest major land mass to Mauritius
is Madagascar, some 1350 km to the west.

Mauritius has a tropical maritime climate. The months of December
to May are hot and humid, and the months of June to November are
cooler and dryer. The hotter months bring the threat of cyclones, which
frequently devastate the island. The whole island is under intense
cultivation pressure, the principal crop being sugar cane ( Saccharum
officinarum L.). Although dense cloud forest once covered the entire
island, today only a small area survives, in the Black River Gorges in
the south-west of the island. This area is stringently protected but
appears to be ultimately doomed due to the encroachment of introduced
competitor species such as Chinese Guava ( Psidium cattleianum
Sabine) and Privet ( Ligustrum walkeri Decaisne) which strangle the
understorey and prevent tree regeneration.

The present checklist of Mauritian butterflies recognises 35 species,
of which only 25 are at all common — that is, likely to be seen by the

Accepted for publication June 1984

146

J. Res. Lepid.

Table 1. Distribution of the Mascarene butterfly species. Subspecies in parenthesis.

Mauritius

Rodrigues*

Reunion*

Madagascar

Africa

Orient

Danaus chrysippus

X

X

X

X

X

X

Euploea euphon

X

Euploea desjardinsi

X

Euploea goudotii

X

Amauris phaedon

X

(X)

Henotesia narcissus

X{narcissus)

X(narcissus)

Xifraterna)

Melanitis leda

X{helena)

X(helena)

X{helena)

X{helena)

X(helena)

X(leda)

Neptis frobenia

X

Neptis dumetorum

X

Hypolimnas misippus

X

X

X

X

X

X

Hypolimnas bolina

(X)

(X)

X

X

Hypolimnas dubius

X(drucei)

X{drucei)

X(dubius)

Junonia rhadama

X

X

X

X

Junonia goudotii

(X)

X

Vanessa cardui

(X)

(X)

X

X

X

Antanartia borbonica

X(mauritiana)

X(borbonica)

X(borbonica)

Phalanta phalantha

X{aethiopica)

X{aethiopica)

X{aethiopica)

X(aethiopica)

Xiphalantha)

Salamis angustina

X(vinsoni)

X(angustina)

X{angustina )

Libythea cinyras

(X)

Papilio manlius

X

Papilio phorbanta

X

Papilio demodocus

X

X

X

X

Catopsilia florella

X

X

X

X

X

Catopsilia thauruma

X

X

Eurema brigitta

X(pulchella)

X(pulchella)

X(brigitta)

Eurema floricola

X(cercs)

X(ceres)

Xifloricola)

Cacyreus darius

X

X

Leptotes pirithous

X

X

X

X

X

X

Lampides boeticus

X

X

X

X

X

X

Zizina antanossa

X

X

X

X

Zizula hylax

X

X

X

X

X

Zizeeria knysna

X

X

X

X

X

Cyclyrius mandersi

X

Virachola antalus

X

X

X

Coeliades forestan

Xiforestan)

X(forestan)

X(f ores tan )

X(arbogastes)

Xiforestan)

Eagris sabadius

X{sabadius)

X{sabadius)

X{andracne)

X(astoria etc.)

Borbo borbonica

X{borbonica )

X(borbonica)

X(borbonica)

X(borbonica)

X(borbonica)

Parnara naso

X(naso)

X(bigutta)

X{potieri)

X(monasi etc.)

Erionota thrax

X(thrax)

X

*N.B. Distribution data from Reunion and (particularly) Rodrigues are from old sources and may be unreliable.

ordinary visitor. The ten remaining species are either casuals, extinct or
nearly extinct, or only recently established. There is a high incidence of
endemism, with nine species or subspecies being confined to the island;
and a further eight species found only on the island and in other parts of
the Malagasy sub-region (Madagascar, the Mascarenes, the Comoro is.,
and the Seychelles).

For this reason alone the butterfly fauna is of great interest. Table 1
lists the known Mascarene butterflies in comparison with their occur-
rence in other regions. Subspecies are given. Although the comparative
data are of biogeographical interest, we have strong reservations on the
completeness of the data for Reunion and Rodrigues. The latter data
were from lists of the last century. The first list of Mauritian butterflies

30(3-4):145-161, 1991

147

was produced by Trimen in 1866, who recognised 26 species. This list
was updated by Manders in 1908, who added five species and deleted
one, bringing the total to 30. The most recent checklist was produced by
J. M. Vinson in 1938, who catalogued all known Mascarene lepidoptera
and largely followed Manders in respect of the Mauritian butterflies.
The present paper reviews our current knowledge of the Mauritian
butterfly fauna, and attempts to correct some taxonomic errors and
ambiguities present in existing literature. In addition, we add two
newly introduced species to the Mauritian list and offer notes on recent
changes in status of other species. The latter notes are based upon
observations made by P. M. H. and J. P. L. Davis during their stay on the
island between 1976 and 1980.

Further fieldwork on the butterfly fauna of this island is very
desirable. Much of Mauritius consists of private sugar estates which
include hills and scrubland, and to which access is restricted. It is
possible that some of the rare or assumed extinct species may still be
present in these areas, awaiting rediscovery.

To aid identification and to place the Mauritian species in the context
of the African fauna, we have cross-referred the species in our checklist
to their appropriate entries in both Carcasson’s 1981 checklist of
African butterflies (published in his ‘Handguide to the Butterflies of
Africa’), and d’Abrera’s 1980 volume on the Afrotropical fauna. These
works are abbreviated CC and DA respectively in the following.

The Annotated Checklist
DAN AID AE

Danaus chrysippus L. 1758 (CC No. 3185; DA p. 152) is common and
widely distributed over the island. It seems to be monomorphic there, all
specimens seen from 1976 to 1980 being of form chrysippus. Manders
(1908) and Vinson (1938) both claim to have seen single specimens of
this species with white hindwings, which may have been form alcippus ;
however, no further specimens of this form from Mauritius have been
brought to our attention. Of the many hundreds of this species bred on
the island a few had some whitening of the veins in the upperside
hindwing, but not sufficiently so to be considered as form alcippus. Form
dorippus has never been recorded — a surprising observation in the light
of the occasional presence of its presumed mimic Hypolimnas misippus
female-form dorippoides (‘ inaria ) on the island. Like the following two
species, adults of this species were often seen to congregate in the
vicinity of the Boraginacean tree Tournefortia argentea L. for several
days after emergence from the pupa, where they appeared to be
imbibing exudations from the ends of broken branches on the ground. It
is presumed that these exudations contained pyrrolizidine alkaloids
which serve as pheromone precursors and may also be involved in
boosting overall bodily toxicity.

148

J. Res. Lepid.

Euploea euphon Fabricius 1798 (CC No. 3204; DA p. 158). The genus
Euploea is of oriental origin, as it does not occur at all in mainland
Africa. Nevertheless, several of the Indian Ocean islands support
endemic species — for example, E. goudotii Boisduval, which is found
only on Reunion; and E. desjardinsi Guerin, which is confined to
Rodrigues. The endemic Mauritian species, E. euphon , is widely distrib-
uted throughout the island, although it is less common than D.
chrysippus and appears to exhibit some preference for the forested areas
of the Black River Gorges and certain coastal areas. The principal larval
foodplant appears to be oleander ( Nerium oleander L.) although
Manders reports it as also feeding upon Ficus repens Rottboell.

Amauris phaedon Fabricius 1798 (CC No. 3201; DA p. 158) is confined
to Mauritius, although there are doubtful reports of specimens from
Madagascar. It is found occasionally in all parts of the island, but is
especially common in the south-west, near the Le Morne peninsula. It is
most plentiful in areas of coastal vegetation containing the tree Tour-
nefortia argentea , in which the adults spend large parts of the day
resting or imbibing juices. The early stages are, regrettably, not known
although Vinson (1938) states that the larvae may feed upon various
Asclepiadaceae, principally Tylopha asthmatica Wight.

SATYRIDAE

Henotesia narcissus Fabricius 1798 (CC No. 3019; DA p. 186) ssp.
narcissus. This subspecies is confined to Mauritius and Reunion, the
other subspecies — ssp. fraterna Butler — being confined to Madagascar
and Anjouan island in the Comoro group, it is common throughout the
island, especially in shady areas in woods and gardens. Manders (1908)
describes the early stages, giving the larval foodplant as various
Bamhusa species and other gramineae.

Melanitis leda L. 1758 (CC No. 2894; DA p. 162) ssp. helena West-
wood. This species is widely distributed throughout the old world
tropics, being represented in Mauritius by the African subspecies
helena. It is common throughout the island, and particularly active at
dawn and dusk. Manders collected a series of 155 specimens of this
species from Mauritius between March 10 and December 31, 1905,
exhibiting seasonal variation. The wet season forms (Dec. -June) have
more prominent ocelli and lighter basal areas beneath than do the dry
season forms (June-Nov.). The series is deposited in the Hope Collec-
tions, Oxford, U.K. Vinson (1938) gives the larval foodplants on the
island as sugar cane and other gramineae, particularly Thysanoloena
maxima Kuntze.

NYMPHALIDAE

Neptis frobenia Fabricius 1798 (CC No. 2552; DA p. 248) is confined
to Mauritius. It is found most commonly in the hills above the Black

30(3-4):145-161, 1991

149

River Gorges and in the Maccabee forest. Manders (1908) reports the
larval foodplants to be various species of Acalypha and Erythrospermum
mauritiana Baker.

Hypolimnas misippus L. 1764 (CC No. 2664; DA p. 214) is a widespread
cosmopolitan species in the island, but can be very scarce in some years.
During the years 1976-1980 occasional males were seen, especially on
the east coast, but rarely inland. No females were seen at all during this
period. After bad cyclones at the end of 1979 and in early 1980,
accompanied by very heavy rainfall, the butterfly became much more
common. Freshly emerged specimens were often seen inland, at Moka,
and females were seen for the first time. Most females were of the form
H. misippus misippus , which is the presumed mimic of Danaus chry-
sippus chrysippus. However, specimens of the form dorippoides (‘maria’)
were also seen and captured. The presumed model of this mimetic form
isD. chrysippus dorippus , which is not present on the island. If we are to
believe the evidence of specimens in the Hope Collections at Oxford,
both forms have persisted in the island since at least the 18th century.
Manders (1908) reports a single specimen of form alcippoides from the
Port Louis Museum in Mauritius.

Hypolimnas bolina L. 1758 (CC No. 2645; DA p. 214) is probably only an
occasional visitor to the island, or possibly even a mistaken record.
Early records are almost certainly attributable to misidentification of
specimens of H. misippus. Manders (1908), however, mentions two
authenticated Mauritian specimens known to him. No specimens were
seen during the years 1976-1980.

Hypolimnas dubius Palisait de Beauvois 1806 (CC No. 2655; DA p.
222) ssp. drucei Butler is a subspecies of the African H. dubius confined
to Madagascar, Mauritius and the Comoro islands. It must be regarded
as extremely rare or extinct in Mauritius. There are two specimens in
the british Museum (Natural History) taken by a J. T. Rawlins in
November and December 1953. No locality is given. Vinson (1938)
reports a specimen from La Mi Voie, Black River, taken in September
1915.

Junonia rhadama Boisduval 1833 (CC No. 2669; DA pp. 229) ( =
Precis rhadama) is found in Madagascar, Rodrigues, Reunion, the
Comoro islands, and Astove island in the Seychelles, in addition to
Mauritius. According to Trimen (1866) the species was introduced into
Mauritius from Madagascar in 1857 or 1858, and spread rapidly
throughout the island. It is found all over Mauritius, but is especially
abundant at Flic-en-Flac on the east coast. During the day it is found in
large numbers in the vegetation and on the rocks above the tide-line.
Vinson (1938) reports the larval foodplant to be various species of
Barleria.

150

J. Res. Lepid.

Junonia goudotii Boisduval 1833 (CC No. 2673; DA p. 230) (= Precis
goudotii) is confined to Madagascar and the Comoro islands. Only a
single specimen — presumably a vagrant — is recorded by Vinson (1938)
from Long Mountain.

Vanessa cardui L. 1758 (CC No. 2694; DA p. 238) is an almost
cosmopolitan species which can be regarded only as exceedingly rare or
accidental in Mauritius. Occasional specimens from the island can be
found in collections, but between 1967 and 1980 only one or two
unconfirmed sightings were made.

Antanartia borbonica Oberthur 1880 (CC No. 2697; DA p. 239) ssp.
mauritiana Manders. The nomotypical subspecies of A. borbonica is
confined to Reunion and the Tamatave region of E. Madagascar. The
other subspecies, mauritiana , is restricted to Mauritius. It is consider-
ably smaller than the nomotypical race. The two have previously been
considered as subspecies of the African species A. hippomene Hiibner
1823.

This species must be regarded as on the verge of extinction, if it is not
already so. Manders (1908) reports that at the turn of the century the
species was confined to the locality of Curepipe, at 1800 ft. (c. 550 m).
Vinson additionally reports (1938) the presence of other specimens from
Moka at 1200 ft. (c. 360 m). The larval foodplant is reported to be Pilea
urtice folia Blume (family urticaceae) by Manders (1906) who gives a
complete description of this species’ early stages. The flight period is
given as February to March, and sometimes as early as September or as
late as May. During four years on the island (1976-1980) no specimens
were seen.

Phalanta phalantha Drury 1773 (CC No. 2704; DA p. 210) ssp. aethio-
pica Rothschild & Jordan. The nomotypical subspecies of this species is
oriental. Subspecies aethiopica is the African and Malagasy form, and is
the one present in Mauritius. The species is common and widespread
throughout the island, especially in sunny patches in gardens and near
coastal vegetation. The larval foodplants are various Flacourtia species.

Salamis angustina Boisduval 1833 (CC No. 2662; DA p. 224) ssp.
vinsoni Le Cerf (= ‘Salamis augustina’ Auct.). The nomotypical sub-
species of this insect is confined to Reunion and Madagascar. Subspecies
vinsoni is found only on Mauritius. It may be distinguished from the
former subspecies by the greater amount of purplish colouring on the
upper side wing surfaces of vinsoni as compared with angustina. It must
be regarded as a very rare species indeed in Mauritius, if not already
extinct. As early as 1866, Trimen noted this species to be very un-
common, having seen only one or two preserved specimens. Vinson
(1938) records that he captured specimens of this species between 1920
and 1923 from the months of April to September. This agrees with
Manders’ description (1908) of the habits of the Reunion subspecies,

30(3-4):145-161, 1991

151

whose flight time he gives as “April and May and again in September”
(between 0900 hrs. and 1000 hrs.). The larval foodplant is given as sugar
cane. Manders (1908) attributes the decline in numbers of this species to
the introduction of Mynah birds from India. The last recorded specimen
was taken by Dr. J. Bolton in August 1929. No locality is given. This
species closely resembles the Danaid species Euploea euphon , of which it
may be a mimic.

LIBYTHEIDAE

Libythea cinyras Trimen 1866 (CC No. 2101; DA p. 409). Only the
type specimen of this species is known. It was given to Trimen in 1865 by
a Mr. Colville Barclay, who allegedly took it at Moka. The specimen is in
the British Museum (Natural History), and is in exceedingly poor
condition.

PAPILIONXDAE

Papilio manlius Fabricius 1798 (CC No. 492; DA p. 22) is confined to
Mauritius. It is closely related to the Reunion species P. phorbanta L.
1771 and the Malagasy species P. epiphorbas Boisduval 1833. All three
are presumed to have evolved from the same line which produced the
‘blue papilios’ of the mainland such as P. nireus L. 1758 and related
species. P. manlius is fortunately common throughout the island and in
no danger of extinction, as it feeds upon various species of Citrus , which
are widely cultivated, although not intensively sprayed with insecti-
cides. The larvae may be found feeding together with the next species, P.
demodocus. They may be distinguished from the latter as P. manlius
larvae are bright green at all stages. A review of the conservation status
of this species appears in the IUVN ‘Papilionidae Red Data Book’ (1985).

Papilio demodocus Esper 1798 (CC No. 505; DA p. 30) is the common
citrus swallowtail of Africa It is found all over the island and is fairly
common. Like the preceding species it is a Citrus feeder, and its
population size appears to fluctuate with fluctuations in Citrus numbers
due to disease. It was probably accidentally introduced into the island at
some time between 1865 and the turn of the century, following its
deliberate importation into Reunion from Madagascar in 1863.

PIERIDAE

Catopsilia florella Fabricius 1775 (CC No. 557; DA p.53). The taxonomy
of this species and others in the same genus has frequently been the
source of much confusion. Although lists prepared of the island’s
Catopsilia species, and specimens in the collection of the Mauritius
Institute Museum, indicate a plurality of forms, varieties and species,
our own work on the island and in various collections shows to our
satisfaction that only two basic species are present on Mauritius. The

152

J. Res. Lepid.

first of these, the pan- African migrant C. florella is common and
widespread throughout the island. Specimens in the Mauritius Insti-
tute labelled ‘C pyranthe' almost certainly result from early taxonomic
confusion with this oriental species. The two species are separable on
the basis of genitalia (Klots, 1929) and also by the dark forewing apical
band, which is always continuous in C. pyranthe but frequently broken
in the Africa C. florella. Both female forms of C. florella , the white form
pyrene and the yellow form florella are present on the island, the white
form being by far the more abundant. The male is always white. The
larval foodplant is Cassia fistula L.

Catopsilia thauruma Reakirt 1866 (CC No. 558; DA p. 53). Specimens
in the Mauritius Institute labelled as the oriental ‘C. pomona are
almost certainly misidentified specimens of this species, which is
restricted to Madagascar and Mauritius. Although the two species have
often been regarded as conspecific (see, for example, Corbet, 1948) the
two are quite clearly distinguishable on the basis of genitalic differences
(Klots, 1929).

Previously this species has been regarded as two separate species, C.
thauruma and C. grandidieri Mabille 1877, or the latter as a subspecies
of the former. The principal grounds for separation appear to be the
smaller size of the male of C. grandidieri , and the extreme dimorphism
of the two females. Whereas dimorphism of the females of this species is
unquestionable — in common with the females of many pieridae, and
most Catopsilia species — examination of long series of C. thauruma
from Madagascar and Mauritius indicate that there is enormous vari-
ation in size of both males and females within and between populations.
The notion of two separate species or subspecies cannot therefore be
sustained on this basis. We hence suggest, in common with other
Catopsilia species, that the females of the two taxa hitherto regarded as
subspecies of C. thauruma should be relegated to the status of ‘female
forms’, i.e. C. thauruma female-f. thauruma and C. thauruma female-f.
grandidieri. Both are illustrated with their male, from Mauritius, in
Figure 2. Both forms appear to occur in both Mauritius and Madagascar,
the only appreciable difference between the two geographic subspecies
being the consistently smaller size of males and females from Mauri-
tius. This diminution is a phenomenon common to the fauna of many
small islands.

The erection by Le Cerf (1916) of the taxon ‘var. mauritiana ’ of C.
grandidieri from Mauritius appears to have been based on too small a
sample (3 specimens) and equivocal characteristics. It should therefore
be ignored as an invalid infrasubspecific category.

The insect is common, although less so than the preceding species, all
over the island — especially where its larval foodplant grows in abund-
ance, e.g. Beau Bassin. The larval foodplant is Cassia siamea Lamarck.

30(3-4):145-161, 1991

153

Eurema brigitta Stoll 1780 (CC No. 566; DA p. 55) ssp. pulchella
Boisduval. This species is distributed throughout the old world tropics
in a confusing variety of subspecies and forms. Subspecies pulchella is
restricted to Madagascar, Mauritius, the Comoro islands and Aldabra,
and has often been regarded as a separate species from E. brigitta. The
female is frequently dusted with black scales over its entire upperside
wing surface, the overall effect being a dull green coloration. Only the
wet season from, which is illustrated, appears to be at all common on the
island. The dry season form is hardly ever seen. The butterfly is
common, particularly in gardens where it flies with the next species.
The larval foodplant is Cassia mimosoides L.

Eurema floricola Boisduval 1833 (CC No. 565; DA p. 55) ssp. ceres
Butler. This species has been considered conspecific with E. hecabe —
which it superficially resembles — for many years. However, genitalial
examination shows the two should be regarded as separate species
(Paulian & Bernardi, 1951). Eurema floricola is distributed throughout
the Malagasy sub-region in a variety of subspecies, i.e. floricola
(Madagascar); aldabrensis Bernardi (Aldabra); anjuana Bulter
(Comoro islands); and ceres , which is confined to Mauritius and Reunion.
The insect is common throughout Mauritius and found in the same sort
of habitat as the preceding species. Vinson (1938) records the larvae as
feeding upon Desmanthus virgatus Willdenow, Caesalpinia bonducella
Fleming and Leucaena glauca Bentham, data which may indicate
ecological separation from the above species.

LYCAENIDAE

Cacyreus darius Mabille 1877 (CC No. 1845; DA p. 544) is confined to
Madagascar, the Comor islands and Mauritius, and has frequently
been regarded in the past as conspecific with the African C. lingeus Stoll
1782, from which it is superifically indistinguishable. However, as
Stempffer (1943) shows, the two species are separable on the basis of
consistent differences in the valves of the genitalia. C. darius is
probably an import to Mauritius, having first been caught on the island
only at the turn of the century. Manders suggests that it may have been
imported into the island upon a species of Coleus , its larval foodplant,
which was brought in from Madagascar and planted out in the Botanic
Gardens at Curepipe. There is a small series of this species in the Hope
Collections at Oxford, taken by Manders at the turn of the century in
these same Botanic Gardens. By 1938 Vinson noted that the species was
rather scarce, and no specimens at all were seen or taken during the
years 1976-1980. Its status today is unclear. It may be extinct on the
island.

v

154

J. Res. Lepid.

Leptotes pirithous L. 1767 (CC No. 1868; DA p. 546) (= Syntarucus
telicanus Lang 1789) is a very common ‘blue’, found all over the island.
It prefers small bushes and trees and it can be seen flying around and
settling on these, especially near the coast and in gardens. Its overseas
distribution includes Africa, Madagascar, much of Asia and Europe. Its
larval foodplants include many legumes such as Cajanus cajan Druce,
and various other species including Plumbago capensis Willdenow, and
even Lantana camara L.

Lampides boeticus L. 1767 (CC No. 1825; DA p. 541) is an almost
cosmopolitan species. Trimen (1866) found it rather scarce on the island,
being confined mostly to gardens where peas were grown. Since then the
increase in population of the island, and the great increase in the
growing of vegetables, especially peas, has led to this species becoming a
common pest. This is a very fast flying ‘blue’ which can be found in any
garden or field in which there are peas or other legumes.

Zizina antanossa Mabille 1877 (CC No. 1902; DA p. 550) (= Z.
perparva Saalmuller 1884) is distributed over the entire continent of
Africa, including Madagascar, and also occurs on Reunion. Manders
(1908) believes it to be an introduction to the island at about the turn of
the century. The species is common throughout Mauritius, and found
frequently on garden lawns and flower beds, where it flies low and
erratically. It is very similar to Zizeeria knysna Trimen. However, Z.
knysna has a black spot in the centre of the cell on the underside of the
fore wing, which Z. antanossa lacks.

Zizula hylax Fabricius 1775 (CC No. 1906; DA p. 551) (= Z. gaika
Trimen 1862) is distributed over the entire continent of Africa, in-
cluding Madagascar and Reunion and is also present in the orient. It is
very common on Mauritius, but can be overlooked due to its small size
and dull coloration. It was first recorded in the island by Manders in
1907, but may of course have been present unnoticed for some time
before. Mamet (1955) gives the larval foodplant as Lantana camara L.

Zizeeria knysna Trimen 1862 (CC No. 1901; DA p. 550) (= Z. lysimon
Hubner 1803) is distributed throughout Africa, Madagascar and the
Seychelles, and is also present in the orient. It is very common in
Mauritius, and often flies with Z. antanossa. It may be distinguished
however, as Z. knysna has a black spot in the centre of the cell on the
underside of the forewing, which is lacking in Z. antanossa. Mamet
(1955) gives the larval foodplants as Cajanus cajan Druce and Pisum
sativum L.

Cyclyrius mandersi Drue 1907 (CC No. 1867; DA p. 546) (= Nacaduba
mandersi) is confined to Mauritius and was described by Druce from
specimens collected by Manders at the turn of the century. It is said to be
restricted to the coast. Specimens taken by Manders in the collection of
the British Museum (Natural History) and the Hope Collections at

30(3-4):145-161, 1991

155

Oxford indicate coastal localities as far apart as Blue Bay in the south-
east of the island; Le Morne Brabant in the south-west; and Flacq on the
north-east coast. The larval foodplant is given as Caesalpinia bondu-
cella Fleming in Manders’ account of the early stages. The insect is
reported to be a high flyer, in contrast to the other Mauritian members
of this family, which tend to fly close to the ground. It is somewhat
surprising in the light of this information that no specimens of this
species were seen or taken in the years 1967-1980 by P. M. H. and J. P. L.
Davis. We cannot, therefore, comment on its current status and it would
be of great interest to know if this endemic is still extant.

Virachola antalus Hopffer 1855 (CC No. 1662; DA p. 515) (= Deu-
dorix antalus) is a species new to the island since the publication of
Vinson’s check list in 1938. David L. Hancock of the National Museum,
Bulawayo, Zimbabwe, informs us that Dr. E. C. G. Pinhey, formerly of
that museum, took one male and two females of this species in May of
1976 at Case Noyale, Relais de la mi Voie, and Riviere du Rempart. P.
M. H. and J. P. L. Davis took further specimens at Moka in 1978 and
later in the Black River Gorges and on the coast. The species would
hence appear to be well established on the island. The flight patterns
and other behaviour is similar to that of L. boeticus, and for this reason it
may have been on the island for some time and been confused with this
other species. The insect is present throughout Africa and in Mada-
gascar and the Comoro islands.

HE SPERHD AE

Coeliades forestan Stoll 1782 (CCNo.ll) ssp. forestall is distributed
throughout Africa, Reunion, Mauritius, Rodrigues, the Seychelles and
the Comoro islands as the nomotypical subspecies. Subspecies arbo-
gastes is present in Madagascar. The species is widely distributed in
Mauritius, but is never very common. It appears to prefer open ground.
The larval foodplants are Canavalia ensiformis. A. P. de Candolle, and
Terminalia catappa L. There is no evidence that the closely related C.
ramanatek Boisduval 1833 has ever occurred on the island, as has been
suggested.

Eagris sabadius Boisduval 1833 (CC No. 54) ssp. sabadius is distri-
buted throughout East and Southern Africa and the Indian Ocean
islands in a number of subspecies. The Mauritian subspecies is the
nomotypical one, also present in Reunion, and is not common on the
island. The insect has a very rapid and erratic flight and rests with its
wings spread flat. The larval foodplant is Hibiscus rosasinensis L.

Borbo borbonica Boisduval 1833 (CC No. 466) spp. borbonica. This
species is found throughout the Afrotropical region and southern
Europe. It is probably the commonest Hesperiid in Mauritius. The
ground colour is fuscous, with a row of angular yellowish spots running

\

156

J. Res . Lepid.

across the forewing. The sexes are similar. The larval foodplants are
various gramineae, especially Paniscum species.

Parnara naso Fabricius 1793 (CC No. 470) ssp. naso (= Parnara
marchalli Boisduval 1833) is widely distributed throughout Africa and
the Orient in a variety of subspecies. The nomotypical form is confined
to Mauritius. It is probably the second most common Hesperiid after the
preceding species, and is superficially similar in shape and coloration. It
may be distinguished by the reduced yellow forewing spotting in this
species, which may not even be present in the male. There is some
sexual dimorphism, the female having a larger fore wing spot and being
paler in coloration than the male, which is a rich chocolate brown. The
larval foodplant is sugar cane.

Erionota thrax L. 1767 ssp. thrax is an extremely large ‘skipper’ with
a forewing length of some 30 mm. The ‘Banana Skipper’ of the orient, it
appears to be a very recent introduction. It was first seen in Mauritius in
1970, and by 1972 its larva was such a common pest on bananas that the
Mauritius Ministry of Agriculture was forced to import three species of
hymenopterous parasitoids against it, from Sabah. The species sub-
sequently became uncommon. Three specimens were taken by J. P. L.
Davis at the Black River Aviary on the west coast in 1979, but it was not
seen elsewhere between the years 1976 and 1980. It current status is
unclear. This species is a notorious coloniser of tropical island habitats,
presumably being imported with agricultural produce.

Acknowledgements. The authors wish to thank the British Museum (Natural
History) for their considerable help and for the loan of specimens, and in
particular Dick Vane-Wright and Philip Ackery, and Cindy North. Our thanks
also go to the acting curator of the Hope Collections at Oxford, Malcolm J.
Scoble, for the further loan of specimens, and to Audrey Smith, Hope Librarian,
for her help in tracking down references. Our thanks to David L. Hancock of the
National Museum, Bulawayo, Zimbabwe, and E. C. G. Pinhey, formerly of that
museum, for information concerning the latter’s collecting trip to Mauritius;
and to the Herbarium of the Botany Dept, at Oxford University, especially Fred
Topliffe, for help in tracking down authorities for larval foodplants. Our thanks
also to Paul Brakefield of Exeter University for invaluable help in obtaining
specimens for photography. Last but not least, the thanks of the junior author to
the senior author and his wife Kathleen and son Julian for having introduced
him to this beautiful island and its butterfly fauna. His additional thanks to
Carl Jones of the Black River Aviary Project for a fascinating week’s collecting
in the Black River Gorges. Photographs by Martin Street, Honiton, Devon.

Literature Cited

CARCASSON, R. H., 1981. Handguide to the butterflies of Africa. Collins, London.
CORBET, A. S., 1948. Observations on the species of Rhopalocera common to

Madagascar and the Oriental region. Trans. Roy. Ent. Soc. Lond. 99, pt.

17:589-607.

30(3-4):145-161, 1991

157

D’ABRERA, A. B., 1980. Butterflies of the Afrotropical region. Landsdowne,
Melbourne.

INTERNATIONAL UNION FOR THE CONSERVATION OF NATURE. Papilionidae Red
Data Book. (1985).

KLOTS, A. B., 1929. The generic status of Catopsilia Hiibner and Phoebis Hiibner,
with a discussion of the relationships of the species and the homologies of the
male genitalia (Lepidoptera, Pieridae). Bull. Brooklyn Ent. Soc. 24(4):203-
214.

LE CERF, F., 1916. Note sur deux Pierides (Lep. Rhopal.) de Pile Maurice et de Pilot
d’Astove. Bull. Soc. Ent. France 1916:112-113.

MAMET, J. R., 1955. A revised food-plant catalogue of the insects of Mauritius.

Bull. Dept. Agric. Mauritius 90. 95 pp.

MANDERS, N., 1906. The variation of the larva and pupa of Vanessa hippomene
Hub. Entomologist 39:41-42.

, 1908. The Butterflies of Mauritius and Bourbon. Trans. Ent. Soc.

Lond. 1907(4):429-454.

PAULIAN, R., 1956. Faune de Madagascar II. Insectes Lepidopteres Danaidae,
Nymphalidae, Acraeidae. Tananarive, Inst. Rech. Sci. Madagascar. 102 pp.
2 col. pis., 110 figs.

&G. BERNARDI, 1951. Les Eurema de la region Malgache (Lep. Pieridae).

Naturaliste Malgache 3:139-154.

STEMPFFER, H., 1943. Contribution a l’etude des Lycaenidae de la faune Ethio-
pienne (Lepidopt. Rhopaloc.) Ann. Soc. Ent. de France. Vol. 91, 1942:117-
134.

TRIMEN, R., 1866. Notes on the butterflies of Mauritius. Trans. Ent. Soc. Lond.

Dec. 1866, JVol. 5, 3rd Ser., Pt. IV:329-344.

VINSON, J. M., 1938. A catalogue of the Lepidoptera of the Mascarene islands.
Mauritius Inst. Bull. 1. Dec. 1938:1-69.

Addendum

Since the submission of this manuscript for publication an important
new record for the island has been brought to our attention. R. I. Vane-
Wright, of the British Museum (Natural History), has informed us that
Dr. J. R. Williams of the Sugar Industry Research Institute of Mauritius
has taken two authenticated specimens of the species Danaus plexippus
at Curepipe, in the central uplands of the island, in April of 1984.
Furthermore, several specimens of this species are reported to have
been reared from larvae taken on an unknown Asclepiad by a junior
employee of the Ministry of Agriculture’s Entomology Division, in
November of 1983. It remains unknown as to whether or not the species
is now firmly established on the island, but if such is the case it
represents an important — and intriguing — extension to the range of
this widespread species. The species has not been recorded from the
mainland of Africa, the Indian Sub Continent, or, hitherto, and Indian
Ocean Islands. Its likely source of introduction must therefore lie in the
realms of conjecture.

158

J. Res. Lepid.

Further Addendum

One more new Mauritian species has recently been reported to us by Dr. J. R.
Williams. This is the Malagasy Hesperiid Ceoliades ernesti Grandidier 1867,
which was introduced into Reunion in the 1950’s. Dr. Williams has taken
several specimens of this species in the Mondrain area of the Black River
Gorges, and it appears to be well established. The most recent reports also
indicate that D. plexippus is now well-established and fairly common through-
out the island.

30(3-4):145-161, 1991

159

Fig. 1. 1 Dana us chrysippus chrysippus, 2. Henotesia narcissus narcissus ,

3. Amauris phaedon female, 4. Euploea euphon female, 5. Neptis
frobenia , 6. Hypolimnas dubius drucei female, 7. Hypolimnas
misippus male, 8. Junonia rhadama, 9. Vanessa cardui, 10. Hypo-
limnas misippus female form misippus, 11. Antanartia borbonica
mauritiana, 12. Junonia goudotii (Madagascar specimen), 13. Hypo-
limnas misippus female form dorippoides, 14. Phalanta phalantha
aethiopica, 15. Mel an ids leda helena

160

J. Res. Lepid.

Fig. 2.. 1. Pa pi Ho demodocus, 2. Pap i Ho man Hus female, 3. Sa/amis angu-
stina vinsoni female, 4. Catopsi/ia florella male, 5. Catopsi/ia florella
female form pyrene, 6. Catopsi/ia florella female form florella, 7.
Catopsi/ia thauruma male, 8. Catopsi/ia thauruma female form
thauruma, 9. Catopsi/ia thauruma female form grandidieri, 10.
Eurema brigitta pulchella male, 11. Eurema floricola ceres male

30(3-4): 145-161, 1991

161

Fig. 3. 1. Cacyreus darius female, 2. Cacyreus darius male, 3. Leptotes

pirithous male, 4. Leptotes pirithous female, 5. Lampides boeticus , 6.
Viracho/a antalus , 7. Zizina antanossa male, 8. Zizina antanossa
female, 9. Zizeeria knysna male, 10. Zizeeria knysna female, 11.
Cyc/yrius mandersi male, 12, Cyclyrius mandersi female, 13. Zizula
hy/ax male, 14. Zizula hylax female, 15. Eagris sabadius sabadius
male, 16. Eagris sabadius sabadius female, 17. Borbo borbonica
borbonica, 18. Parnara naso naso male, 19. Parnara naso naso female,
20. Coe/iades forestan forestall, 21. Erionota thrax thrax, 22. Zizina
antanossa (underside), 23. Zizula hylax (underside), 24. Zizeeria
knysna (underside)

Journal of Research on the Lepidoptera

30(3-4):162-174, 1991

Convergent Evolution in Western North American
and Patagonian Skippers (Hesperiidae)

Arthur M. Shapiro

Department of Zoology and Center for Population Biology, University of California,
Davis, CA 95616

Abstract. The western North American skippers Polites sabuleti and
Pyrgus scriptura, typically found in saline or alkaline semiarid habi-
tats, are pheno typically remarkably similar to Hylephila zapala and
Pyrgus seminigra from similar habitats in Argentine Patagonia. Their
behavior is effectively identical, and the two Pyrgus deviate behavior-
ally from their close relatives. The host plants of P. sabuleti and H.
zapala are identical; those of the two Pyrgus are subspecies or sister-
species. Placement in distinct genera makes P. sabuleti and H. zapala
unambiguously convergent. Phylogenetic relationships within Pyrgus
are unclear, but P. scriptura and P. seminigra appear to belong to
different, allopatric species-groups or lineages and to have evolved
their similarities by parallelism or convergence.

The shape of a crystal is determined solely by the molecular forces, and
it is not surprising that dissimilar substances should sometimes assume
the same form: but with organic beings we should bear in mind that the
form of each depends on an infinitude of complex reactions.. .It is
incredible that the descendants of two organisms, which had originally
differed in a marked manner, should ever afterward converge so closely
as to lead to a near approach to identity throughout their whole
organization.

— Charles Darwin, On the Origin of Species (1859)

Darwin objected not to the reality of convergent evolution, with which
he was thoroughly familiar, but to the idea that it could progress so far
that competent systematists would be unable to distinguish convergence
from homology (resemblance due to common ancestry). Yet debates of
this sort continue to enliven the zoological literature. Biochemical-
genetic methods (Sibley and Ahlquist 1987) have resolved some of them.
A good example of the problem is the striking similarity of much of the
high-Andean and Patagonian butterfly fauna to that of the temperate
and boreal Holarctic (Descimon 1986). Although historical-biogeographic
scenarios deriving the Andean fauna from the Holarctic have been
current for nearly a century, the matter is still unresolved for most taxa
(Shapiro 1991). Paradoxically, the greatest ambiguity lies at relatively
high taxonomic levels (genera and above). When individual species show
remarkable similarity but demonstrably belong to different lineages and
cannot be each other’s closest relatives, a diagnosis of convergence is
assured. This is especially true when the resemblance extends to behav-
ior and ecology, implying adaptive significance rather than the effects of

30(3-4):162-174, 1991

163

mere chance. Thus there is great uncertainty as to whether the Andean
Yramea is closely related to the Holarctic Boloria , but no doubt that the
uncanny phenotypic similarity of an undescribed oreal Yramea from
northwestern Argentina to the boreal Boloria improba Butler is conver-
gent.

Among such species-level convergences, a few spectacular cases have
been described which strain the biologist’s credulity. Cody ( 1974, pp. 165-
168) summarizes a number of avian examples, including the near-
legendary American meadowlark Sturnella (Icteridae) and African pipit
Macronyx (Motacillidae) which fooled Linnaeus, and which are still
virtually indistinguishable even as hand specimens (Friedmann 1946).
Using a picture book such as Lewis (1973), it is possible to synthesize
numerous cases of seemingly striking convergence in butterflies, but
without behavioral and ecological information they are not very credible.

In the course of field work in temperate North and South America, I
have twice been struck by apparent convergence equivalent to the
Sturnella-Macronyx case. Both instances involve Hesperiidae. Perhaps
unsurprisingly, the South American taxa are obscure and poorly docu-
mented. Both cases involve saline or alkaline, semiarid habitats. Al-
though the two North American species are frequently sympatric (espe-
cially in California), their Patagonian equivalents occur in similar
habitats but widely separated from each other. In both cases my initial
encounter with the Patagonian species elicited a shock of recognition —
equivalent to encountering an old friend in a very unlikely place.

Polites sabuleti and Hylephila zapala

Polites sabuleti Bdv. (type locality “California”), the Sandhill Skipper,
is fairly widespread in western North America (fig. 1), breaking into a
bewildering array of subspecies and local races in the Sierra Nevada and
Great Basin. The nominate, lowland Californian subspecies occurs
primarily in saline or alkaline seeps or on sand with a high water table,
where its native host, the perennial grass Distichlis spicata (L.) Greene,
is commonly the aspect dominant. Other consistent plant associates are
semisucculent Chenopodiaceae. The life history ofP. sabuleti was briefly
described by Comstock (1929).

The genus Polites is wholly Nearctic.

Hylephila zapala was described by Evans (1955, pp. 313-314) from a
pair collected at Zapala in the Province of Neuquen by P. Kohler and
forwarded by K. Hayward. Evans identifies these with Hayward’s figures
(1951, pi. IV and XVI) of Hylephila boulleti Mab. This same pair had
earlier been called H. peruana Draudt by Hayward (1934, pp. 111-112),
which he corrected to boulleti a few years later (1941, p. 281)(peruana is
considered a Peruvian subspecies of boulleti). In Hayward’s final cata-
logue (1973) he used the name zapala and gave its range as “Neuquen
(and) Chubut.” The true boulleti is also listed from far to the north in
Catamarca, following Evans. This usage is nearly correct. A member of

164

J. Res. Lepid.

Fig. 1 . Distribution of Polites sabuleti (all subspecies) in North America, from Scott
1 986, and of Hylephila zapala in South America.

Fig. 2. Distribution of Pyrgus scriptura in North America, from Scott, 1 986, and of P.
seminigra in South America.

the boulleti complex, probably undescribed, occurs at high elevation in
the Cumbres Calchaqmes and Sierra de Aconquija, a very high pre-
Andean range in Tucurnan and Catamarca. The rather dissimilar H.
isonira mima Evans, also listed by Hayward (1973) from Catamarca,
occurs below it in the same ranges (also in Salta and in the cordillera Real
in San Juan). I have not found any Chubut specimens of H. zapala , but
the habitat is common enough there.

The currently documented range of H. zapala is thus contained within
Neuquen, from Chos Malal (866m) in the far north to Zapala (1012m) in
the central part of the Province. It is locally common in marshy alkaline
seeps in high desert, localities inevitably dominated by Distichlis spicata
and various Chenopodiaceae. It does not occur in cooler, moister, less
saline inundated meadows ( mallines ) farther west, as at Loncopue
(892m), where it is replaced by the mesic-Patagonian and Chilean H.
signata Blanch, and H. fasciolata Blanch. H. zapala is at least double-
brooded (November and January).

Fig. 3A, B shows habitats of Californian P. sabuleti : 3C, D Patagonian
H. zapala.

Hesperiine skippers are often confusingly similar, but fig. 4A, B
demonstrates thatP. sabuleti and H. zapala transcend the ordinary. The
detailed resemblance extends to the individual spots forming the com-
plex pattern of the ventral hindwing beneath. These species are much
more similar in pattern than are most montane races ofP. sabuleti to the

30(3-4):162-174, 1991

165

lowland California phenotypes. Naive observers presented with the two
species have consistently noted the more prolonged apex of the Hylephila
forewing, as well as differences in the shape of the male stigma and
occasionally the shorter antenna in Hylephila.

No detailed description of the behavior of P. sahuleti has been pub-
lished. Scott (1986, p. 444) states that “Males perch all day in low grassy
spots to await females,” a more or less generic description of skipper
epigamic behavior. The mating behavior of Hylephila phyleus Drury has
been described thoroughly by I. Shapiro (1975, 1977). This, the only
North American Hylephila , sometimes co-occurs with P. sahuleti and is
useful as a comparison with it and its own congener, H. zapala. Although
H. phyleus will breed on Distichlis, it is neither restricted to it nor
particularly common in saline areas and alkali seeps. Unlike P. sahuleti ,
it is not effectively restricted to perching on Distichlis turf or nearby bare
soil. All three species are perchers, but H. phyleus males often select
roosts several cm above ground level, on projecting vegetation or litter;
H. zapala , like P. sahuleti , consistently perches on the ground (>100
observations of each). Both H. zapala and P. sahuleti oviposit singly on
the undersides of Distichlis blades, flying between bouts. H. zapala has
not been reared.

Pyrgus scriptura and P . seminigra

Pyrgus scriptura Bdv. (type locality “California”), the Least Checkered
Skipper, is indeed the smallest Pyrgine skipper in North America and
rivals the African and Middle Eastern species of Spialia for global
honors. Although its range is fairly extensive (fig. 2), throughout this
range it is local and rarely taken far from its host plant. Only two hosts,
both Malvaceae, are recorded: Sphaeralcea coccinea (Pursh) Rydb. in
Colorado (Ferris and Brown 1980) and Malvella leprosa ssp. hederacea
Torr. ( =Sida hederacea (Dougl.)Torr.) in California. These two plants are
not closely related but are very similar in habit, both being low, spread-
ing, perennial and pubescent. The life history of P. scriptura remains
unpublished, although it has been reared many times. Minno (1981)
studied its bionomics; Dusheck (1984), its host plant relations. Dusheck
describes its habitat in California as “the drying borders of alkaline
marshes in the Sacramento and San Joaquin Valleys, as well as vacant
lots and railroad tracks,” and of M. leprosa ssp. hederacea as “heavy
compacted... clay soils... more or less moist saline areas below 1800m.”
She found that females would not oviposit on other (non-pubescent)
mallows in the laboratory, but larvae were easily reared on them.
Whatever the eco-evolutionary basis for monophagy in the field, it is
neither toxicity nor nutritional inadequacy of the unused mallow species.

Pyrgus seminigra was described by Hayward (1933, p. 273 and pi.
XXIX, mistakenly cited in the text as pi. XXIV) as a “form” of what he
called Erynnis emma Stgr., differing from the typical in lacking the
marginal row of light spots and the white discal dot on the hindwing

166

above. The figured specimen indeed presents a solidly black hindwing
above. Second-brood specimens are darker than first, but nearly all I
have seen do have at least traces of the discal dot. The habitat is given as
Chubut and the south of the Province of Buenos Aires.

By 1941 Hayward called it Pyrgus bocchoris Hew. form seminigra , with
no additional data; likewise Hayward 1948. Evans (1953, p. 216) treats
seminigra as a junior synonym ofP. bocchoris ssp. cuzcona Draudt, type
locality Cusco, Peru. Like Hayward (1941), he treats emma as a junior
synonym of bocchoris. Hayward (1973) follows Evans on all points, listing
bocchoris from the northern, high -Andean Provinces of La Rioja,
Catamarca and Tucuman and seminigra from La Rioja, Tucuman and
Jujuy — without mentioning the southern records in the original descrip-
tion! Evidently his concept of his own taxon had changed.

MacNeill (pers. comm.) treats seminigra as a subspecies of cuzcona ,
which he separates from bocchoris. All of these entities are very similar
in facies and in genitalia. Some specimens of the common, weedy Chilean
subspecies of bocchoris , trisignatus Mab., are nearly indistinguishable
from Patagonian seminigra , though they average significantly larger. I
have elected to treat seminigra as a species in this paper, given the
uncertainties in its affinities and especially its geographic isolation in
northeastern Patagonia. This is an extraordinary disjunction from the
southern Peruvian highlands, not duplicated in any other organism
known to me. Moreover, temperate seasonality has been a very difficult
barrier for tropical high-Andean butterflies to overcome (Shapiro 1991).
It is also difficult to construct a scenario that would put seminigra where
it is and nominate bocchoris in the high Andes of northwestern Argen-
tina, between it and cuzcona , if these two are indeed more closely related
to each other than either is to bocchoris.

The ranges of Pyrgus scriptura and P. seminigra are given in fig. 2.

Seminigra is abundant in eastern Chubut (Trelew, Puerto Madryn,
Rawson, Uzcudun), north to the vicinity of Viedma and Carmen de
Patagones. It seems always to occur with the very widespread, but less
common Pyrgus fides Evans. (Typically, the taxonomy of this entity is
also confused. Hayward, 1933 first figured it as trisignatus and recorded
it from the Sierras of Cordoba (P. Kohler). In 1941 he corrected the name
to fides amd gave records from Cordoba, Mendoza, Neuquen, Rio Negro
and San Luis, repeating all of this in 1973. Evans (1953) correctly placed
trisignatus as the lowland Chilean subspecies of bocchoris : it is not
known from Argentina).

Pyrgus seminigra is restricted to alkaline seeps, mostly on compacted
clay soils, floristically and physiognomically equivalent to P. scriptura
habitats in California (fig. 3, A, Bus. E, F). It is confined to the immediate
vicinity of its only known host, Malvella (=. Sida)leprosa (Ortega) Krapov.
This plant is apparently identical to Californian hederacea except in
flower color (cream in hederacea , sulphur yellow in leprosa). It ranges
from “Mexico to Patagonia, in saline soils” (Cabrera 1953, p. 310), often

30(3-4):162-174, 1991

167

very widely disjunctly. These two plants, now placed in Malvella Jaub.
& Spach., have no very close relatives in the Americas. Like P. scriptura,
P. seminigra often occurs on waste ground, along railroad tracks, and in
very degraded or abused sites.

As in the Polites-Hylephila case, many checkered skippers are pheno-
typically similar, and the Pyrgus pattern is very conservative. However,
these two species (fig. 4C, D) are the smallest and darkest in their
respective faunas, and the same pattern elements tend toward obsoles-
cence in each. Their resemblance is not, however, especially close on the
ventral hindwing. What attracts attention in this pair, however, is not so
much their phenotypes as their behavior and ecology.

Scott (1986, p. 495) says of P. scriptura : “Males patrol and sometimes
perch all day in swales or gullies to seek females.” Minno ( 1981) contrasts
the behavior of scriptura with sympatric Pyrgus communis Grote: “ Pyrgus
communis... patrol by flying between six and twelve inches (15-30 cm) off
the ground, stopping frequently to feed and occasionally to bask. Pyrgus
scriptura assumes a somewhat different patrolling style in that males fly
closer to the ground and land frequently on vegetation or bare soil to
bask.” This comparison is quite accurate. P. scriptura seldom fly >10 cm
above the ground. They never engage in the nearly circular reconnoiter-
ing flights characteristic ofP. communis. While the flight of communis is
fast and direct, scriptura appears almost indolent when not alarmed,
flying with a zigzag, skipping motion which is reminiscent of a small
Bombyliid (Diptera). The flight ofP. scriptura is most exaggerated in the
small, late summer broods.

The flight and behavior of P. seminigra resemble those of scriptura
closely, contrasting analogously with the much stronger flight and
higher perching of sympatric fides. Males of the two Patagonian species
are easily distinguished, but ovipositing females may not be. Pyrgus fides
breeds on a variety of Malvaceae, including weedy naturalized Malva
ssp., erect annual Sida (in Neuquen and Mendoza), and M. leprosa.
Seminigra , like scriptura , appears strictly monophagous afield. Chilean
bocchoris are associated with weedy mallows, but remarkably the life
history is unpublished.

The eggs ofP. seminigra , like those ofP. scriptura , are laid singly on the
under surfaces of mature leaves. P. communis will lay eggs almost
anywhere on a host. P. fides is not well-known.

Although Emmel and Emmel (1973) remark that P. scriptura occurs
“sparingly” in the Colorado and Mojave deserts of southern California
and is “never taken in large numbers,” this is not true in the California
Central Valley. There, it is one of three abundant butterflies in its
unusual habitat, P. sabuleti and Brephidium exilis Bdv. (Lycaenidae)
being the others. In eastern Chubut, P. seminigra is likewise abundant.
The phenomenon of high population densities in species of low-diversity
faunas, so-called “density compensation,” was first described for islands
(MacArthur, Diamond and Karr 1972). In skippers it often results in

168

J. Res. Lepid.

suppression of territorial perching behavior, as described by Shapiro
1970, p. 120. Such is effectively the case here.

Discussion

Phylogenetic Considerations. Hylephila is an Andean genus, reaching
the highest altitudes of any Hesperiine skippers in the Americas. The
weedy lowland species H. phyleus probably entered North America in the
Great American Interchange or thereafter (i.e., within the past 3 MY),
and its range expanded to its climatic limits with the introduction of the
weedy grass Cynodon dactylon (L.) Pers. It now ranges from New York
City and northern California to extreme northern Patagonia. MacNeill
and Herrera are currently revising the genus. There is no basis available
to reconstruct the phylogenetic relationships among species, except to
say that H. zapala does not appear to be a primitive member of the genus,
especially close to the genus Polites, or the sister-species of P. sahuleti.
Ergo, its resemblance to P. sahuleti must constitute convergence.

Although Pyrgus is desperately in need of global revision and offers
unusually interesting opportunities for phylogenetic reconstruction, no
such project seems to be in progress. Early subdivisions of Pyrgus were
based on the male secondary sexual characteristics, which, however,
appear extraordinarily labile albeit not within species. On genitalic as
well as biogeographic grounds, P. seminigra belongs to an Andean cluster
of taxa. P. scriptura is somewhat aberrant but seems most closely related
not to the Andean Pyrgus but to P. ruralis Bdv. (male costal fold present)
and P. xanthus Edw. (absent, as in scriptura ). The reduced patterns of
both species are derivative in their lineages, and the derivative character
of P. scriptura is underscored by its seasonal dimorphism, the spring
phenotype being virtually identical to the single phenotype ofP. xanthus.
Thus, P. scriptura and P. seminigra cannot be sister-species.

Host Relations. I can find no record of the hosts of P. cuzcona or
nominate P. hocchoris (or of Chilean trisignatus , though it is known to eat
Malvaceae). There are plenty of mallows in the high-Andean flora up to
4500m (Halloy 1983 discusses an extreme case). The hosts of high-
Andean P. hocchoris are probably species of Nototriche or Maluastrum.
P. fides is also a mallow feeder. Thus all of the Argentine Andean and
Patagonian Pyrgus investigated so far are mallow feeders. In western
North America, both P. ruralis and P. xanthus feed on herbaceous
perennial Rosaceae. The Rosaceae-Malvaceae duality is pronounced in
the swarm of sibling species of this lineage in Europe, some of which feed
on one, some on the other, and one (P. maluae L.) perhaps on both
(Higgins and Riley 1970). Mallow feeding has presumably arisen several
times in the Holarctic Pyrgus , and is derivative within its lineage in P.
scriptura.

All Hylephila and Polites are presumably grass feeders. Distichlis
occurs naturally on both continents. It is the most common oviposition
substrate of all of the Andean and Patagonian Hylephila I have studied

30(3-4):162-174, 1991

169

from Arequipa, Peru south. It may well be the ancestral host of the genus.
In Polites, turfgrasses are not usual substrates; only sabuleti and its
desert races feed on Distichlis . The montane races feed mainly on
bunchgrasses such as Festuca idahoensis Elmer. One may thus infer that
P. sabuleti has diverged more from its relatives in its host relations than
has H. zapala.

“ Parallelism ” and Convergence . Traditionally, evolutionary biologists
and systematists have distinguished between these two terms, depend-
ing on the degree of relationship between the taxa. Simpson (1961)
defined parallelism as “the development of similar characters separately
in two or more lineages of common ancestry and on the basis of, or
channeled by, characteristics of that ancestry.5’ The repeated “discovery”
of Malvaceae in the Holarctic Pyrgus described above, would be a typical
case. Eldredge and Cracraft (1980) argue against the concept on grounds
of parsimony and intrinsic ambiguity; it is operationally unfeasible to
demonstrate a homologous genetic basis for allegedly parallel traits
except by hybridization experiments (which are ruled out in cases of full
speciation). Shapiro (1986) showed by such experiments that seasonal
polyphenism was developed in genetically non-homologous ways by two
members of the same polytypic species. If this may occur, it is dangerous
to make assumptions about genetic homology at all. I therefore treat the
resemblances between Pyrgus scriptura and P. seminigra as convergent.
Historical Biogeography . If we are not dealing with disjunctions between
sister-taxa, it is not necessary to “account” for the geographic relation-
ships shown in figs. 1 and 2. There are in fact floristic affinities between
western North America and the Patagonian region. The best-known are
those between the Sonoran desert and the Argentine monte (Solbrige7 al.
1977); the taxa described here do not properly pertain to these biota.
Pyrgus is almost certainly a Holarctic element in the Andean Fauna, but
the diversity of the genus in South America argues against it being as
recent there as the Great American Interchange.

Determinism and “ Adaptive Syndromes .” Both situations described
here represent apparently integrated syndromes or suites of characters,
including morphology, wing pattern, behavior, habitat and host plant. Is
any single factor the trigger for the development of such syndromes? The
two Pyrgus discussed here are very similar to the various species of the
Old World genus Spialia , found in arid and semiarid (occasionally mesic)

(Overlay) Fig. 3. Habitats of western North American and Patagonian skippers. A, B,
habitats of both Polites sabuleti and Pyrgus scriptura in spring, Solano
County, central California, with Distichlis spicata and Malvella leprosa ssp.
hederacea. V, 1 979. C, habitat of Hylephila zapala at Chos Malal, Neuquen.
1.1981. D, probable type locality of H. zapala, Zapala, Neuquen, 1.1981. E,
habitat of Pyrgus seminigra, Trelew, Chubut. The shrubs are naturalized
Tamarix. XII. 1989. F, same site as E showing large clone of Malvella
leprosa ssp. leprosa growing intermixed with Atriplex hastata. P. seminigra
was very abundant here. XII. 1989. All photos by AMS.

170

J. Res. Lepid.

30(3-4):162-174, 1991

171

Fig. 4. Dorsal and ventral surfaces of convergent skippers. A, Male and female Hylephila zapala, Chos Malal, Neuquen, Argentina, 1.1986. B,
male and female Polites sabuleti, West Sacramento, Yolo Co., Calif., VI. 1981 . C, Male Pyrgus seminigra, Trelew, Chubut, Argentina,
XII. 1989. D, Female Pyrgus scriptura, Suisun Marsh, Solano Co., Calif., IX. 1974. Photos by S.W. Woo.

172

J. Res. Lepid.

habitats in southern Europe, Africa, and the Near East. All species of
Spialia are confusingly similar among themselves, and indeed may be
indistinguishable on the wing. Their behavior, as described in various
regional works, is also very similar. However, their host plants are
taxonomically diverse, embracing Convolvulaceae, Sterculiaceae,
Malvaceae and Roseceae (Larsen 1974, Larsen and Larsen 1980). Thus
the identity of the host cannot be the determining factor (unless specia-
tion and host-plant specialization have occurred too recently for much
phenotypic differentiation to have occurred). Yet, as noted above, Pyrgus
scriptura uses two phenotypically similar, but not closely related, mal-
lows not used by other Pyrgus and growing in unusual habitats. Clearly
some kind of determinism is operating, but it is maddening when such
elusive selection factors seem able to produce remarkably precise, de-
tailed resemblances.

The most definitive study of convergence in a functional context is by
Mares (1980). Every biology student learns that widespread convergence
has occurred in the morphology of desert granivorous mammals, embrac-
ing not only various rodent groups but members of other lineages, even
marsupials, as well. Mares examined entire faunas using quantitative
morphometries and was able to demonstrate convincingly that morphol-
ogy was correlated with, and presumably functional in, feeding ecology.

The principal barrier to such sophisticated methods being brought to
bear on, say skippers is the lack of an ecological data base. The checkered
skippers appear ideally suited to such treatment, combined with a
rigorous phylogenetic analysis - but there is not even any anecdotal
literature on the biology of most of the species, and even common North
American and European species have never had their life-histories
published.

Cody (1974) points out that structurally simple habitats seem to
produce a high frequency of convergence. The occurrence of two striking
cases of convergence in different skipper subfamilies in precisely the
same, simple community in two hemispheres, as reported here, hardly
seems accidental. While not all alkaline seeps in temperate North and
South America have any skippers, there seem to be remarkably few ways
to be a skipper there at all.

Acknowledgments. Don MacNeill has determined all my South American
skippers for many years, but is not responsible for any errors of interpre-
tation. He and John Burns have been constant sources of wisdom on the
Hesperiidae, their genitalia and their evolution. Field studies reported
upon here were funded by NSF grants DEB-76- 186 11 and BSR-83-
06922. Specimen photos are by Samuel W. Woo. I thank June McCaskill,
UCD Herbarium for help with the botanical literature.

30(3-4): 162-174, 1991

173

Literature Cited

Cabrera, A.L. 1953. Manual de la Flora de los Alrededores de Buenos Aires.
Editorial Acme, Buenos Aires. 589 pp.

Cody, M.L. 1974. Competition and the Structure of Bird Communities. Princeton
University Press, N.J. 318 pp.

Comstock, J.A. 1929. Studies in Pacific Coast Lepidoptera (continued). Bull. So.
Calif. Acad. Sci. 28: 22-32.

Descimon, H. 1986. Origins of Lepidopteran faunas in the high tropical Andes,
pp. 500-532 in F. Vuilleumier & M. Monasterio, eds., High Altitude Tropical
Biogeography. Oxford University Press, New York.

Dusheck, J. 1984. Larval host suitability and oviposition preference in two
checkered skippers, Pyrgus communis and Pyrgus scriptura (Hesperiidae).
Unpublished M.A. thesis, University of California, Davis. 66 pp.

Eldredge, N., and J. Cracraft. 1980. Phylogenetic Patterns and the Evolutionary
Process, Columbia University Press, New York. 349 pp.

Emmel, T.C. and J.F. Emmel. 1973. The Butterflies of Southern California.

Natural History Museum of Los Angeles County, Calif. 148 pp.

Evans, W.H. 1953. A Catalogue of the American Hesperiidae in the British
Museum (Natural History). Part III. British Museum, London. 246 p.

1955. A Catalogue of the American Hesperiidae in the British Museum

(Natural History). Part IV. British Museum, London. 499 pp.

Ferris, C.D., and F.M. Brown. 1980. Butterflies of the Rocky Mountain States.

University of Oklahoma Press, Norman. 442 pp.

Friedmann, H. 1946. Ecological counterparts in birds. Sci. Monthly 63: 395-398.
Hallo y, S. 1983. Algunos datos ecologicos para Nototriche caesia Hill, Malvacea
altoandina, en las Cumbres Calchaqufes, Tucuman, Lilloa 36: 85-104.
Hayward, K.J. 1933. Lepidopteros argentinos. Familia Hesperiidae. III. Rev.
Soc. Ent. Arg. 5: 219-275.

1934. Lepidopteros argentinos. Familia Hesperiidae. IV. Rev. Soc. Ent.

Arg. 6: 97-233.

1941. Hesperiidarum Argentinae Catalogus. Rev. Mus. La Plata (N.S.)

Secc. Zook 2: 227-340.

1948. Genera et Species Animalium Argentinorum . Insecta, Lepidoptera.

I. Guillermo Kraft, Buenos Aires. 389 pp.

1951. Genera et Species Animalium Argentinorum. Insecta, Lepidoptera.

II. Guillermo Kraft, Buenos Aires. 388 pp.

1973. Catalogo de los Roploceros Argentinos. Opera Lilloana 23: 1-318.

Higgins, L.C., and N.D. Riley. 1970. A field Guide to the Butterflies of Britain
and Europe. Houghton Mifflin, Boston. 380 pp.

Larsen, T. 1974. Butterflies of Lebanon. Nat. Council for Sci Res., Beirut. 256

pp.

Larsen, T., and K. Larsen. 1980. Butterflies of Oman. Bartholomew Books,
Edinburgh. 80 pp.

Lewis, H.L. 1973. Butterflies of the World. Follett, Chicago. 312 pp.

Lindsey, A.W., E.L. Bell., and R.C. Williams, Jr. 1931. The Hesperioidae of
North American. Denison University Bull., J. Sci. Lab. 26: 1-142.
MacArthur, R.H., J.M. Diamond, and J.R. Karr. 1972. Density compensation in
island faunas. Ecology 53: 330-342.

Mares, M.A. 1980. Convergent evolution among desert rodents: a global
perspective. Bull. Carnegie Mus. Nat. Hist. 16: 1-51.

174

J. Res. Lepid.

Minno, M.C. 1981. The population biology of tropical and temperate butterflies
of the genus Pyrgus (Hesperiidae). Unpublished M.S. thesis, University of
California, Davis. 50 pp.

Scott, J.A. 1986. The Butterflies of North America. Stanford Univ. Press, Calif.
583 pp.

Shapiro, A.M. 1970. The biology of Poanes viator (Hesperiidae). J. Res. Lepid.
9: 109-123.

1986. The genetics of polyphenism and its role in phylogenetic

interpretation of the Tatochila sterodice species-group (Pieridae) in the
Andean-Neantarctic region. J. Res. Lepid. 24, suppl.: 24-31.

1991. The zoogeography and systematics of the Argentine Andean and

Patagonian Pierid fauna. J. Res. Lepid. 28: 137-238.

Shapiro, I. 1975. Courtship and mating behavior of the fiery skipper, Hylephila
phylaeus (Hesperiidae). J. Res. Lepid. 14: 125-141.

1977. Interaction of population biology and mating behavior of the fiery

skipper, Hylephila phylaeus (Hesperiidae). Amer. Midi. Nat. 98: 85-94.

Sibley, C.G., and J.E. Ahlquist. 1987. Avian phylogeny reconstructed from
comparisons of the genetic material, DNA. pp. 95-122 in C. Patterson, ed.,
Molecules and Morphology in Evolution: Conflict or Compromise? Cambridge
University Press, Cambridge.

Simpson, G.G. 1961. Principles of Animal Taxonomy. Columbia Univ. Press,
New York. 247 pp.

Solbrig, O.T., W.F. Blair, F.A. Enders et al. 1977. The biota: the dependent
variable, p. 50-66 in G.H. Orians and O.T. Solbrig, eds., Convergent Evolution
in Warm Deserts, Dowden, Hutchinson & Ross, Stroudsburg, Penna.

Journal of Research on the Lepidoptera

30(3-4):175-195, 1991

Multivariate and Phylogenetic Analyses of Larval
and Adult Characters of the Editha Complex of the
Genus Lycaena (Lepidoptera:Lycaenidae)

Gordon F. Pratt1, David M. Wright2, and Gregory R. Ballmer3

Abstract. The Editha Complex of the genus Lycaena , L. editha, L.
xanthoides, and L. rubidus, are united by biological and morphological
similarities which are not shared by other North American Lycaena.
The specific status of parapatric L. editha and L. xanthoides has been
questioned, since populations with intermediate wing characteristics
have been noted. Statistical analyses indicate significant larval char-
acter differences among populations of each species. Independent
discriminant analyses of adult, first instar, and mature larval charac-
ter sets of the three taxa, L. editha, L. x. xanthoides, and L. x. dione ,
gave congruent taxonomic classification with most populations. All but
one misclassification involved L. xanthoides, and these are believed
due to phylogenetic relationships of populations rather than ‘gene
flow.’

Phylogenetic analyses of 17 populations of the Editha Complex,
using 30 morphological characters and L. heteronea as an outgroup,
indicate that Lycaena rubidus ferrisi is probably the most primitive
taxon. Both L. editha and L. x. xanthoides appear to have evolved
independently from anL. x. dione ancestor, probably through vicariant
events caused by the glacial periods of the Pleistocene.

Introduction

Analyses using C14 dating of Pollen from core samples and macrofossils
of pack rat ( Neotoma ) middens from various regions of western North
America, have revealed great changes in plant biogeography during the
Pleistocene (Martin & Mehringer, 1965; Wells, 1983). For instance,
during the most recent Wisconsin glacial, subalpine conifers extended
down to basal elevations of the Great Basin; desert vegetation, which was
replaced by more mesic plants, moved into lowland pockets and to
southern desert refugia. Because of the complex topography of the West,
dispersal routes as adaptations to the climatic changes of the Pleistocene
have been proposed for a number of western birds, mammals, plants
(Wells, 1983) and butterflies (Porter & Shapiro, 1991; Austin & Murphy,
1987). These routes were based on island biogeography, the phyletic
relationships of neighboring populations, and the potential routes avail-
able from these regions.

The Pleistocene probably greatly affected the evolution and biogeogra-
phy of the Lycaena editha Complex, which consists of L. editha (Mead),
L. rubidus (Behr), and L. xanthoides (Boisduval), (Klots, 1936; Ballmer

1 Entomology and Applied Ecology Department, University of Delaware, Newark, DE
19716

2 100 Medical Campus Drive, Lansdale, PA 19446

3 Department of Entomology, University of California, Riverside, CA 92521

176

J. Res. Lepid.

& Pratt, 1989b), and primarily confined to western North America.
Lycaena xanthoides dione (Scudder), the only member of this group found
east of the Rocky Mountains, extends into the Great Plains. Although the
three species are regionally sympatric, they are confined to different
elevations. For example, along the eastern slopes of the Sierra Nevada,
in the vicinity of Sherwin Summit, L. xanthoides occurs mostly below
2,000 meters, L. editha generally above 3,000 meters, while L. rubidus
occurs between them from 2,000 to 3,000 meters elevation (Ballmer &
Pratt, 1989b; Ballmer & Pratt, pers. obs.). Certainly these elevation
limits, like those of the desert and subalpine vegetation, must have
changed drastically during the most recent Wisconsin glaciation.

A number of problems complicate phylogenetic studies of the Editha
Complex: 1) there is a paucity of good phylogenetic characters, 2) there
may be ‘gene flow’ between taxa creating either intermediate or ‘primi-
tive’ taxa, and 3) because most of the characters which separate these
taxa are continuous rather than discrete, there is a high probability of
convergence. Intermediate populations between L. editha and L.
xanthoides have been reported from Mather and Dunsmuir, California
(Scott, 1980). Four suspected hybrid specimens with L. rubidus have
been reported, one with L. x. dione and three with L. editha (Scott, 1980;
Anonymous, 1986; & Crowe, 1970).

Scott (1980) compared different populations of L. editha and L.
xanthoides using 15 wing characters. He found that only two of those 15
characters separated L. editha from L. xanthoides : wing length and
ventral hindwing spot size. The populations from Mather and Dunsmuir,
California were intermediate in these two characters. Scott (1980) also
determined that L. x . dione was the most distinct of the different
comparisons, since it differed from both L. editha and L. xanthoides on
the basis of eight wing characters. Unfortunately no statistics were used
to determine the significance of these characters.

The use of first instar setal patterns has been proposed for elucidating
the phylogenetic relationships of the Lycaenidae (Clark & Dickson,
1956). These characters differ greatly from those of later instars, since
first instars have only primary and subprimary setae, whereas in third
and fourth instars only secondary setae can be discriminated (Wright,
1983; Ballmer & Pratt, 1989b). Primary and subprimary setae are
distinguished from secondary setae by their numbers and positions,
which are fixed in specific locations on the larva; secondary setae are
variable in both number and location (Ballmer & Pratt, 1989b). Many of
the morphological structures found in mature Lycaena larvae are not
found in first instars, such as mushroom, dendritic, and other specialized
setae (Ballmer & Pratt, 1989b). Mature larval characters, representing
the opposite end from that of first instars of a spectrum of characters
expressed in the larval stage, are useful in distinguishing!/, editha from
L. xanthoides (Ballmer & Pratt, 1989b).

The two closest relatives to the Editha Complex are Lycaena heteronea
(Boisduval) and L. gorgon (Boisduval). In fact these two species are so

30(3-4):175-195, 1991

177

morphologically similar to the Editha Complex that Miller and Brown
(1979) misrepresented their phylogenetic relationships. They placed L.
rubidus with L. heteronea in the genus Chalceria, and L. editha and L.
xanthoides along with L. gorgon in Gaeides, even though male genitalic
structures indicate that L. gorgon and L. heteronea belong to one group,
while members of the Editha Complex comprise another species group
(Klots, 1936). Details of larval biology and morphology support this
genitalic arrangement of the species (Ballmer & Pratt, 1989b).

In this paper we employ discriminant and phylogenetic analyses of
larval and adult wing characters to investigate the status of the different
species of the Editha Complex, with emphasis on L. editha and L.
xanthoides. Lycaena heteronea was used as an outgroup comparison
because it shares the presence of dendritic setae on the mature larva with
that of the Editha Complex; these setae are lacking in all other North
American Coppers (Pratt, Ballmer, & Wright, pers. obs.). Based on
various phylogenies of the taxa at a population level, we hypothesize the
different dispersal routes and vicariant events which may have led to the
historical evolution of the different members of this complex.

Materials and Methods
OVIPOSITION AND REARING

Larvae were primarily reared from ova obtained from captive females; some
were field collected. Females of L. x. xanthoides from Mojave River Forks were
reared from larvae and mated in the lab (as in Ballmer and Pratt, 1989a); mated
females of other populations were field-collected. Dried leaves of Rumex acetosella
L., Rumex crispus L., and Rumex salicifolius Weinm. were used as oviposition
substrates. Egg diapause was terminated by refrigeration at 4°C for 3-5 months
(as in Pratt and Ballmer, 1986). Within 7-10 days after removal from 4°C and
incubation at 25°C, first instars eclosed and were reared on R. crispus until
mature. Mature larvae were preserved and injected, as in Ballmer and Pratt
( 1989b). First instar larvae were primarily obtained by dissection from ova, while
those that eclosed were fed R. crispus and distended and fixed in first instar
distention fluid (10 ml glacial acetic acid, 1 ml glycerin, 0.1 ml Triton X-100®, and
89 ml water).

FIRST INSTAR CHARACTERS

First instar characters (see Fig. 12 in Wright, 1983) are as follows: number of
SD1 and SD2 setae on prothoracic shield, of SD2 setae on T2-A7 and A9, of SD3
setae on T2, T3, and A3-6, of SD4 setae on A1-A7, of SD1 setae on A9, of lateral
setae and dorsal lenticles on A8; lengths of D1 and D2 setae on T2 and Al, of LI
and L4 setae on T2, L4 seta on T3, and L3 seta on Al; and mean crochet number
per proleg on A3-A6. SD3 and SD4 are derived subprimary setae common to
lycaenids. Frequently short and clubbed, SD3 is located near proprioceptor MD1
on thoracic and abdominal segments; SD4 is located near the spiracle on
abdominal segments. Permanent slide mounts of first instar cuticle were pre-
pared following brief KOH ( 10%) digestion of internal soft parts and cleansing of
cuticle. All characters were analyzed with a binocular microscope (200-400X),
and measurements were made with an ocular micrometer. The locations and
sample sizes are shown in Table 1 and Figure 1.

178

J. Res. Lepid.

Table 1 . Locations and Sample Sizes of the Different Populations for First Instar

Characters

Taxon

L. x.

Location*

xanthoides

Sacramento
Branscomb Lake
Mojave River Forks
Acton

Sherwin Summit
Silver Canyon

L. x. dione

Idledale

Milford

Brookfield

L. editha

Sagehen Mdws

Mt Barcroft

Tioga Pass

Winter Park
Dunsmuir

L. r. rub id us

Tioga Pass

L. r. ferrisi

Ditch Camp

L. heteronea

Chuchupate Cmpgd
Warner Valley

County

State

N

Sacramento

CA

9

Mendocino

CA

6

San Bernardino

CA

7

Los Angeles

CA

9

Mono

CA

8

Inyo

CA

10

Jefferson

CO

3

Seward

NB

12

Waukesha

Wl

6

Mono

CA

8

Mono

CA

7

Mono

CA

9

Grand

CO

6

Siskiyou

CA

12

Mono

CA

8

Apache

AZ

6

Ventura

CA

8

Lassen

CA

6

*The locations are shown in Figure 1 .

MATURE LARVAL CHARACTERS

Mature larval characters (see Fig. 1 in Ballmer and Pratt, 1989b) are as follows:
number of dendritic setae on T1 and A1-A8, number of non-sensory/non-mush-
room and mushroom setae on the prothoracic shield, head width, head and leg
pigmentation using L. xanthoides (Mojave River Forks) and L. editha (Tioga
Pass) as standards, length of the longest dorsal seta on Al, and mean lateral
crochet number on prolegs A3-A6. Measurements were made with an ocular
micrometer using a binocular dissecting microscope. The locations and sample
sizes are shown in Table 2 and Figure 1.

ADULT CHARACTERS

Adult characters (Figure 2) are as follows: length of the Cu2 vein in the
forewing and hindwing; medial width of the dorsal Cul aurora; lengths of the
dorsal M3 discal and Cu2 basal shadows of the forewing; longest length and
perpendicular width of the basal macule on the hindwing; longest lengths of the
Ml and M2 discal macules of the hindwing, and of the M3 and Cul discal macules
of the forewing. These measurements were made with an ocular micrometer
using a binocular dissecting scope. The locations and sample sizes are shown in
Figure 1 and Table 3.

DISCRIMINANT ANALYSIS

Discriminant Analyses (Hand, 1981) were performed onL. x. xanthoides, L . x.
dione, and L. editha using the SAS program. Three reference populations of the

30(3-4):175-195, 1991

179

Figure 1 Locations of the various populations of Lycaena : open circles = L
heteronea, closed circles = L. editha, closed triangles = L. x. xanthoides,
open triangles = L. x. dione, and open squares = L. rubidus. The locations
are as follows: 1=Mojave River Forks, 2=Acton, 3=Frazier Park, 4=Hunter
Mt, 5=Silver Canyon, 6=Mt Barcroft, 7=Sherwin Summit, 8=Sagehen
Meadows, 9=Tioga Pass, 1 0=Tioga Pass, 1 1 =Warren Canyon, 1 2=Sacra-
mento, 13^Branscomb Lake, 14=Mt Lassen, 15=Mt Bidwell, 16=Ball Mt,
1 7=Dunsmuir, 1 8=Ditch Camp, 1 9=Westcliffe, 20=Winter Park, 21 =ldledale,
22=Nebraska, and 23=Wisconsin.

taxa were used in these analyses, the remainder were classified (as test un-
knowns) as to their likely membership into one of these reference populations.
Some reference populations had to be lumped with neighboring populations,
since their sample sizes were considerably smaller than the number of charac-
ters. These reference populations are as follows: L. x. xanthoides from Sacra-
mento CA (mature larvae and adult female) and Sacramento and Branscomb
Lake (first instar); L. x. dione from Nebraska (mature larva), all of the adult
females (adult female), and Nebraska and Wisconsin (first instar); and L. editha
from Tioga Pass and surrounding areas (mature larva and adult female) and

180

J. Res. Lepid.

Table 2. Locations and Sample Sizes of the Different Populations for Mature

Larval Characters

Taxon

L. x.

Location*

xanthoides

Sacramento

Mojave River Forks
Hunter Mountain
Branscomb Lake
Silver Canyon
Sherwin Summit

L. x. dione

Milford

L. editha

Tioga Pass

Mount Barcroft
Dunsmuir

L. r. rubidus

Tioga Pass

L. heteronea

Mount Bidwell
Warren Canyon
Frazier Mountain
Westcliffe

County

State

N

Sacramento

CA

20

San Bernardino

CA

15

Inyo Co

CA

3

Mendocino

CA

2

Inyo

CA

6

Mono

CA

4

Seward

NB

20

Mono

CA

16

Mono

CA

14

Siskiyou

CA

11

Mono

CA

9

Modoc

CA

11

Mono

CA

4

Kern

CA

2

Custer

CO

15

* The locations are shown in Figure 1 .

Tioga Pass and Mt. Barcroft (first instar). F our characters were removed from the
first instar analysis: the number of the SD1 and SD2 setae on the pro thoracic
shield and the number of SD3 and SD2 setae on the metathorax and the ninth
abdominal segments, respectively. These setae are commonly absent on both L.
editha and L. x anthoides] their presence or absence is more important in
discriminating these taxa from L. rubidus and L. heteronea.

PHYLOGENETIC ANALYSIS

All of the morphological characters of the different populations were statisti-
cally compared using the general linear models procedure of SAS. Different
statistical classes were defined using a T-test (P<0.05) for each morphological
character for the different populations. Populations within each class were
significantly different from every other population outside of the class. The
number of classes were maximized and ordered according to decreasing magni-
tude of the means of the morphological character. The first class was given a value
of one, the next class a value of 2, etc. Two additional adult characters (iridescence
and color of the male dorsal wing surface) were added to the phylogenetic
analysis. These two characters were discrete and coded without statistics. The
iridescence was coded as present=2 or absent=l. The color gray was coded as 1,
orange as 2, and blue as 3.

These morphological scores were analyzed by PAUP (Phylogenetic Analysis
Using Parsimony, version 2.4, Swofford, Illinois Natural History Survey). Mulpars
and global branch swapping were performed with the FARRIS optimization.
Since the characters were continuous, the Weights Scale option was used.

30(3-4):175-195, 1991

181

Figure 2 A) Ventral wings of male Lycaena editha from Dunsmuir. Gul is cubital vein
1 , Ml is the medial vein 1 , M2 is the medial vein 2, M3 is the medial vein 3.
B) Ventral wings of female L. editha from Dunsmuir. C) Ventral wings of
female L editha from Mt Barcroft, White Mts. Cu2 is the cubital vein 2. D)
Dorsal wings of female L. editha from Dunsmuir.

Lycaena heteronea was used as an outgroup in most analyses, so in these cases
no prior assumptions were made about the polarity of character states (primitive
versus advanced). Those character states which were closest to that of L.
heteronea were therefore accorded primitive status by the analyses.

182

J, Res. Lepid.

Table 3. Locations and Sample Sizes of the Different Populations for Adult

Female Characters

Taxon

Location*

County

State

N

L. x. dione

Winnipeg

Manitoba

2

Omaha

NB

6

Blue Springs

Jackson

MO

2

Lees Summit

Jackson

MO

2

L. x. xanthoides

Sacramento

Stone Lakes

Sacramento

CA

9

area

West Sacramento

Yolo

CA

1

American R. Parkway

Sacramento

CA

10’

L. x. xanthoides

Southern

Lake Flenshaw

San Diego

CA

1

CA area

Mojave R. Forks

San Bernardino

CA

3

Solemint

Vista Grande

Los Angeles

CA

1

Ranger Station

Riverside

CA

1

Tahquitz Lodge

Riverside

CA

1

L. x.

xanthoides

Los Altos

Santa Clara

CA

4

L. x .

xanthoides

Sherwin Summit

Mono

CA

2

L x .

xanthoides

Hunter Mt

Inyo

CA

8

L. editha
editha

Tioga Pass

Mono

CA

17

Dunsmuir

Shasta

CA

12

* The locations are shown in Figure 1 .

Results

THE MORPHOLOGICAL CHARACTERS
The Editha Complex separates from Lycaena heteronea on the basis of
two first instar characters, the crochets and D2 length on Al. None of the
taxa of the Editha Complex separate as clearly from the remainder.
Lycaena rubidus is significantly different fromL. editha, L. x. xanthoides,
and L. x. dione (but not from L. heteronea ) on the basis of the prothoracic
shield SD1 and SD2 setae. Lycaena. x. xanthoides separates from the
other taxa on the basis of the D1 and D2 setal lengths on Al. Because of
variability, bothL. editha andL. x. dione are only significantly different
for one first instar character (L3 on Al). This character separates L. x.
dione fromL. x. xanthoides and L. editha , but not fromL. rubidus and L.
heteronea.

30(3-4):175-195, 1991

183

Three mature larval characters separate L. heteronea from members of
the Editha Complex: the number of dendritic setae on Al, the ratio of the
number of dendritic setae on A7 to A8 and the length of the dorsal setae
on Al. Within the Editha Complex, L. r. rubidus (Warren Canyon) is
distinct from all other populations in the number of dendritic setae on A3-
6 and non-sensory/non-mushroom (secondary) setae on the prothoracic
shield. Lycaena editha populations are distinct from most L. x. xanthoides
in having more dendritic setae on A2 and a darker head capsule and legs.

In addition to smaller wing size, two adult characters separate L.
editha from L. xanthoides : the ratio of the length of the hindwing M2
discal macule to the forewing Cul discal macule (>1 inL. editha but <1
inL. xanthoides) and the width of the dorsal orange aurora along Cul. Of
these two characters only the macule ratios yield significantly distinct
classes as discussed in the Materials and Methods. The mean ratio of the
M2 macule to the Cul macule is less than or approximately equal to one
in Lycaena gorgon , L. nivalis (Boisduval), L. heteronea , L. rubidus , L.
phlaeas (L.), andL. xanthoides (Pratt, pers. obs.). The mean ratios for L.
editha specimens from Tioga Pass and Dunsmuir were 1.33 and 1.54,
respectively, and were significantly different from the other Lycaena ,
and each other. In L. x. dione the orange aurora is wider than in three of
the five L. x. xanthoides populations, but is similar to that of the Hunter
Mt and Southern California populations.

In discriminant analyses all individuals of the reference populations
were classified in posterior tests to their proper taxon, showing that the
discriminant functions properly discriminate these three taxa. Mature
larval characters gave the best discrimination of the test unknowns. Out
of 55 unknowns, only two misclassifications were obtained, two L. x.
xanthoides larvae from Silver Canyon were misclassified as L. x. dione.
The first instar analysis misclassified 11 out of 63 unknowns. Two of
three L. x. dione , and two of eight and seven of ten L. x. xanthoides from
Sherwin Summit and Silver Canyon, respectively, were misclassified as
L. editha. Out of 33 adult female unknowns, nine were misclassified, one
of 12 L. editha from Dunsmuir was misclassified as L. x. xanthoides and
eight L. x. xanthoides were classified as L. x. dione (5 of 8 from Hunter Mt
and 3 of 7 from Southern California). Complete data of the morphological
characters can be obtained from the first author.

PHYLOGENETIC ANALYSIS

Twenty eight morphological characters formed significantly different
classes and were used in the phylogenetic analyses along with two
discrete adult wing characters. Some continuous characters did not form
significantly different classes as defined in the Materials and Methods,
although there were significant differences amongst the populations.

The phylogenetic trees produced by PAUP are based on the Wagner
Algorithm, which assumes that the best estimate of a phylogenetic tree
is that which has the fewest character changes (most parsimonious).

184

J. Res. Lepid.

First Instar Characters

— dione Wisconsin

- dione Nebraska

dione Colorado

xanthoides Sherwin

1— xanthoides Mojave

I xanthoides Acton

I — xanthoides Sacramento

c

xanthoides Branscomb

xanthoides Silver Cyn
— editha Sagehen
rubidus Warren Cyn

r - editha Winter Park
— editha Dunsmuir
editha Mt Barcroft
editha Tioga Pass

— rubidus Ditch Camp

, heteronea Lassen

heteronea Ventura

Figure 3 Phylogenetic Analysis (Distance Wagner Tree) of the members of the L.
editha complex using first instar characters and L. heteronea as an
outgroup. Length is 22.05 and the consistency index is 0.590.

Wagner trees show the interrelationships amongst the operational
taxonomic units (populations) better than IJPGMA derived trees, since
they do not assume that taxonomic units exhibit equal evolution through
time. In other words some populations can be more primitive
(plesiomorphic) than others. Not only are the interbranch lengths impor-
tant with these phylogenies, but so too are the branch lengths of the
individual taxonomic units, since they indicate how advanced
(apomorphic) the units are. From an analysis of nine taxonomic units,
slightly more than 4,000 different phylogenies or arrangements of the
units are possible. This number increases geometrically with the number
of taxonomic units being investigated. It is not surprising, therefore, that
there may be many phylogenies which are equally parsimonious in an
analysis of more than nine taxonomic units.

Three equally parsimonious trees were produced from the 14 first
instar characters, their topologies, or branching structures, were identi-
cal to Figure 3. All populations of L. x. xanthoides and L. x. dione
(Colorado) were advanced and clustered together, whereas L. r. ferrisi
(Ditch Camp) and L. editha (Tioga Pass) were primitive. Lycaena r.

30(3-4): 175-195, 1991

185

Mature Larval Characters

dione Nebraska

xanthoides Silver Cyn

editha Dunsmuir

editha Tioga Pass

— editha Mt Barcroft

rubidus Warren

xanthoides Sherwin
xanthoides Sacramento
xanthoides Branscomb L
xanthoides Mojave R

Cyn

L xanthoides Acton

r heteronea Tioga Pass
' heteronea Ball Mt

— heteronea Colorado

heteronea Ventura
heteronea Mt Bidwell

heteronea Warren Cyn

Figure 4 Phylogenetic Analysis (Distance Wagner Tree) of the members of the L.
editha complex using mature larval characters and L. heteronea as an
outgroup. Length is 15.65 and the consistency index is 0.767.

rubidus (Warren Canyon) was advanced compared to L. r. ferrisi (Ditch
Camp) with respect to first instar characters, and clustered amongst the
L. editha populations. The other populations ofL. x. dione and L. editha
were intermediate, yet did not cluster together. The Colorado L. x. dione
population appears to be distinct from the two other L. x. dione popula-
tions and is more closely related to L. x. xanthoides than is L. editha.
Lycaena editha (Sagehen Meadows) and L. xanthoides (Silver Canyon),
were the most advanced and most primitive of their species, respectively.

Twenty-one phylogenies were constructed from the 13 mature larval
characters. All showed L. xanthoides as most primitive, withL. editha as
intermediate, and L. rubidus as the most advanced species (Figure 4).
Lycaena x. dione clustered closely with L. x. xanthoides , but as more
primitive. The major differences amongst the 21 trees were the precise
arrangement among theL. xanthoides populations. The arrangement of
L. xanthoides Silver Canyon was the same with respect to L. editha and
L. rubidus , in all 21 trees.

These phylogenies also show an interesting relationship amongst the
L. heteronea populations. Those L. heteronea populations which exhibit

186

J. Res. Lepid.

Mature Larval Characters with Dark Legs and Head as Primitive

dione Nebraska

xanthoides Sherwin S

xanthoides Mojave R

xanthoides Acton

L xanthoides Branscomb

— xanthoides Sacramento

xanthoides Silver Cyn
— editha Dunsmuir

editha Tioga Pass

editha Mt Barcroft
rubidus Warren Cyn

- ANCESTOR

Figure 5 Phylogenetic Analysis (Distance Wagner Tree) of the members of the L.
editha complex using mature larval characters and a hypothetical ancestor
(characters of L heteronea, but with a dark head capsule and legs). Length
is 14.4 and the consistency index is 0.764.

well formed macules on the ventral hindwings (Tioga Pass, Ball Mt, and
Colorado) cluster separately from those with no visible macules on the
ventral hindwings. The Tioga Pass L. heteronea (on Eriogonum nudum
Doug, ex Benth.) was less than 1 km from the Warren Canyon population
ofL. heteronea (on Eriogonum umbellatum Torr.). The same two popula-
tions also differ in other morphological features and in ant attendance
(Ballmer and Pratt, 1992).

Since the phylogenies derived from mature larval characters differed
dramatically in the placement ofL. rubidus from those derived from first
instar characters, phylogenies were constructed with a hypothetical
ancestor modified from L. heteronea. This was to test whether the
polarization of the taxonomic units could be easily reversed by changes
in character states to those which could have been primitive for the
complex. The two partially correlated characters, the dark coloration of
the head capsule and legs, are quite variable amongst the Lycaenidae.
Populations of the same species can differ in these two characters
(Ballmer & Pratt, 1989b). A hypothetical ancestor (with a dark head
capsule and legs) causes the polarity of the taxa to be completely reversed
(Figure 5).

Since phylogenies can be affected easily by only a couple of changes in
the polarity of character states, a single convergence in a character state
could cause a reversal in the polarity of a phylogeny. The more characters
used in a phylogenetic analysis, the more robust the analysis will be and
the less likely that a single character will have this affect. Therefore a

30(3-4):175-195, 1991

187

All Characters
Without dione Colorado

dione Nebraska

xanthoides Silver Cyn

xanthoides Sherwin
xanthoides Mojave R
— xan Acton
xan Sacramento
L xanthoides Branscomb
editha Dunsmuir
r editha Tioga Pass
L editha Mt Barcroft
rubidus Warren Cyn

L

heteronea Ventura

rubidus Ditch Camp

Figure 6 Phylogenetic Analysis (Distance Wagner Tree) of the members of the L.
editha complex using first instar, mature larval, and adult characters and L.
heteronea from Ventura as an outgroup. Length is 43.726 and the consis-
tency index is 0.663.

phytogeny was constructed using all 30 morphological characters and
Lycaena heteronea (Ventura) as an outgroup. Since mature larval char-
acters were not available for L. r. ferrisi (Ditch Camp), these characters
were coded as missing. The L. x. dione Colorado population was not
included, since it was based on only three first instar larvae. Only one
Wagner Tree was produced from this analysis (Figure 6). Lycaena
rubidus ferrisi (Ditch Camp) was the most primitive taxon amongst the
Editha Complex. All of the L. x. xanthoides populations clustered to-
gether, as did allL. editha , and the twoL. rubidus populations clustered
at the base. This phylogeny suggests thatL. x. dione gave rise toL. editha
and L. x. xanthoides independently.

DISCUSSION

Since the number of characters used for the first instar and mature
larval derived phylogenies were small, 14 and 13 respectively, there are
a number of possible explanations of why the two phylogenies were quite
different. One of these, convergence of character states with the outgroup
was examined (Figure 5) and may be a partial explanation. Another
explanation is that different operational units were examined in the two
analyses. It would also seem likely that the mature larval characters
contained different evolutionary information from that of first instars
and that an analysis including all of the characters, would be a more
robust representation of the historical evolution of the complex.

188

J. Res. Lepid.

Some first instar characters may be correlated with mature larval
characters, such as lengths of primary and secondary setae. By combin-
ing these character sets an uneven weighting due to these correlated
characters would be obtained in a final analysis. However, with Euphilotes
Mattoni a correlation of setal lengths in first instars to mature larvae
appears unlikely, since all combinations in lengths have been observed
(Pratt, 1988). Euphilotes enoptes mojave (Watson & Comstock) first
instars have very short primary setae, while mature larvae have long
dorsal secondary setae. Euphilotes enoptes ancilla (Barnes &
McDunnough) is the reverse with long primary setae and short dorsal
secondary setae. Euphilotes enoptes dammersi (Comstock & Henne) has
both short primary and secondary setae, whereas E. battoides baueri
(Shields) has both long primary and secondary setae.

Within the Editha Complex there are three closely related entities
(. Lycaena rubidus , L. xanthoides, and L. editha). Recently, Scott (1980)
reduced L. editha to a subspecies of L. xanthoides. The Dunsmuir L.
editha population was proposed by Scott (1980) as intermediate between
these taxa. In this study we find that although the Dunsmuir population
is intermediate in wing size, it is not with larval and adult characters.
With discriminant analyses of these characters, the Dunsmuir popula-
tion clearly classifies as anL. editha. Also, a character which separates
all adult L. editha from other Lycaena (ratio of macule lengths) is not
intermediate in the Dunsmuir population.

Scott (1980) found only two morphological characters that distin-
guished L. editha from L. xanthoides ; these characters were the size of
the wings and the macules. He showed that the M2 macule size for the
Dunsmuir population is intermediate betweenL. editha andL. xanthoides.
This may be due to the way the macule was measured (parallel to a wing
vein), rather than its maximum length. Scott’s method is satisfactory
when the M2 macule is circular or when it is elongate and its long axis
is parallel to vein M2, as inL. editha from the White Mountains (Figure
2D). However, Scott’s method gives an erroneous value for macule width
in specimens with elongate macules whose long axes are not parallel to
the vein, as often found in L. editha from Dunsmuir (see Figure 2C).

From the phylogeny based on first and last instar larvae and adults
(Figure 6), both L. editha and L. x. xanthoides appear to have evolved
from L. x. dione. The depicted arrangement of populations within and
between taxa, may also reflect an historical pattern of evolution. For
instance, in the area where L. x. xanthoides shared the most recent
connection with extant L. x. dione populations, the latter could be more
similar to L. x. xanthoides than are other L. x. dione populations.
Therefore, considering the geographic distributions of the two taxa and
of the populations in this study (Figure 1), L. x. dione from Colorado (the
most southwestern) should be closest phenetically to L. x. xanthoides
from Silver Canyon, Sherwin Summit, and Hunter Mountain (the most
eastern). The L. x. dione population from Colorado is closest to L. x.

30(3-4):175-195, 1991

189

xanthoides from Sherwin Summit and Silver Canyon according to a first
instar phylogeny (Figure 3), which is the only complete data set for the
three dione populations. Discriminant analyses of mature larvae from
Sherwin Summit and adults from Hunter Mountain misclassify 2 of 6
larvae and 5 of 8 adults, respectively as L. x. dione. This supports the
hypothesis that L. x. xanthoides evolved from or was founded by L. x.
dione, rather than by L. editha.

The elevation differences between L. editha andL. xanthoides popula-
tions in the Sierra Nevada and Inyo Mountains generally make the two
taxa asynchronous in flight periods. Along the east slope of the Sierra
Nevada L. xanthoides occurs from about mid-May to mid- or late June at
2,000 meters, whereas L. editha generally doesn’t occur before early to
mid- July at 3,000 meters. Although it is possible that an early male of L.
editha may encounter a late female of L.x. xanthoides, by that season the
L. x. xanthoides female should have been mated.

Still ‘gene-flow’ may be occurring between L. editha and L. xanthoides
at Silver Canyon and perhaps Sherwin Summit, since according to a
discriminant analysis of first instars they are phenetically closer to L.
editha than are other L. x. xanthoides populations. It is also possible that
these relationships are due to retention of plesiomorphic characters,
since according to discriminant analyses two of three first instars of L.x.
dione from Colorado, also misclassify as L. editha. In any case, these
populations cluster more closely with L. xanthoides than with L. editha
(Figures 3-6).

Lycaena heteronea was chosen as an outgroup because it belongs to the
most closely related group (along with L. gorgon) to the Editha Complex,
and is the only other North American Lycaena known to possess dendritic
setae in later instars (Ballmer & Pratt, 1989b). The mature larval
characters which make L. xanthoides appear close to L. heteronea may be
due to retained plesiomorphic characters, convergent derived charac-
ters, or a combination of both (Figure 4). However, additional characters
of mature larvae shared by L. heteronea and L. gorgon are absent in the
Editha Complex and all other North American Lycaena (Ballmer & Pratt
1989b). These character states and those of first instar body and cranial
chaetotaxy (which are shared with other North American Coppers)
indicate that the Editha Complex is more plesiomorphic (older) than L.
heteronea (Wright pers obs.).

Although there is a possibility that L. xanthoides and L. editha may be
exhibiting some regional or site-specific gene-flow, all populations exam-
ined can be assigned to one or the other species in the complex. With
respect to larval characters, L. ruhidus appears to be at least as close to
L. editha as it is to L. xanthoides. There does not appear to be continuous
blending of populations ofL. xanthoides from the south or low elevations
with L. editha in the north or high elevations. Isolated populations of L.
editha from Tioga Pass (Sagehen Meadows, Winter Park, and Mt Barcroft)
cluster with other L. editha populations and isolated populations of L.

190

J. Res. Lepid.

xanthoides (Hunter Mt, Silver Canyon, and Sherwin Summit) cluster
with other L. xanthoides populations. It is possible that there are narrow
blend zones between the species, but at present these have not been
identified.

Lycaena ruhidus ferrisi is more primitive and significantly different
fromL. r. ruhidus (Warren Canyon) on the basis of seven of 14 first instar
characters used in the phylogenetic analysis. Perhaps these two taxa are
separate species as has been proposed by Johnson and Balogh (1977).
This distinction may also be due to either a sample bias or more likely to
the geographic isolation between the two taxa. More populations of L.
ruhidus need to be sampled in order to arrive at a species level decision
regarding these two taxa.

DISPERSAL ROUTES AND VICARIANT EVENTS OF THE EDITHA

COMPLEX

Lycaena ruhidus

Lycaena ruhidus ferrisi was shown to be the most primitive taxon in the
Editha Complex. This is particularly interesting since this taxon is
presently allopatric with all other N orth American Lycaena (Scott, 1986).
The formation of L. r. ferrisi therefore could have been the initial
vicariant event which separated the Editha complex from its sister taxon
and allowed it to evolve in isolation.

The dispersal routes open to L.r. ferrisi for the subsequent formation
ofL. ruhidus at Warren Canyon could have been long and precarious. The
species may have moved through the mountain ranges to the north of
where L. r. ferrisi occurs (WH in Figure 7) into Colorado, Wyoming,
Montana, and then westward to the Cascades of Washington and Oregon
and south into the Sierra Nevada. This movement could have resulted in
a series of founder events which caused the differentiation of L. ruhidus
at Warren Canyon. Since intermediate populations of L. ruhidus were
not used in the first instar phylogenetic analysis, it is not surprising that
the character steps of this population would be more parsimonious with
other derived populations of other taxa. This phylogenetic discrepancy
disappeared with a more robust analysis.

L. x. dione

The most likely method by which L. x. dione formed from L. r. ferrisi is
during a glacial period, perhaps the first of the four which occurred
during the Pleistocene. This would have caused L. r. ferrisi to move down
in elevation, perhaps to the Colorado Plateau (C in Figure 7). When the
glacier retreated, the Colorado Plateau, where no Lycaena presently
resides, probably became more xeric and an unfit habitat for the food
plants, Rumex sp. Perhaps some individuals retreated up the mountains,
whereas others at the eastern edge of the plateau went east into the Great
Plains (GP in Figure 7). At least this Colorado Plateau population would
have become adapted to Prairie habitats during the glacial period. Since

30(3-4):175-195, 1991

191

Figure 7. The map illustrates the present geographic regions of grasslands and
mountains of western North America. The dark areas represent mountain
ranges and the grasslands are outlined, while the remaining areas of the
West are largely desert. The following abbreviations: C=Colorado Plateau;
CR=Golorado Rocky Mountains; GB=Great Basin Desert; GP=Great Plains;
l=the edge of the Glacier during the Wisconsin Glacial period; IN=lnyo and
White Mountains of California; M=Mojave Desert; S=Sierra Nevada;
SV=Sacramento Valley and San Joaquin Valley; W=Wasatch Range;
WH=White Mountains of Arizona; XR=the present southern limit of L. x.
dione.

192

J. Res. Lepid.

the present southern limit of the range of L. x. dione (XR in Figure 7) is
north of the Colorado Plateau, L. x. dione must have moved north as the
glacier retreated, which may have further isolated L. rubidus from L. x.
dione.

L. editha

Two biological adaptations must have occurred in the ancestral L. x.
dione which formed L. editha : first it became adapted to high elevations
and/or cool environments; second, and perhaps most important, it be-
came adapted to Rumex paucifolius Nutt. These two changes probably
coincided, since R. paucifolius largely occurs in cool environments,
presently in alpine environments. BothL. rubidus and L. xanthoides use
large broad leaved Rumex species and not R. paucifolius (Ballmer &
Pratt, 1989b). The geographic range of L. editha is entirely confined
within the range of this food plant (Scott, 1986; Hitchcock & Cronquist,
1973; Scoggan, 1978). In areas where L. editha is not associated with R.
paucifolius , as at Dunsmuir, the butterfly primarily uses the closely
related species Rumex acetosella L. and/or R. angiocarpus Murb. These
three Rumex species share similar floral and other characteristics, and
key out in sequential couplets (Hitchcock & Cronquist, 1973). Both R.
acetosella and R. angiocarpus are believed to be introduced from Europe
(Munz, 1974). Lycaena cupreus (W. H. Edwards), a high altitude species
also adapted to R. paucifolius , like L. editha , sometimes occurs at lower
elevation in association with R. acetosella (Emmel & Pratt, pers. obs.).

A likely mechanism for these changes in a population of L. x. dione ,
involves entrapment during a glacial period in a blind prairie canyon or
valley. This canyon must have been long, broad, and ran generally north-
south. The blind end would have been blocked by mountains to the south,
with high mountains to the east and the west blocking movement in all
directions but the north. The closer to the southern edge of the glacier, the
colder the canyon would have become. Such a prairie canyon occurs in the
Rocky Mountains of Montana, the Red Rock River Canyon (Figure 7).

As the glacier moved south, probably L. x. dione moved south in
response. Eventually the glacier may have even cut off any possible
movement to the north, so that an isolated pocket of L. x. dione was
formed. In such an isolated pocket the large Rumex species, on which L.
x. dione was adapted, may have been gradually extirpated, being re-
placed by the colder adapted R. paucifolius. Therefore only those butter-
flies that went through a host shift to this new host survived.

The glacial period probably lasted a long time; for instance, the recent
Wisconsin lasted for over 50,000 years (Martin & Mehringer, 1965). This
is a lot of generations of selection upon a new host race. When the glacier
receded, the new host race probably followed its food plant up into the
Rocky Mountains, where it remained isolated from L. x. dione at high
elevations. Probably during the following glacial period L. editha dropped
down into the Great Basin desert (GB Figure 7), which was just south of

30(3-4): 175-195, 1991

193

the Rocky Mountains of Montana. During the Wisconsin Glacial, subal-
pine plants extended down to basal elevations (Wells, 1983). At that time
L. editha could have expanded its range throughout the Great Basin.
Now L. editha is found throughout connecting mountain ranges that
border the Great Basin desert.

L. x. xanthoides

Lycaena x. dione is adapted to the prairie habitat of the Great Plains
andL. x. xanthoides is adapted to grassland habitats of California (SV in
Figure 7). Although L. x. xanthoides does not occur exclusively in these
grassland habitats, its range encompasses the region. Only one other
Copper occurs in the grasslands of California, Lycaena helloides
(Boisduval), and it also inhabits the prairie habitat of the Great Plains
(Scott, 1986). It is possible, therefore, that L. x. xanthoides came from L.
jt. dione through a prairie corridor which led from the Great Plains to
California. Because of the Rocky Mountains and Great Basin desert of
the west, such a corridor probably would have had to have run from
western Texas through southern New Mexico and Arizona (south of the
White Mountains) and ended in the Mojave desert of California (WH and
M in Figure 7).

Present distributions of three lycaenid species suggest that such a
prairie corridor could have once existed, Phaeostrymon alcestis (W. H.
Edwards), Everes comyntas (Godart), and Lycaeides melissa (W. H.
Edwards) have populations adapted to the prairie habitats of the Great
Plains. Phaeostrymon alcestis occurs through much of the southern
regions of the Great Plains and its range overlaps that of L. x. dione. This
butterfly occurs in 2 disjunct regions through this region from the White
Mountains south (Scott, 1986). Everes comyntas also occurs throughout
the Great Plains and in disjunct areas south of the White Mountains.
This species extends its range into the grasslands of California and north
into the grasslands of Washington and is absent from the Rocky Moun-
tains and most of the Great Basin, and shows no likely east-west
connection to the north (Scott, 1986). Lycaeides melissa which occurs
through most of the west, occurs in this region south of the White
Mountains as disjunct populations, whereas the butterfly’s range out-
side of this region is largely continuous (Scott, 1986).

Although L.x. dione does not presently occur south to the gateway of
this hypothetical prairie corridor, the northern half of its range was
covered in glacier during the Wisconsin. It would seem likely that during
the Wisconsin, L. x. dione probably extended south to this gateway. Not
only would this prairie corridor have been cooler during a glacial period,
but moister and more favorable to Rumex species. When the glacier
retreated this corridor probably became too dry for Rumex species and
the range of L.x. dione moved north. Now L. x. dione and L. x. xanthoides
are isolated by the Rocky Mountains and Great Basin desert.

194

J. Res. Lepid.

Conclusion

One of the remarkable results of this study are the number of statistical
character differences found in different populations of the same species.
Perhaps this is not too surprising in L. rubidus and L. editha, which
historically may have formed isolated geographic populations during the
Pleistocene, but these differences are even apparent in the first instars
of different populations of L. x. dione. The Great Plains presently are
rather uniform, with little if any geographic barriers to gene-flow. On the
other hand, pockets of isolated prairie could have occurred during the
Wisconsin glaciation. For instance, the eastern peninsula of grassland
habitat of the Great Plains could have been isolated during a glaciated
period (see Figure 7).

By increasing the number of carefully chosen populations, number of
characters, and sample sizes, a more refined phylogenetic analysis could
be performed which could answer more specific dispersal questions
within the Editha Complex. Genitalic characters could also be examined
and scored in much the same way as the larval characters, particularly
since character differences have been observed in the L. rubidus subspe-
cies (Johnson & Balogh, 1977). By increasing the sample sizes for some
of the larval characters, greater statistical differences may be obtained
which would create more characters that form significantly different
classes. Such analyses may demonstrate vicariant events which could be
timed by a known geological history of the regions. These analyses may
also help illuminate potential vicariant events of other butterflies as well
as other insects.

Acknowledgements. We thank Susan S. Borkin for ova of L. x. dione from
Wisconsin, Jim Scott for ova of L. editha and L. x. dione from Colorado, Malcolm
Douglas for ova of L. r. ferrisi from Arizona, Art Shapiro for females of L. x.
xanthoides from Sacramento and ova ofL. heteronea from Ball Mt (CA), Sterling
Matoon for the ova of L. xanthoides from Branscomb Lake (CA), Rudy Mattoni
for ova of L. xanthoides from Acton and L. heteronea from Chuchupate Camp-
ground (CA), Adam Porter for ova of L. xanthoides from Silver Canyon (CA).
Additional thanks are due to Adam Porter, Art Shapiro, and John Emmel for
reading and commenting on an early version and two anonymous reviewers
which helped in the final manuscript.

Literature Cited

Anonymous, 1986. Season Summary 1985, Colorado Section. News Lep. Soc. 2: 23.
Austin, G. T. & D. D. Murphy, 1987. Zoogeography of Great Basin butterflies:
Patterns of distribution and differentiation. Great Basin Naturalist, 47:186-
201.

Ballmer, G. R. & G. F. Pratt, 1989a. Instar number and larval development in
Lycaena phlaeas hypophlaeas (Boisduval) (Lepidoptera:Lycaenidae). J. Lep.
Soc. 43:59-65.

, 1989b. A survey of the last instar larvae of the Lycaenidae (Lepidoptera)

of California. J. Res. Lepid. 27: 1-80.

30(3-4): 175- 195, 1991

195

. 1992. Quantification of Ant Attendance of Lycaenid Larvae. J. Res. Lepid.

30:95-112.

Clark, G. C. & C. G. C. Dickson, 1956. Proposed classification of South African
Lycaenidae from the early stages. J. Ent. Soc. South Africa 10,195-215.

Crowe, C. R., 1970. A possible new hybrid copper. J. Res. Lep. 8: 51-52.

Hand, D. J. 1981. Discrimination and Classification, New York: John Wiley &
Sons, Inc.

Hitchcock, C. L. & A. Cronquist. 1973. Flora of the Pacific Northwest. University
of Washington Press, Seattle Washington, pp. 90-92.

Johnson, K. & G. Balogh, 1977. Studies in the genus Lycaena. 2 Taxonomy and
evolution of the Neararctic Lycaena rubidus complex, with description of a
new species. Bull. Allyn Museum 43: 1-62.

Klots, A. B., 1936. The interrelationships of the species of the genus Lycaena
Fabricius (Lepidoptera, Lycaenidae). Bull. Brooklyn Ent. Soc. 31: 154-171.

Martin, P. S.,& P. J. Mehringer, Jr. 1965. Pleistocene pollen analysis and
biogeography of the Southwest. Pages 433-45 1 in H. E. Wright and D. G. Frey,
editors. The Quaternary of the United States. Princeton University Press,
Princeton, New Jersey, USA.

Miller, L. D., & F. M. Brown, 1979. Studies in the Lycaeninae (Lycaenidae) 4. The
higher classification of the American coppers. Bull. Allyn Mus. 51: 1-30.

Munz, P. A. 1974. A Flora of Southern California, University of California Press,
Berkeley, CA.

Porter, A. H. & A. M. Shapiro, 1991. Genetics and biogeography of the Oeneis
chryxus Complex in California. The Journal of Research on the Lepidoptera.
28:263-276.

Pratt, G. F. 1988. The Evolution and Biology of Euphilotes Biotypes. Unpublished
doctoral dissertation, University of California, Riverside, 653 pp.

Scoggan, H. J. 1978. The Flora of Canada, Bonnie Livingston, editor. National
Museums of Canada, publisher, Ottawa, Canada, p. 639.

Scott, J. A., 1980. Geographic variation in Lycaena xanthoides. J. Res. Lep. 18:
50-59.

1986. The Butterflies of North America. Stanford University Press,

Stanford, California.

Wells, P. V. 1983. Paleobiogeography of montane islands in the Great Basin
since the last Glaciopluvial. Ecological Monographs, 53:341-383.

Wright, D. M., 1983. Life history and morphology of the immature stages of the
bog copper butterfly Lycaena epixanthe (Bsd. & Le C.) (Lepidoptera:
Lycaenidae). J. Res. Lep. 22: 47-100.

Journal of Research on the Lepidoptera

30(3-4):196-208, 1991

A morphological search for the sound mechanism of
Hamadryas butterflies (Lepidoptera: Nymphalidae)

Julian Monge-Najera

Museo de Zoologia, Universidad de Costa Rica, Costa Rica

Francisco Hernandez

Facultad de Microbiologla, Universidad de Costa Rica, Costa Rica

Abstract. At least seven locations have been proposed for the sonic
mechanism of Hamadryas butterflies, which has not been identified to
date. Using light microscopy, five species of Hamadryas (including a
“mute” population) were compared with Siproeta stelenes, Anartia
fatima and Ectima thecla, which do not emit audible sound. Three
macrostructures were found exclusively in Hamadryas: the abdominal
rami, and in the forewings, a hemispheric membrane in the costal cell
and the swollen base of the subcostal vein. Only the third appears to be
related to sound emission (possibly, percussion by clapping the wings).
Thus, the fore wings appear to be the most feasible location for the
sound emission mechanism in Hamadryas.

Introduction

Despite Darwin’s mention of the phenomenon in his books (Darwin,
1839; 1871), the loud sound emissions of Hamadryas — a genus of
Neotropical butterflies — are not widely known by entomologists. At least
seven locations of the thorax, forewings and abdomen have been pro-
posed for the still unidentified sound mechanism, but experimentation
has proven difficult (Swihart, 1967; Monge-Najera, 1991).

Considering the intensity and frequency of the sound, the structures
which produce it must be relatively large (Swihart, 1967; Cromer, 1978)
and possibly detectable in a study of the external gross morphology.
Surprisingly, such a study had never been attempted before.

This paper compares external structure in eight ecologically related
Neotropical species: I) five sound-emitting Hamadryas species from
Costa Rica: H. amphinome, H. feronia, H. guatemalena, H. glauconome
and H. februa\ specimens of H. februa from a Venezuelan population
which does not emit audible sound (Otero, 1988) were also included; II)
three species which do not emit sound: a) Siproeta stelenes , similar to
Hamadryas in size, territorial behavior and diet; b) Anartia fatima , also
territorial but smaller and nectarivorous and c) Ectima thecla , a member
of a genus that is phylogenetically and ethologically close to Hamadryas
(Jenkins, 1985).

30(3-4):196-208, 1991

197

Methods

The insects were fixed in KAAD for at least a week and preserved in 75%
ethanol. Scales and debris were eliminated with a brush and by manual or
ultrasonic agitation while submerging the body (6 min) and the wings (3 min) in
commercial bleach (approximately 2.5 % sodium hypochlorite). After a rinse in
distilled water, the specimens were dried with the help of a vacuum evaporator
or in air.

All illustrations are based on camera lucida drawings. To facilitate observa-
tion, some translucent parts were coated with gold as for scanning electron
microscopy and observed with a light dissection microscope. Ink was added to
check possible communication among veins.

A total of 97 specimens of eight species were studied, as follows: Hamadryas
februa(5 9,30 6),H. feroniai 2 9,5 6),H. guatemalena (1 9,8 8),H. glauconome
(1 9,1 8),H. amphinome (1 9,1 8 ), S. stelenes (2 9 , 24 6 ), A. fatima (3 9,11 8)
and E. thecla (2d).

Results

The structure of forewings and the general thorax and abdomen plans
are similar in all taxa, independent of sex (Figs. 1-9). The only outstand-
ing characteristics are the presence of the hypandrium and rami, in male
E. thecla and Hamadryas spp., respectively (Figs. 7 C, 8 A,B and 9 A,B),
and three structures of the forewings: Vogel’s organ, and in Hamadryas,
the costal cell membrane and the swollen base of the subcostal vein (Figs.
2 and 7 B).

The hypandrium and rami are structures associated with the male
genitalia. The hypandrium has the shape of a curved lamina and the rami
are a pair of rods (Figs. 7 C, 8 A,B and 9 A,B).

“Vogel’s organ” (a term coined by Otero, 1988) is developed in both sexes
of Hamadryas and less defined in the other genera. This organ (Fig. 7 B)
was found in satyrids early in the century (Vogel, 1912) and occupies the
base of the Cu vein. It has a rigid cap mounted on a flexible ring. There
are four chambers under the cap; ink added to them does not reach the
adjoining veins, suggesting a lack of direct connection.

The costal cell membrane (present in Hamadryas only) has the shape
of an elongated coppola and is located in the wing base (Fig. 7 B). This
membrane, which is inflated in the living insect, can be easily ruptured
(and thus deinflated) during manipulation of the specimen. In Hamadr-
yas, the strong subcostal vein is highly “swollen” and reaches a diameter
about 2-3 times that of the equivalent vein of silent species.

All Hamadryas have three structures that are lacking in the species
which do not emit sound: a) in the abdomen, the rami and b) in the
forewings, the costal cell membrane and the swollen base of the subcostal
vein.

Discussion

To evaluate the plausibility of the role of each structure in sonic
emission, it is useful to consider how sound is produced by organisms.

198

J. Res . Lepid.

parapatagium
patagium
p/eurosternum
epistermun
coxa i
epimeron g
mer on 2 -
eucoxa

qnepisternum 2
Dasalare z
s ubalare 2
axilore cord 2
scutellum 2
scutum 3
scutellum 3
subalare 3
epimeron 3
eucoxa 3
epimeron 3
episkternum 3

Fig. 1 . Lateral view of thorax. A. Nomenclature (from Scott, 1985). B. Male E. thecla.
C. Male E. fatima. D. Female E. fatima (wings removed). E. Male S.
stelenes. F. Female S. stelenes . Bars 1 mm.

30(3-4): 196-208, 1991

199

Fig. 2. Lateral view of thorax. A. Male H. februa. B. Female H. februa. C. Male H.
feronia. D. Female H. feronia. E. Male H. guatemalena. F. Female H.
guatemalena. Bars 1mm.

200

J. Res. Lepid.

Fig. 3. Dorsal view of thorax. A. Nomenclature (from Scott, 1985). B. Male E. thecla.
C. Male E. fatima. D. Female E. fatima (wings removed). E. Male S.
stelenes. F. Female S. stelenes. Bars 1 mm.

30(3-4): 196-208, 1991

201

Fig. 4. Dorsal view of thorax. A. Male H. februa. B. Female H. februa. C. Male H.
feronia. D. Female H. feronia. E. Male H. guatemalena. F. Female H.
guatemalena. Bars 1mm.

202

J. Res. Lepid.

Fig. 5. Ventral view of thorax. A. Nomenclature (from Scott, 1985). B. Male E. thecla.
C. Male E. fatima. D. Female E. fatima (wings removed). E. Male S.
stelenes. F. Female S. stelenes. Bars 1 mm.

30(3-4): 196-208, 1991

203

Fig. 6. Ventral view of thorax. A. Male H. februa. B. Female H. februa. C. Male H.
feronia. D. Female H. feronia. E. Male H. guatemalena. F. Female H.
guatemalena. Bars 1mm.

204

J. Res. Lepid.

Fig. 7. Dorsal (A) and ventral (B) sides of the forewing base of male H. februa. Veins:
Sc subcostal, R radial, Cu cubital, A anal. Other structures: BA basalare, H
hypandrium, MCC membrane of costal cell, SA subalare, VO Vogel’s
Organ. From above: dorsal, lateral and ventral view of abdomen in male E.
thecla (C), and A. fatima (D: male, E: female). Bars 1 mm.

30(3-4): 196-208, 1991

205

Fig. 8. From above: dorsal, lateral and ventral view of abdomen in: A. Male S.
stelenes. B. Male H. februa. C. Female S. stelenes. D. Female H. februa.
Bars 1 mm.

206

J. Res. Lepid.

Fig. 9. From above, dorsal, lateral and ventral view of abdomen in: A. Male H. feronia.
B. Male H. guatemalena. C. Female H. feronia. D. Female H. guatemaiena .

Bars 1 mm.

30(3-4): 196-208, 1991

207

There are three basic mechanisms, all known in the Lepidoptera (Monge-
Najera and Morera, 1987): vibration of a filament, vibration of a mem-
brane and percussion.

None of the structures found appears capable of vibrating as a filament.
The costal cell membrane and Vogel’s organ could act as a vibrating
membrane but they lack the necessary muscle and the second also occurs
in “silent” species (Figs. 1 and 2). Swihart (1967) experimentally showed
the auditory function of the costal cell membrane; the function of Vogel’s
organ, proposed to be a hearing organ by Vogel himself (1912), has not
been tested (it may be specialized for detecting predatory bats, as
suggested by its rigidity and smaller size; see Cromer, 1978).

Percussion could be carried out by any mobile structure hitting against
another; this includes the antennae and legs (both never suggested as
sound organs in the literature on Hamadryas ) and the rami and wings.
The antennae appear too soft for loud percussion and the legs do not
participate, since -very infrequently- perching individuals produce sound
while the legs are motionless (but they clap the wings as they emit sound
during perching; JMN, personal observation). The rami do not show
articulations (Jenkins, 1983) and are more probably structures used by
females “to evaluate the male” during copulation attempts, as suggested
in general for complex, rigid sexual structures by Eberhard (1985). The
same applies to the hypandrium of the mute Ectima.

In contrast with the other structures, the swollen base of the subcostal
vein in the forewings (exclusively present in species that emit sound)
could be a reinforcement for percussion. A strong venation may allow
sound production when the insect claps its wings (less probably, sound
may result from chitin flexing or snapping during a modified wing beat).
Interestingly, the “mute” Venezuelan Hamadryas do not present any
defined morphological difference from sound emitting Hamadryas. Their
silence may be an ethological rather than a morphological characteristic.

In conclusion, the forewings are the most feasible location for the sound
emission mechanism in Hamadryas, possibly in relation to clapping, but
the exact nature of the mechanism remains to be identified.

Acknowledgments. We thank many colleagues for assistance and advice, par-
ticularly Luis Otero (Universidad Central de Venezuela) and Peter Dobbeler
(Berlin Herbarium), who provided specimens and a translation, respectively.
Carlos Valerio, Juan Chavarria and William Eberhard (Universidad de Costa
Rica) gave valuable advice and commented on the manuscript.

Literature Cited

Cromer, A.H. 1978. Ffsica para las ciencias de la vida. Reverte, Barcelona,
Spain. 541 p.

Darwin, C.R. 1839. A naturalist voyage. Murray, London.

. 1871. The descent of man and selection in relation to sex. Murray, London.

Eberhard, W.G. 1985. Sexual selection and animal genitalia. Harvard University
Press, Cambridge. 244 p.

208

J. Res . Lepid.

Jenkins, D.W. 1983. Neotropical Nymphalidae I. Revision of Hamadryas. Bull.
Allyn Mus. 81: 1-146.

. 1985. Neotropical Nymphalidae. IV. Revision of Ectima. Bull. Allyn Mus.

95: 1-31.

Monge -NAjera, J. 1991. Cryptic Butterflies, Hamadryas of Panama (Lepidoptera,
Nymphalidae): Their biology and classification. In D. Quintero and A. Aiello
(editors). Insects of Mesoamerica and Panama: Selected Studies. Oxford
University Press, Oxford.

Monge-Najera, J., & B. Morera-Brenes. 1987. Emision de sonido en insectos: un
enfoque evolutivo. Biocenosis 3: 18-21.

Otero, L.D. 1988. Contribution a la historia natural del genero Hamadryas
(Lepidoptera: Nymphalidae). Ph.D. Thesis, Universidad Central de
Venezuela, Maracay, Venezuela.

Scott, 1985. The phylogeny of Butterflies (Papilionoidea and Hesperoidea). J.
Res. Lepid. 23 (4): 241-281.

Swihart, S.L. 1967. Hearing in butterflies (Nymphalidae: Heliconius,Ageronia).
J. Insect Physiol. 13: 469-476.

Vogel, R. 1912. Uber die Chordotonalorgane in der Wurzel der
Schmetterlingsflugel. Z. Wiss. Zool. 100: 210-244.

Journal of Research on the Lepidoptera

30(3-4):209-220, 1991

Extirpation and Recolonization of the Buckeye,
Junonia coenia (Nymphalidae) Following the
Northern California Freeze of December, 1990

Arthur M. Shapiro

Department of Zoology and Center for Population Biology, University of California,
Davis, CA 95616

Abstract. The Buckeye, Junonia coenia, seems to have been eradi-
cated from its northern California range east of the San Francisco Bay
area by a severe freeze the fourth week of December 1990. Reinvasion
of the Sacramento Valley began at the end of June and many areas had
been reoccupied by the end of the 1991 season. Population growth was
approximately exponential during the reoccupation, but at season’s
end numbers were estimated at only 10% of average. No colonization
was observed in the montane Sierra Nevada.

Introduction

In 1972 Ehrlich et al. observed that "... we believe that extinctions of
local populations may be relatively commonplace in most temperate
butterfly species, if not in most temperate animal species. That they are
not more commonly observed and reported may well be an artifact of
misapprehensions about the size of evolutionary and population-dy-
namic units... as well as the difficulty of ‘proving’ that a population of
small, vagile organisms is no longer maintaining itself in an area.” Since
that time, extinction processes affecting threatened or endangered
species have become an important topic of concern in ecology and
conservation biology. Such butterflies are typically monophagous, steno-
topic, univoltine and philopatric (Arnold 1981, 1983). Extinctions of
common, “weedy” species are rarely noted, although in theory they
should be very common. The Buckeye, Junonia (or Precis ) coenia Hbn.
(Nymphalidae), is about as different from an endangered Lycaenid as a
butterfly can be; it is oligophagous, eurytopic, multivoltine and disper-
sive or even migratory. Such a species may be thought of as having no
“permanent” populations at all, only constantly dynamic local manifes-
tations of a large-scale or even global “metapopulation” (Andrewartha
and Birch 1954, Levins 1970). Buckeyes may breed in a given locality for
only one generation before dispersing, or being forced to disperse by
seasonal changes in the ruderal vegetation. The larger entity — the
metapopulation — may, however, be very persistent. In lowland Califor-
nia J. coenia is regarded as ubiquitous, common, and weedy (Garth and
Tilden 1986; Tilden and Smith 1986).

The droughty winter of 1989-90 allowed the Buckeye to overwinter
successfully in many atypical localities, as evidenced by the flight of
adults in late winter and early spring where this is a rare event. In the
California Central Valley the regular pattern of seasonal population
growth was advanced by 1-2 generations, depending on locality. By

210

J. Res. Lepid.

autumn 1990 populations were very large in most areas, and adults were
still flying as late as early December at the Suisun Marsh. Within three
weeks the coldest weather in several decades gripped California. Al-
though it cannot be rigorously proven, this freezing episode seems to have
destroyed all of the Buckeye metapopulation east of the East Bay hills.
This paper presents the available data on both the disappearance of the
Buckeye and its gradual recolonization-reoccupation of its inland range
in 1991.

The Normal Situation

Junonia coenia ranges throughout the Central Valley, San Francisco
Bay Area, foothills of the Coast Ranges and northern California moun-
tains, and western foothills of the Sierra Nevada. Southward it extends
across the Transverse Ranges and is common in coastal Southern
California, but scarce in the deserts. At higher elevations in the northern
mountains and Sierra Nevada it is a sporadic visitor from midsummer
through autumn. When it arrives early enough it may breed successfully
at least up to 2000m, but is unrecorded as overwintering at such
altitudes. On the Sierran east slope it is often collected as singletons on
flowers of Rabbitbrush ( Chrysothamnus , Comositae) along highways
(e.g., 89 and 395) in autumn. In some years it descends the Truckee River
drainage to Reno, NV but its occurrence east of the Sierran crest is always
chancy and short-lived.

Beginning in 1972, a transect has been maintained across north-
central California parallel to Interstate Highway 80 for studies of
butterfly phenology and faunsitics. At present there are 10 stations on
the transect. The number and identity of butterfly species flying at each
is recorded on a biweekly sampling schedule throughout the season.
Stations on the transect have been monitored from 6 to 20 yr. In addition,
butterfly phenology is monitored closely on the floor of the Sacramento
Valley at a network of stations not on the transect. I am thus in a good
position to detect anomalies — such as the disappearance of a common
species over a wide area for several months. I have in addition presence-
absence data from more or less regular trips to the north end of the State.

The Buckeye has been present at all low-elevation (below 1000m)
transect stations in all years of record. That, however, does not mean it
is a continuous resident. Shapiro ( 1974, p. 120) stated that it “is restricted
to bottomlands early in the season, but generally distributed by late vi.
It is not certain thatP. coenia overwinters in the (Sacramento) Valley at
all. It is abundant in the foothill canyons 3-6 wk before it appears on the
Valley floor.”

This assessment has held up. The first J. coenia on the Valley floor are
usually seen in III or IV and occur in riparian habitat. They may have
followed the streams down from the hills and are always spotty and rare;
the species is almost never common before mid to late V.

The first specimens seen in foothill canyons are very small and clay-

30(3-4):209-220, 1991

211

colored beneath. This phenotype is produced in autumn everywhere, but
infrequently seen in the Central Valley in spring. Table 1 demonstrates
the high variance in first-flight dates for low-elevation stations since
1982, as well as the pattern of early appearance in the foothills (repre-
sented by Gates Canyon) as compared to the Valley, with the Suisun
Marsh usually latest of all. Gates Canyon is an E-W oriented Coast Range

CD

"O

0

CD

0

Q

U

CD

0

JC

eo

0

CO

■Q

■4—*

o>

C/3

0

"D

C

0

00

0

-Q

0

0
■*— •
CO

0

>

o

"5

O

o

SZ

O

C

0

CC

c

o

>>

C

0

o

C/3

03

0

0

C/3

0

C

3

0

3

CO

X

X

X

X

X

X

X

X

X

X

» t

CO

T—

03

CM

CM

T“

d

cd

T”

CM

V“

i

i

T”

*1 —

1

22. V ■

I

>

1.VIII ■

14.1V

17.IV-

3. VII

19.11 -

- roe

10.VI -

16. VIII

> > == =

c\i c\i cd °o

CM CM CM 1—

~P ID -n "O

03

X

O

X

0

CM

X

cd

0

T—

v

d

CM

"O

>

A

A

i

CO

1

o

z

cd

>

>

—

>

CM

— :

LO

03

~0

CM

CM

O

O

0

0

0

>

O

O

0

>4

0

0

>

O

o

0

>%

0

0

>

O

o

0

>%

Q)

"0

>

0 0 co 0 'r_

X

0

03

X

X

X

X

CM

>

iJ)

d

CM

A

id

CM

CM

d

CM

1

i

__ _

EE

—

>

zz

>

cd

Lf3

Is"

CO

X

X

X

X

X

X

03

X

X

CM

cd
*1 —

CM

A

03

1

>

=

>

>

>

—

CM

CM

d

CM

03

T"

00

20.1

d

CM

X

X

cd

X

X

X

X

X

X

cd

CM

cd

cd

id

l —

C\l

00

>

>

>

>

>

1

>

>

>

00

d

d

03

d

cd

l

T—

T”

T-

CO

CM

03 C\J

WW^IO(ONCOO)0'-
00 00 00 00 00 00 00 00 030)

0303030)030303030303

o

CM

C

4-*

O

0

if)

c

0

c

o

if)

~D

»_

o

o

0

0

0

o

0

CD

03

0

00 i?
00 • -
03 0
i- O

c

X &

_ O)

® E

§ ®

"O -0

f 2

E ®

s §

o &
z <

212

J. Res . Lepid.

canyon and itself averages 2-4 wk later for the Buckeye than the nearby
N-S oriented Cold Canyon, which is monitored in spring but not part of
the transect.

The Buckeye does occasionally appear at Valley sites as early as I,
especially in drought years. There is no year-to-year pattern of repeated
overwinter survival at particular Valley sites. I interpret the picture
presented in table 1 as follows: the Buckeye is eradicated from most
Valley sites in “normal” winters; the probability of successful overwinter-
ing is enhanced in drought years, but even then it is a rare and stochastic
event. The Valley is recolonized during spring, mostly from foothill
populations but perhaps with some contributions from locally overwin-
tered animals. The population grows rapidly and always peaks in
September-October, coinciding with the flowering time of Coyotebrush
(. Baccharis pilularis, Compositae), a favorite nectar source, and then
tails off rapidly in November. In most years two or three apparent
“waves” of migrating Buckeyes are seen passing through Davis during
the summer and fall. Their general direction is S to N or SW to NE, but
their sources are unknown and they are rarely abundant.

Junonia is a tropical and subtropical genus. In the eastern United
States the northern boundary of J. coenia fluctuates from year to year,
apparently due to cold intolerance. Opler and Krizek (1984) summarize
a vast literature thus; “The Buckeye ranges from southern Canada
through most of the United States to northern Mexico. Throughout most
of its range the species cannot survive the winter, and year-round
populations occur only in the southern portions of its range. It can survive
the winter on coastal dunes as far north as North Carolina.” Clark ( 1932)
reported some overwintering at Washington, D.C., but his spring indi-
viduals could just as easily be immigrants. No one has even alleged the
existence of a southward migration in autumn in the northern 2/3 of the
range. Presumably, then, all reproduction in the north is ultimately in
vain.

Northern California is the northernmost area where J. coenia is
considered a permanent resident. Dornfeld (1980) treated it only as an
immigrant into southern Oregon. The events of 1990-91 can be viewed as
a temporary shift downslope and southward of the average threshold for
successful overwintering near the northern edge of the species range.

The Catastrophe

On December 16-17, 1990 a low-pressure system from the Gulf of
Alaska moved SE into northern British Columbia and Alberta, trailing
a cold front. This is a normal sequence of events, but the context in which
it occurred was not. A block in the northern hemisphere atmospheric
circulation had ponded up a huge mass of stagnant, cold high pressure in
the Arctic, and it eventually had to break out of confinement. Blocking
high pressure remained over the Atlantic Ocean. The 16-17.XII system
set the stage for the escape; when on 19.XII a second system moved
southward on the Pacific coast, the pressure pattern allowed the bottled-

30(3-4):209-220, 1991

213

up Arctic air to rush straight southward instead of being deflected to the
SE east of the Sierra Nevada, as normally occurs. From the 21st through
the 24th, air direct from the Yukon flowed down into the Central Valley
on gusty NNE winds induced by a low-level jet stream. The wind eased
on Christmas Day but picked up again on the 26th, blowing until the end
of the month. This was the worst cold wave in California since 1932, and
in some areas since 1913. The worst night was the 22nd, setting dozens
of records up and down the State. In many localities this was the coldest
night ever recorded. Sensible temperatures (wind-chill temperatures)
frequently dropped below zero F (-17.8 C). Of 190 California weather
stations reporting deviations from 30-yr monthly mean temperatures in
December, 187 were negative and the average deviation was greater
than -3°F. Examples of climatological data for this period appear in
tables 2-4.

Economic consequences were severe, reflecting the impact of such
unusual temperatures on exotic species. The entire citrus crop was lost,
and many trees were killed or badly damaged. The severity of damage
was exacerbated by the desiccating effects of the wind; on the 22nd mid-
afternoon humidities were near 10% at many stations, and overnight
rebounded only to 40%. Vegetable crops were damaged, and recently-
planted fields required replanting. Damage to exotic ornamentals was
much more extensive than in the shorter freeze of 1972. Among woody
genera severely damaged were Eucalyptus , Greuillea and Brachy chiton
(all from Australia). All specimens of Canary Island Pine, Pinus
canariensis, in the Davis-Sacramento and Chico areas lost most of their
needles. They appeared dead, but nearly all recovered. Standing dead
and damaged Eucalyptus in the East Bay hills contributed to the
destructive wildfire of 20-2 l.X. 1991.

The Aftermath: Transect Data

Numbers of individual butterflies are not counted as part of the
transect study; only presence/absence data are taken. However, when I
realized that the Buckeye was absent at all stations, I resolved to count
all individuals of that species until numbers approached normal and this
became impractical. That point was never reached in 1991.

On 31. Ill I saw a single female at Willow Slough, Yolo County, 5 km N
Davis, on the floor of the Sacramento Valley. That is a normal time and
place for a first Valley record of the year, but it is very unusual for no
Buckeyes to have been seen by then in the foothills, and such was the
case. (In 1985 the Buckeye was seen in the Valley before the foothills.)
This was the last individual seen for 90 days! On 29. VI I observed two
males in riparian habitat in North Sacramento, Sacramento Co. — one
worn, one much fresher. 32 days later I saw another male in West
Sacramento, Yolo Co. None was seen at the Suisun Marsh until 10.VIII
( 1 male): at Rancho Cordova, Sacramento Co. , until 16. VIII ( 1 male); and
in the foothills at Gates Canyon until 17.VIII (also 1 male). The cluster
of mid-VIII sightings seems not to be coincidental. Single females in

Table 2. Daily T observations (in °F) for the second half of December, 1990 for selected stations affected by the freeze. From

Climatological Data, California, Vol. 90 #12, NOAA, Asheville, N.C.

214

T

00

LO

00

CO

"3T

or

LO

(D

O'

CD

r-

CO

CM

o

LO

CO

-M-

CM

CM

LO

CM

LO

CM

o

CM

O'

CM

LO

CO

LO

CM

-o

CD

•*_

o

O

00

T—

CD

O

r-

00

T—

O'

CD

O'

o

CO

00

CO

L.

3

CO

CD

CM

LO

CM

CM

or

CM

LO

CM

O'

CM

LO

CO

O'

CM

o

o

o

CD

CD

r-

T—

O

O

r-

CM

00

o

Is-

00

r-

CM

LO

00

O'

"cO

C\J

CM

LO

CO

LO

CM

LO

CM

LO

CM

o-

CM

LO

CO

O'

CM

>

D

CO

LO

CO

00

00

LO

O

LO

O'

O'

CM

r-

O'

o

Is-

CD

CO

C\J

LO

CM

LO

CM

CM

LO

CM

LO

CM

LO

CM

LO

O'

LO

CM

CD

E

o

Is-

CO

CO

CM

00

CD

LO

r-

r-

O'

CO

CD

T—

LO

CM

00

CO

CM

LO

CM

LO

CM

LO

CM

LO

CM

LO

CM

LO

CM

LO

CO

LO

CM

0)

1_

CD

JZ

CD

00

CM

o

CD

T—

CO

CD

CD

CM

O'

CD

00

LO

LO

Is-

CM

$

CM

CM

LO

CM

LO

CM

LO

CM

LO

CM

O'

CM

LO

CO

o-

CM

c

B

LO

00

O

o

CO

00

CM

O

CO

T—

—

00

O'

CM

r-

LO

O

c

3

CM

CM

LO

CM

CM

LO

CM

LO

CM

O'

CM

LO

CO

O'

CM

o

Is-

CM

00

CD

CD

CO

CM

00

CD

CD

CM

CD

o

c

CM

T—

"O'

CM

•M-

O’

CM

o

CM

O'

CM

O'

CM

O'

D

co

CO

Is-

CD

CD

CD

CM

o

00

CD

LO

Is-

CD

CM

00

O'

LO

c

CO

to

o

CM

CO

CO

CM

O'

00

•t —

00

O'

CM

CO

CM

CO

00

r-

Is-

Is-

CO

r-

00

CD

CD

CM

CD

00

Is-

T“

CO

»«— *

r"

CM

CO

CO

CO

00

i —

CO

CO

CO

CM

CO

CD

O

-5,

1 —

CO

i —

'M'

LO

CM

00

LO

O'

CO

CM

i —

O

00

CD

CD

~o

CM

CM

CM

CO

CM

CO

CM

O"

CM

O'

CM

O'

CM

CO

CO

o

O

Is-

CD

CO

CO

LO

r-

o

LO

00

O'

CD

O'

O

CD

d

n

CM

CM

CO

CM

O'

LO

CM

O'

CM

O'

CO

O'

>s

C

CO

CD

o

Is-

CD

O

CM

r-

T —

LO

CM

CM

o

o

T_

CM

O

LO

CO

CO

'M'

CO

LO

00

LO

CO

LO

CO

LO

O'

LO

CO

CD
CD
•+— «
CO

00

CO

cd

CD

1 —

CM

CM

o

CM

CO

O'

O

CD

o

CM

00

C D

LO

CM

LO

CO

LO

CO

LO

CO

CD

00

LO

00

LO

O'

LO

CM

o

■*—>

c

Is-

—

CM

CD

CO

CO

00

CO

LO

CD

CM

00

CD

o

CM

Is-

Q)

LO

CO

LO

CO

LO

CO

LO

CO

LO

CO

LO

00

LO

O'

LO

CM

o

CO

~o

CO

CD

O

LO

00

CD

CO

00

LO

CD

T—

CD

CD

Is-

CO

O'

O'

O

_Q

LO

CO

CO

LO

CO

LO

CO

LO

CO

O'

00

LO

O'

O'

CO

_d

CD

X

<

z

X

<

Z

X

<

Z

X

<

z

X

<

Z

X

<

Z

X

<

z

X

<

Z

V.

CO

2

tT

o

CL

u.

2

2

Airport M

2

c

D

CD

D

CO

O

« 4 — •

<:

<

o

CO

O

o

CD

E

C

£3

ID

O

Id

(0

CD

J0

C

C

. . CM

Z (/)

TD

CD

E

CO

S

o

CD

CO

"O

CO

V-.

LL.

CD

o >

t=

o

C

O

C

o

CO

a.

LU

H

jZ CO
< Q

"co

LL

CO

cn

CO

>

LJ

CO

CO

<

1—

Q

C/)

c

0) _J

— i

r. Res. Lepid.

.angston). d representative Sierran foothill site (360m).

30(3-4):209-220, 1991

215

Table 3. Departure of mean temperature from 30-yr means and lowest
temperature (°F) registered for selected stations affected by the freeze,
December 1990. From Climatological Data, California, Vol 90# 12, NOAA,

Asheville, N.C.

STATION:

Departure From Mean T:

Lowest T Recorded:

Angwin (Pacific Union College)

N.A.

16 F

Flealdsburg

-4.1

14

Napa State Hospital

-6.2

14

Santa Rosa

-2.9

18

Auburn

-4.3

17

Chico University Farm

N.A.

17

Clearlake, 4 mi SE

N.A.

6

Fairfield

-4.0

17

Placerville

-1.4

13

Sacramento (Airport)

-4.6

18

Sacramento (City)

-3.8

18

Vacaville

-3.1

17

Willows 6 mi W

-4.1

11

Winters

-3.7

15

Woodland

-3.9

19

Davis 2 mi SW

-4.9

16

Palo Alto

-5.7

20

Redwood City

-5.4

19

San Francisco Airport

-2.8

27

Santa Cruz

-4.0

19

Watsonville Water works

-5.8

12

Modesto

-2.3

19

Stockton Fire Station #4

-5.3

15

Turlock

N.A.

19

apparent migratory flight were seen in Davis on 19.VIII (WNW to ESE)
and 25.VIII (S to N), and a male was seen on flowers on 6. IX. These were
the first Buckeyes of the year in Davis — two to three months later than
average.

Numbers continued to increase in late summer and early autumn
(table 5) in approximately the usual seasonal pattern, but remained far
below average for calendar date. Since counts are not normally done, I
can only guess that numbers were only about 10% of “average” in the late
summer. Junonia coenia is normally present in every flower garden in
Davis in September and October. In 1991 1 never saw one in my garden,
and sightings were so few that all were carefully noted. Other observers
report similar experiences (see below).

Not one Buckeye was seen above 1000m in all my 1991 field work,
although I spent 81 days afield in the Sierra Nevada and another 35
elsewhere in the mountains. On 19.X one male and two females were
observed flying up the South Yuba River canyon near Washington,
Nevada Co., at ca. 900m. By then, successful breeding was out of the
question, and butterfly activity had ended there by 16.XI.

216

J. Res. Lepid.

Table 4. Minimum temperature records for Sacramento, December 1878 through
1987 with comparison to 1990 minima. Degrees Fahrenheit. From NOAA
Technical Memorandum NWS WR-65 (Revised). Compare December 9-15, 1932.

Date

Record Low, Year

Low, 1990

Date

Record Low, Year

Low, 1990

1

32, 1929

29

17

28, 1928

33

2

30, 1906

31

18

28, 1924

32

3

32, 1918

33

19

25, 1924

30

4

29, 1909

34

20

27, 1928

23

5

32, 1972

36

21

26, 1928

22

6

29, 1891

34

22

25, 1928

18

7

28, 1891

33

23

28, 1930

20

8

27, 1972

31

24

25, 1879

19

9

23, 1932

32

25

26, 1891

22

10

22, 1932

39

26

25, 1879

23

11

17, 1932

46

27

27, 1878

26

12

21, 1932

35

28

26, 1930

25

13

23, 1932

36

29

24, 1878

27

14

23, 1940

32

30

28, 1962

20

15

26, 1932

39

31

24, 1915

24

16

26, 1892

38

The Aftermath: Data From Others

In September I wrote to a number of northern and central California
Lepidopterists soliciting Buckeye data. A total of 18 respondents contrib-
uted to the synthesis which follows. Their experiences range from
extensive field tripping to observing their own gardens. Information was
received covering most of the area from Trinity and Siskiyou Cos. on the
N to Inyo and Stanislaus Cos. on the S. A selection of observations
follows; it must be stressed that none of those responding had found
normal or nearly normal numbers of Buckeyes, and several had seen
none at all to that time.

W. Swisher (Santa Rosa): Only 1 or 2 between I and VI; perhaps
20% of “normal” numbers by IX.

J. Mori (Modesto): None until 9.X (7 observed that date; 4 of these
apparently migrating S to N).

W. Patterson (Sacramento): “In II-III, 1990 common along Middle
Fork American River (Sierra foothills); none in 1991. One 1 mi
E Pilot Hill, Eldorado Co., 12.V.91. One at Del Puerto Canyon,
Stanislaus Co. (Inner Coast Range), 30. VIII. 91.” (A Buckeye at
Del Puerto Canyon would not normally be noteworthy at any
time of year — AMS.)

L. Smith (Sacramento): None in Sacramento until one each, 7. IX
and 13. IX.

30(3-4):209-220, 1991

217

Table 5. Individual counts of Buckeyes at low-altitude transect stations, 1991 .

RANCHO

SUISUN

GATES

NORTH

WEST

CORDOVA

MARSH

CANYON

SACRAMENTO

SACRAMENTO

1.1-0

16.1-0

20.1-0

19.1-0

17.1-0

18.1-0

22.1-0

30.1-0

26.1-0

23.1-0

29.1-0

9.11-0

12.11-0

7.11-0

28.1-0

10.11-0

18.11-0

17.11-0

14.11-0

6.11-0

20.11-0

6. Ill - 0

24.11-0

23.11-0

16.11-0

13.111 -0

27.111 -0

14.111 -0

16.111 -0

25.11-0

29.111 -0

2.1V -0

28.111 -0

30.111 -0

8. Ill - 0

4.1V -0

10.1V - 0

3.1V -0

7.1V -0

22.111 -0

14. IV - 0

28. IV - 0

13. IV - 0

21 .IV - 0

30.111 -0

cn

<

i

o

14. V - 0

28. IV - 0

cn

<

o

9.1V -0

16. V - 0

25. V - 0

11.V-0

16. V - 0

17. IV - 0

30. V - 0

4. VI - 0

25. V - 0

28. V - 0

5.V-0

19. VI - 0

20. VI - 0

3. VI - 0

12. VI - 0

19. V - 0

1 .VII - 0

6. VII - 0

20. VI - 0

29. VI - 2

30. V - 0

16. VII - 0

21 .VII -0

9. VII - 0

15. VII - 0

15. VI - 0

1 .VIII - 0

10.VIII - 1

4. VIII - 0

1 .VIII - 1

29. VI - 0

16. VIII - 1

23. VIII - 0

17. VIII - 1

16. VIII - 1

14. VII - 0

31 .VIII -0

28. VIII - 6

3. IX - 0

2. IX - 2

31 .VII - 1

15. IX - 4

7. IX - 1

11. IX- 11

15. IX - 1

15. VIII - 2

28. IX - 1

16. IX - 4

25. IX - 20

27. IX - 3

27. VIII - 3

13.X -8

30.IX- 15

18.X- 10

11.X -0

10. IX - 8

27.X- 10

14.X -35

4. XI - 25

27.X - 8

24. IX - 5

9. XI - 0

1.XI -0

22. XI - 2

11. XI - 1

11.X -4

24. XI - 2

15. XI - 2

1 1 .XII - 0

23. XI - 1

23.X - 2

9. XII - 0

29. XI - 0

24. XII - 3a

4. XII - 0

5. XI - 0

12. XII - 0

20. XII - 0

23. XI - 1

19. XII - 0

1 .XII - 0

20. XII - 5a

a Apparently new emergence, much later than average.

C. Nice (Davis): One each in Davis 29.VTI, 20.VIII; three in West
Sacramento, Yolo Co., 18. DC (about 3.5 km N of my West
Sacramento site — AMS).

G. Kareofelas (Davis): In extensive travels in northern California,
only 3 Buckeyes seen in 1991: 1 at Grindstone Overlook, Glenn
Co. (Inner North Coast Range), 4.VIII; 1 at Suisun City, Solano
Co., 17. IX; 1 5 km S Rio Linda, Sacramento, Co., 22.XI.

S. North (Areata): A month late on North Coast and in Trinity Co.

J.R. Tucker (Redding): None in Trinity foothills - Weaverville area
in spring.

D. Giuliani ( teste O. Shields): Several seen migrating W to E, along
ridge running N from summit of Mt. Tamalpais, Marin Co.,

218

J. Res. Lepid.

Table 6. Biweekly total sightings of individual Buckeyes along the transect, 1991.

2-week interval beginning:

3. VI 0

1 7. VI 2

1 .VII 0

15. VII ..0

29. VII 3

12. VIII 5

26. VIII 12

9. IX 28

23. IX 44

7.X 60

21.X 20a

4. XI 28

18. XI ...6

2. XII ......0

a Cold wave 22. X and heavy rain 25-26. X depressed counts.

19.X; several more also migrating W to E, Point Reyes Penin-
sula, 20.X.

Very detailed data were received from two respondents, viz.:

O. Shields (Mariposa): None seen in Mariposa Co. until mid-IX. At
Jerseydale, Mariposa Co., 19 12. IX, IS 23. IX, 19 l.X, 29 4.X,
then continuously present until 2. XI, commonest in mid-X
when up to 15 were counted on 16.X, with some evidence of
migratory movement toward the SW 16-24.X. Also 2 seen at
Mariposa, 7.XI. Also 2 seen in Monterey Co., 8. IX.

R. Langston (Kensington): “Buckeyes are usually observed in
small numbers November through February in most winters in
the Bay Area.” He provides winter counts at San Bruno
Mountain for 1989-90 and 1990-91. The last specimen seen in
1990 was on 28.XI: zero seen in 12 days in XII, zero in 16 days
in 1.91; the first post-freeze sighting was on 15. Ill and numbers
in IV were low. “By mid-IX the Buckeye in ‘fair’ numbers on
SBM and noted in Kensington. However, quite scarce on 3
Xerces (Fourth of July) counts where usually much
commoner:... Berkeley (15), Mt. Diablo (4), and South San
Francisco Bay (only 1!).”

Discussion

This level of coverage is far spottier than one would have arranged had
there been advance warning. Still, it is good enough to persuade one that
Junonia coenia experienced a severe die-off during the December 1990

30(3-4):209-220, 1991

219

Fig. 1 . Sequence of first Buckeye observations in 1 991 at transect localities (sampled
biweekly) and localities with a resident observer.

Localities: 1 San Bruno Mountain 15.111

2 North Sacramento 29. VI

3 Davis 29. VII

4 West Sacramento 31. VII

5 Suisun Marsh 10. VIII

6 Rancho Cordova 16. VII

7 Gates Canyon 17. VIII

8 Jerseydale 12. IX

9 Modesto 9.X

10 Washington 19.X

freeze and gradually reoccupied much of the lost territory during the
second half of summer 1991. The meteorological records indicate that the
freeze was most severe inland, so it is not surprising that Buckeye
survival seems to have been better at the coast (Langston and J.A.
Powell). It is not, however, certain that recolonization of the Sacramento
Valley proceeded from the Bay Area by way of the Sacramento-San
Joaquin Delta, since recolonization of the Suisun Marsh was so late in
comparison with metropolitan Sacramento. Mori reports no Buckeyes at
Modesto until early October, the same time they became common in the
Sierran foothills in Mariposa Co. (Shields). The apparent wave of immi-
grants observed at Davis in mid-August showed no clear-cut directional-
ity. It may be that recolonization resulted from the spreading out of
several independent foci where some overwinter survival had occurred,
rather than being traceable to a single source. At any rate, the various
local populations had effectively coalesced at least near Sacramento by

220

J. Res. Lepid.

the end of the season, as normally happens several months earlier (fig.

1).

As far as I know, there are no quantitative data demonstrating that
Buckeye population growth in normal Central Valley summers is an
exponential process, but my subjective impression after 20 yr is that it is.
When the total number of observations on the transect is summed for
each biweekly sampling cycle, the result for 1991 is very nearly an
exponential process, with a doubling time of nearly 3 wk during the
interval late Vl-early X (table 6). This can only be suggestive, but it does
suggest that Buckeye populations grew in reasonably normal fashion in
1991 but with a substantial handicap due to their late start, which they
could not overcome before the season came to an end. Just as an early
start due to a dry, mild winter seems to have led to an outstanding
Buckeye flight in 1990, the freeze of December 1990 set up the Buckeye
with unusually low numbers of individuals attempting to hibernate over
the winter of 1991-92.

Acknowledgments. I am pleased to thank all the individuals quoted in the text,
as well as P. Cherubini, R. Dowell, C. Dunlap, A. Ludtke, J.A. Powell, S.O.
Mattoon and R. Wells. This study forms part of CA Agricultural Experiment
Station project CA-D*-AZO-39994-H, “Climatic Range Limitation in
Phytophogous Lepidopterans.”

Literature Cited

Andrewartha, H.G. & L.C. Birch. 1954. The Distribution and Abundance of
Animals. Chicago: University of Chicago Press. 782 pp.

Arnold, R.A. 1981. A review of endangered species legislation in the U.S.A.
and preliminary research on six endangered California butterflies
(Lepidoptera: Lycaenidae). Beih. Veroff. Naturschutz Landschaftspflege
Bad.-Wurtt. 21: 79-96.

1983. Ecological studies of six endangered butterflies (Lepidoptera:

Lycaenidae): island biogeography, patch dynamics, and the design of
habitat preserves. Univ. Calif. Publ. Ent. 99: 1-161.

Clark, A.H. 1932. The butterflies of the District of Columbia and vicinity.
Bull. U.S. Nat Mus. 157: 1-337.

Dornfeld, E.J. 1980. The Butterflies of Oregon, Forest Grove, OR: Timber
Press 276 pp.

Garth, J.S. & J.W. Tilden. 1986. California Butterflies. Berkeley: University
of California Press. 247 pp.

Levins, R. 1970. Extinction, pp. 77-107 in M. Gerstenhaber, ed. Some
Mathematical Questions in Biology. Providence, RI: American
Mathematical Society.

Opler, P.A., & G.R. Krizek. 1984. Butterflies East of the Great Plains: An
illustrated Natural History. Baltimore: Johns Hopkins University Press.
294 pp.

Shapiro, A.M. 1974. The butterfly fauna of the Sacramento Valley, California.

J. Res. Lep. 13: 73-82, 115-122, 137-148.

Tilden, J.W., & A.C. Smith. 1986. A field Guide to Western Butterflies.
Boston: Houghton Mifflin. 370 pp.

Journal of Research on the Lepidoptera

30(3-4):221-224, 1991

Notes on the Immature Biology of Two Riodinine
Butterflies: Metacharis ptolomaeus and Napaea
nepos orpheus (Lycaenidae)

Curtis J.Callaghan

Av. Suba 130-25 Casa 6, Bogota, Colombia

Introduction

This paper presents field and laboratory observations on the biol-
ogy of two riodinid butterflies from southeast Brazil: Metacharis
pfo/omae ws(Fabricius, 1793) andNapaea nepos orpheus (Westwood, [1851])
that supplements information published previously (Zikan 1953;
Callaghan 1985).

Metacharis ptolomaeus ranges throughout southeast Brazil from south-
ern Bahia (Pernambuco?) south along the coast and inland to eastern
Minas Gerais, Santa Catherina and Parana States from sea level to
about 1000 meters, and is particularly common along the coast.

On August 30, 1988, near Barra de Sao Joao, Rio de Janeiro State. (site
described in Callaghan 1985), I observed a female M. ptolomaeus ovipos-
iting on the leaves of a large tree around 1500 hours. Eggs were laid
singly and well diszxxpersed on separate branches of the foodplant, later
identified as Heisteria sp.(Olacaceae), previously reported in error as
Lacistema sp. (Callaghan, 1985) The female dragged her abdomen on the
substrate before ovipositing. Three other captured females were induced
to oviposit on food plant placed in a plastic box, yielding 9 eggs.

Imature Stages

EGG: Diameter 0.5mm, height 0.2mm. Color shiny bronze. Sides
covered with network of ridges forming hexagonal figures, with a small
tubercle at each intersection. Duration 12 days. N=9

FIRST INSTAR LARVA: (fig. 1) Length 1-1. 5mm, Thorax and abdomen
light yellow, pubescent. Head dark brown with numerous small setae on
front; headcapsule width 0.13mm. Prothoracic shield with high trans-
verse ridge from which 4 long, black setae project cephalad, followed
posteriorad by 4 longer setae then by two shorter ones, these turning
slightly caudad then cephalad; T2 with two tubercles dorsad from which
two long, black setae extend, first curving cephalad then caudad. T3
through A8 with 2 pairs of long, black setae dorsad, each emerging from
a small tubercle, and three similar setae extending laterally from the
base of the dorsal plate on each side; A10 covered by an anal shield from
which 6 long black setae extend caudad. Spiracles indistinct; found on Tl,
ventrad on A1 and laterally on A2-A8. n= 5

SECOND - FOURTH INSTAR LARVA: (fig. 2) Color light green, body
thickened laterally and dorsad, tapering caudad and covered with long
white lateral setae, giving larva an arctiid like appearance. Head brown,

222

J. Res. Lepid.

many short setae on front. Prothoracic shield with numerous white setae
extending cephalad, followed by a transverse ridge with 10 thick, blunt
setae with small barbs along shaft; T2 with two tubercles from which
extend two similar barbed setae; T3-A8 with 2 pairs of tubercles dorsad,
the first pair long with black bulbs on the tips, curving caudad, second
shorter with a smaller bulb; A1 and A6 with dark area dorsad across base
of setae. Anal shield with 4 long setae extending caudad. Spiracles found
laterally on Tl, ventrally on A1 and laterally on A2 through A8, that on
A2 higher than the rest.n=2

FINAL INSTAR LARVA: (fig. 3) Length 15mm, Color dark green
dorsad, light green ventrad with white mottling. Head dark brown; head
capsule 2mm wide. T1-T3 with white dorsal marks; Setae same as on
second instar, those on prothoracic shield and T2 with black bulbous tips
similar to those on T3 through A8; base of setae on A6 raised higher on
black tubercles; many white setae extending laterally from base of dorsal
plate. End of anal shield with 4 long non bulbous setae extending caudad.
White lateral spiracles on Tl and A2 through A8 with a row of white
lenticle patches dorsad; spiracle on A1 ventrad.

Prepupal stage lasts 2 days, larva turning lighter green color and
remaining motionless on leaf. Larval development time, all instars 42
days. n=l.

PUPA: (fig.4 ) Length 13mm, width 5mm. Light green with black marks
on wing cases, elongated and pointed caudad; Tl with slightly bifurcated
crest; dorsal surface covered with small, mushroom shaped tubercles
with tiny projecting teeth; spiracles on A2 dorsad, those on A4-A8 lateral;
pupa secured by cremaster and silk threads along entire ventral surface,
without girdle. Duration 10 days. n=l.

The first instar larvae fed between the veins on the ventral leaf surface
and in later instars on the entire leaf. All instars rested for long periods,
especially during molts. The larvae expelled frass forcibly, by raising the
end of the abdomen, flopping it down, and ejecting frass 150 to 200mm
from the larva. The larvae were cannibalistic. Three first instar larvae
were devoured by their larger siblings, which may explain the dispersed
ovipositing behaviour. I observed no behaviour or organs indicating
myrmecophily. The larvae were covered with long setae and showed no
evidence of myrmecophilous organs. Just to be sure, I placed Camponotus
ants with the Metacharis larvae and the ants totally ignored them.

Napaea nepos consists of three subspecies. Nominate nepos ranges over
the Amazon basin, intergrading into subspecies tanos Stichel in Bolivia,
and orpheus, which inhabits mountain areas above 900 meters in
southeast Brazil, from Espirito Santo, eastern Minas Gerais south to
Parana and Santa Catherina States.

Zikan (1953) recorded larvae of orpheus on Oncidium sp. and an
“Erdorchidae”(ground orchid). I found two final instar larvae of this
species on Zygopetalus orchids in my garden in Petropolis, Rio de
Janeiro, 900m. The larvae rested among the roots of the plants during the
day and fed at night. They were brought into the lab and raised.

30(3-4):221-224, 1991

223

Figures: Illustrations of M.ptolomaeus life history. 1 . First instar larva. 2. Third instar
larva 3. Flead and thorax of final instar larva. 4. Pupa- lateral, dorsal and
ventral views.

224

J. Res. Lepid.

FINAL INSTAR LARVA: Length 11mm. Larva dorsally rounded,
abdomen narrowing caudad to flat anal shield. Color dark green, dorsad
a purple spot between each segment, bordered laterally by two pairs of
white spots. Head is light brown with numerous short setae and large
labrum and heavily sclerotized, toothed mandibles; head capsule width
3 mm. Prothoracic shield extends over head with many long, barbed setae
projecting cephalad. Thorax and abdomen covered with short, bunched
setae, the longest being on the prothoracic shield projecting cephalad and
along the base of the dorsal plate and at end of anal shield. Spiracles
brown, lateral on T1 and on A1 to A8, those on T1 and A8 larger than
others; that on A2 and T1 slightly more dorsad than rest. One white
lenticel patch dorsad of each spiracle. n=2
PUPA: Length 13mm, width 5mm. Abdomen and thorax light green,
wing cases paler green. T1 with notched crest outlined in black. Spiracles
on segments A2 through A7 with that on A3 hidden beneath wing cases.
Pupa secured by cremaster only, resting at 45 degree angle from sub-
strate. Duration: 16 days. n=l.

Although ants were found feeding on the orchid extrafloral nectaries,
none were observed to take interest in the N. nepos larvae. This observa-
tion, plus the lack of myrmecophilous organs and long setae suggest that
like M. ptolomaeus , its relationship with ants is defensive.

Acknowledgements: I thank Drs. A. Mattos and K. Brown of the Universidade
Estadual de Campinas (SP) for identification of the foodplant of M. ptolomaeus
and Drs. Don Harvey of the U.S. Natural History Museum, Washington D.C.
and Philip DeVries of the University of Texas, Austin, for their helpful com-
ments on the manuscript.

Literature Cited

Callaghan, C.J., 1985(86). Notes on the biology of three riodinine species:
Nymphidium lisimon attenuatum, Phaenochitonia sagaris satnius,
Metacharis ptolomaeus (Lycaenidae.Riodininae) J. Res. Lep. 24(3).

Zikan, J.F. 1953. Beitrage zur Biologie von 19 Riodininen Arten (Riodinidae-
Lepidoptera), Dusenia:IV (5, 6), 402-413.

Journal of Research on the Lepidoptera

30(3-4):225-236, 1991

The Effects of Temperature and Daylength on the
Rosa Polyphenism in the Buckeye Butterfly, Precis
coenia (Lepidoptera: Nymphalidae)

Kelly C. Smith

Departments of Zoology and Philosophy Duke University, Durham, NC 27708

Abstract. In North Carolina, Precis coenia that emerge during the
Summer months exhibit a ventral hindwing (VHW) with well-defined
reddish-brown and brown pattern elements on a light tan background.
During late Summer and early Fall, however, individuals begin to
appear with poorly defined or obscured pattern elements on a dark
reddish-brown background. The present study shows that the Fall
(. rosa ) color morph can be induced by either low rearing temperatures
or short daylengths. The effect of such conditions seems to be cumula-
tive throughout the larval life, although animals are much more
sensitive during the last 24 hours of larval life and immediately after
pupation.

Introduction

Seasonal polyphenism involves a repeating pattern of changing pheno-
types which is under the control of an environmental factor (Shapiro,
1976; Tauber, et. al., 1986). It is important to differentiate this from
genetic polymorphism, where the phenotypic differences in a population
are the result of genetic differences. The Buckeye butterfly, Precis coenia
(Lepidoptera:Nymphalidae), has long been known to exhibit a seasonally
polyphonic color pattern (Clark, 1932; Klots, 1951; Mather, 1968). The
background coloration of the ventral hind wing (VHW) surface changes
with the season: the Summer {tinea) morph has a pale tan background
with well-defined pattern elements while the Fall {rosa) morph has a
dark reddish-brown background with indistinct elements (Figure 1). The
rosa morph becomes predominant during the F all and early Winter, with
the exact time of its appearance depending on local conditions.

The control of the rosa polyphenism had not been studied in detail,
though several authors have speculated on its mechanism. Clark (1932)
suggested that high humidity played a role while Howe (1975) proposed
a photoperiodic control. At the very least, the control mechanism did not
seem to be straightforward, a fact which prompted Shapiro (1976) to
report no obvious environmental correlations at all. The present study,
however, shows that the rosa morph can be induced by exposing devel-
oping larvae either to low temperatures or short daylengths.

Materials and methods

Precis coenia stock was derived from a laboratory population maintained at
Duke University and constituted from animals collected over a period of years in
the region around Durham, NC. A random sample of the eggs produced by this

226

J. Res. Lepid.

Figure 1 . Ventral wing patterns: 1 a) shows the ventral wing pattern of the Summer
or Linea morph (score = 0). The backgound is light tan and the pattern
elements are quite distinct. 1b) shows the ventral pattern of the Winter or
Rosa morph (score = 3). The background coloration is uniformly reddish-
brown and the pattern elements are no longer distinct.

population was used to found a breeding colony which was maintained under
controlled environmental conditions at a temperature of 27°C and a daylength of
16 hrs. light : 8 hrs. dark (16L:8D). To assess the influence of daylength and
temperature on the expression of the rosa morph, eggs were removed from the
breeding colony and randomly assigned to one of nine controlled temperature and
daylength regimes.

Animals assigned to “seasonal” conditions (from June through November of
1990) were reared in a semi-shaded area of an outdoor cage. Those assigned to
regimes of 27°C, 16L:8D were reared in the same constant temperature room as
the breeding colony. All other regimes were maintained in incubators (accuracy
of ± 1°C) with lighting provided by 15W fluorescent bulbs on clock timers. All
larvae were mass-reared in plastic boxes on an artificial diet containing pow-
dered Plantago lanceolata leaves (Smith, 1991).

Results

Adults from each experimental treatment as well as from the breeding
colonies were scored as to VHW background coloration. The scoring
system assigned each animal a score from 0 (the linea Summer morph)

30(3-4):225-236, 1991

227

to 4 (an extreme rosa morph) and is summarized in Table 1. The response
of each treatment was then defined in terms of two response character-
istics: mean score and % response (percentage of animals with scores >
0).

BREEDING COLONY

Non-Zmea morphs (score > 0) were encountered in the breeding colony.
However, the incidence of such morphs was relatively low (14%) and
there were no individuals with more than intermediate expression of the

Table 1 . Scoring system: System used used to quantify the degree of Rosa
expression. Mean score and % response (percentage of animals with score > 0)
were the two response characteristics used to compare treatments. All animals
were scored at least twice and in no case did group means vary by more than 5%.

SCORE MORPH DESCRIPTION

0 Linea Light tan VHW background with

distinct reddish-brown pattern
elements.

1 Light Intermediate Predominantly tan VHW background

with pinkish tinge proximal to the mid-
wing umbral band and along the
trailing edge of the hindwing.

2 Dark Intermediate Predominantly reddish-brown VHW

background with tan areas still visible
distal to the mid-wing umbral band.
Pattern elements in the reddish-brown
areas somewhat obscured.

3 Rosa Uniformly reddish-brown VHW

background and wing edges with red
or pink regions distal to the mid-wing
umbral band. Pattern elements
severely obscured.

4 Extreme Rosa Solid reddish-purple VHW background

with vestigial pattern elements and
distal wing edges a dark purple.

228

J. Res. Lepid.

Emergence Date Emergence Date

Figure 2. Response to seasonal conditions: The graphs show the moving mean
scores (9 day interval centered on date of emergence) of animals raised
under seasonal conditions in an outdoor cage as compared to those under
the controlled conditions of the breeding colony (27°C, 1 6L:8D). The mean
scores of the two groups are initially quite similar but begin to diverge in mid-
August and become maximally dissimilar by October. 2a shows the
response of the two groups as compared to the seasonal daylengths. 2b
shows the response of the two groups as compared with the moving mean
(9 day interval) of average daily temperatures.

Breeding colony scores (27°C, 16L:8D)

seasonal colony scores (outdoor cage)

— — Temperature / daylength

rosa morph (score > 2). Based on these findings, the response character-
istics of the breeding colony (27°C and 16L:8D) were used as a baseline
of comparison for other treatments.

SEASONAL CONDITIONS

Individuals reared outdoors under seasonal conditions showed a marked
increase in both the incidence and degree of rosa correlated with short-
ening daylengths and falling temperatures (Figure 2). The mean scores
of animals raised under the conditions of the breeding colony (27°C,
16L:8D) and those raised outdoors under seasonal conditions remained
quite similar from July through early August. However, by mid-August
the responses began to diverge, with the population under seasonal
conditions exhibiting a progressively higher incidence and degree of rosa
coloration.

TEMPERATURE AND PHOTOPERIODIC REGIMES

While the seasonal exposures suggested that environmental cues
might influence the onset and intensity of the rosa morph, the concurrent
decreases in daylength and temperature made differentiation between

Mean temperature (moving average, °C)

30(3-4):225-236, 1991

229

Summary of Response for all Stable Treatments
Temperature

Photo 18°C 21 °C 24°C 27°C

16L

A

SCORE =3.1 RSP=10Q%
ERROR= 0.26 N=7

SCORE =1.:
ERROR= 0.1

J RSP=81%
10 N=73

SCORE =0.34 RSP=27%
ERROR. 0.06 N.134

0
SCORE =0.11 RSP=14%
ERROR= 0.01 N=644

14L

SCORE =0.74 RSP=52%
ERROR= 0.06 N=187

0
SCORE =0.07 RSP=7%
ERROR= 0.04 N=41

12L

SCORE =2.5 R$F'.yf.%
ERROR. 0.06 N.211

0

SCORE =0.63 RSP=39%
ERROR= 0.10 N=76

10L

SCORE =2.5 RSP=97%
ERROR= 0.08 N.123

\~2 estival

inter

1 1 rosa

SCORE = MEAN SCORE RSP = % WITH INTERMEDIATE OR FULL RESPONSE

ERROR = STANDARD ERROR (SCORE) N = SAMPLE SIZE

Figure 3. Summary of all temperature and daylength regimes: The extent and
intensity of response for each of the nine controlled temperature and
daylength regimes as shown by proportionately shaded pie charts. The
percentage of Linea (score = 0) individuals is shown by the unshaded
portions, intermediate individuals (score = 1 or 2) by the crosshatched
portions and Rosa individuals (score = 3 or 4) by the completely filled
portions. The top row shows response to different temperatures at the
breeding colony daylength of 1 6L:8D while the rightmost column shows the
response to different daylengths at the breeding colony temperature of
27°C. Below each piechart is additional information for that treatment:
mean score and standard error, % response and sample size.

the effects of these two factors impossible. In order to study the effects of
temperature and daylength separately, each was manipulated indepen-
dently under controlled conditions in laboratory incubators. The results
of all such manipulations are summarized in Figure 3.

The mean score and % response of animals subjected to progressively
shortened daylengths rose sharply, even when temperature was held at

230

J. Res. Lepid.

0

0

10

12

14

16

Daylength

Figure 4. Response to daylength: The response characteristics of groups exposed
to different daylengths at the breeding colony temperature of 27°C. The
critical daylength (50 % response) lies between 10L:14D and 12L:12D at
27°C.

mean score

% response (% animals with score > 0)

the breeding colony value of 27°C (Figure 4). It was found that the critical
daylength (the daylength at which 50% of the population shows a
response) lies between 10L:14D and 12L:12D at 27°C. A comparable
increase in mean score and % response was noted in response to lowered
temperatures, even when daylengths were held at the breeding colony
value of 16L:8D (Figure 5). In this case, the critical temperature was
found to lie between 21°C and 24°C at 16L:8D.

Two balanced 2X2 matrices were constructed from the data in Figure
3: one including daylengths 14L:10D and 16L:8D with temperatures
24°C and 27°C, another including daylengths 12L:12D and 16L:8D with
temperatures 21°C and 27°C. Both matrices were subjected to 2 X 2
ANOVA analysis and each revealed highly significant temperature,
daylength and temperature-daylength interactive effects (P < 0.001).

RECIPROCAL TRANSFER STUDIES

Reciprocal transfers of larvae at various developmental stages were
carried out in order to investigate the critical period(s) for the induction
of the rosa morph. Larvae were staged into one of five developmental
classes: first instar, second instar, fifth (terminal) instar, pre-pupation

30(3-4):225-236, 1991

231

4

1.0

0.8

c

<1)

3

2

0

0

18

21

24

27

Daylength

Figure 5. Response to temperature: The response characteristics of groups exposed
to different temperatures at the breeding colony daylength of 1 6L:8D. The
critical temperature (50 % response) lies between 18°C and 21 °C at
16L:8D.

mean score

% response (% animals with score >0)

and post-pupation. The pre-pupation class included those animals that
were hanging by their anal prolegs in preparation for pupation ( 18-24
hours prior to onset of pupation). The post-pupation class included those
animals in which pupation had occurred, but sclerotization of the pupal
casing was not yet complete (approximately 1 hour after onset of pupa-
tion).

In one series of transfers, animals were removed from breeding colony
conditions (27°C, 16L:8D) and subsequently reared under conditions
known from previous experiments to produce a high incidence of the rosa
morph (21°C, 12L:12D). The results of these transfers are summarized
in Figure 6.

Animals transferred early in larval life (first instar) showed response
characteristics very similar to those whose entire life was spent at 21°C,
12L:12D, indicating that the presence of temperature and daylength
cues during egg and early larval stages of life is not crucial to the
development of the rosa morph. Animals transferred in the interval near
pupation - from 24 hours prior to pupation (fifth instar) to 1 hour post-
pupation - showed a clear response to such conditions. However, the
response was intermediate between the two untransferred control popu-

232

J. Res. Lepid.

3

2.5

2

Q)

</) 1.5

c

03

0)

1

0.5

0

IND 1st 2nd 5th BP PP NON
Stage at Transfer

Figure 6. Response to transfer from conditions of the breeding colony: Response
characteristics of animals transferred at various stages of development
from the conditions of the breeding colony (27°C, 16L:8D) to conditions
know to produce a high incidence and degree of Rosa coloration (21 °C,
12L12D). Individuals transferred during the first instar (1ST) show a
response comparable to the untransferred control population under con-
stant conditions of 21 °C, 1 2L:1 2D (IND). Individuals transferred during the
fifth instar (5TH) show a response comparable to those transferred before
pupation (BP) and post-pupation (PP). These are in turn intermediate
between the untransferred breeding colony (NON) and 21 °C, 12L:12D
(IND) control groups, indicating both a cumulative process of Rosa induc-
tion and a heightened sensitivity to conditions near pupation.

mean score

% response (% animals with score >0)

lations. The fact that their response characteristics are significantly
lower than that of animals reared at 21°C, 12L:12D for most or all of
larval life suggests that larval conditions exert a cumulative effect. On
the other hand, the fact that their response characteristics are signifi-
cantly higher than that of animals raised exclusively under the condi-
tions of the breeding colony suggests a relatively high sensitivity to
conditions at or near pupation.

A series of reciprocal transfers was also conducted in which individuals
reared at 21°C and 12L:12D were transferred to the conditions of the
breeding colony (27°C, 16L:8D) at various stages of development. The
results of these transfers are summarized in Figure 7.

Again, animals transferred to the conditions of the breeding colony
very early in larval life (first instar) exhibited response characteristics

30(3-4):225-236, 1991

233

1.0

0.8

0.6

0.4

0.2

Figure 7. Response to transfer to conditions of the breeding colony: Response
characteristics of animals transferred at various stages of development
from conditions know to produce a high incidence and degree of Rosa
coloration (21 °C, 1 21:1 2D) to the conditions of the breeding colony (27°C,
16L:8D). Individuals transferred during the first instar (1ST) show a re-
sponse comparable to the untransferred breeding colony control at 27°C,
18L:8D (NON). Individuals transferred during the second instar (2ND)
retain some residual response as compared to the breeding colony control
(NON) despite being exposed to conditions of 21 °C, 1 2L:1 2D for only a few
days longer than first instar transfers (1 ST). Individuals transferred during
the fifth instar (5TH), before pupation (BP) and post-pupation (PP) show a
progressively stronger retention of the Rosa response. Note the divergence
in response between individuals transferred before pupation (BP) and
those transferred post-pupation (PP), despite a difference in exposure to
the 21 °C, 12L12D conditions of only 19-25 hours. This indicates a rapid
“fixation” of the Rosa response near pupation. Moreover, the retention of
response by post-pupation transfers (PP) at a level comparable to the
untransferred 21 °C, 1 2L:1 2D control population (IND) indicates a complete
fixation of the response within an hour of pupation.

mean score

% response (% animals with score >0)

indistinguishable from the breeding colony control population. However,
those transferred only a few days later (second instar) seem to retain a
slightly heightened response. Animals transferred later show progres-
sively stronger retention of response characteristics, with those trans-
ferred in the post-pupation stage retaining response characteristics
indistinguishable from the untransferred control population at 21°C,
12L:12D. Of special interest is the fact that the post-pupation transfers

234

J. Res. Lepid.

show markedly stronger response characteristics than animals trans-
ferred immediately before pupation, even though an interval of only 19-
25 hours separates the two stages.

Discussion

Short daylengths have been shown to trigger the appearance of the rosa
morph in Precis coenia. This is not surprising since many, if not most,
insects respond in some fashion to photoperiodic cues (Beck, 1980;
Tauber et. al, 1986). In particular, photoperiodic cues are thought to
predominate in the polyphenisms of four families of Lepidoptera, includ-
ing the Nymphalidae (Shapiro, 1976; Brakefield & Larsen, 1984).

Low rearing temperatures alone seem to be as effective as short
daylengths in producing the rosa morph. While temperature is a common
modifier of photoperiodic induction, examples of systems with true
multiple induction are relatively rare. In particular, although multiple
induction systems have been shown to exist in other insects (e.g. , aphids),
the only other clearly documented example among the Lepidoptera is by
Shapiro (1982) - although there is also some evidence that the related
species Precis octavia and Junonia villida may have similar induction
patterns (McLeod, 1968; James, 1987).

The reason for multiple induction in the rosa system is unclear. In
general, it is thought that photoperiodic cues are more reliable indicators
of seasonal change than highly variable temperatures. Thus, it is often
argued that only in areas where reliable photoperiodic cues are unavail-
able (e.g. , equatorial regions and foggy montane habitats) will reliance on
temperature cues evolve (Beck, 1980; Shapiro, 1984; Tauber, et. al.,
1986). For example, as an equatorial species P. octavia is exposed to a
daylength with little seasonal variation, so the appearance of a tempera-
ture-regulated mechanism for the polyphenism is easily explained. But
P. coenia is a temperate species and is subjected to seasonal daylength
fluctuations of more than 3 hours in North Carolina - ample range for the
development of a predominantly photoperiodic induction if such a cue
were inherently superior to temperature.

Shapiro (1978) has suggested that seasonally polyphonic systems in
general should tend to have redundant induction mechanisms. If so, the
dearth of examples in the literature may be due to an overzealous focus
on the importance of photoperiodic cues. On the other hand, Shapiro has
also argued (1976) that seasonal polyphenisms are character traits
which represent a high degree of phylogenetic information. If it turns out
that multiple induction of seasonal polyphenisms is relatively rare, then
perhaps the rosa system represents an early stage in the evolutionary
transition from temperature to photoperiodic control in a species with a
tropical ancestry.

The reciprocal transfer studies seem to indicate that the induction of
the rosa morph is a cumulative process beginning early in larval life,
accelerating during the terminal larval instar and culminating within 1

30(3-4):225-236, 1991

235

hour of the onset of pupation. The protracted nature of rosa induction is
unusual since the majority of insects respond to seasonal cues only
during a tightly defined critical period, although there other are ex-
amples of cumulative induction patterns - as in Pectinophora gossypiella
and Ostrinia nubilalis, as well as in the pierid butterflies Pieris rapae and
Pieris brassicae (Beck, 1962; Barker, et. al., 1963; Shapiro, 1968).

The complexity of the rosa polyphenism may help explain why elucida-
tion of its seasonal cues has so long eluded researchers. Although the
present study solves this particular puzzle, it raises many new ones to
take its place. Obvious questions remain concerning the exact physiology
of the system and these are the focus of ongoing research. Perhaps most
intriguing however, are questions concerning the evolutionary ecology of
rosa. There is interesting evidence that the rosa system has a strong
genetic basis and exhibits clinal variation in the wild (Smith, 1991) but
the adaptive significance (or insignificance) of the rosa morph is un-
known, as are many other parameters of evolutionary importance such
as migration and gene flow patterns. Klots’ (1951) observation concern-
ing the Buckeye is just as true today as it was 40 years ago: “A great deal
of careful work and thorough analytical study is needed.”

Acknowledgments. I am greatly indebted to Fred Nijhout of the Duke University
Zoology department in whose lab this study was carried out and without whose
support its completion would not have been possible. Special thanks also to
Arthur Shapiro as well as Robert Brandon and Debbie Roundtree for reading
and commenting on early drafts.

Literature Cited

Balciunas, J. and K. Knopf (1977) Orientation and flight speeds and tracks of
three species of migrating butterflies. The Florida Entomologist. 60(1):37-
39

Barker, J. F. and W. S. Herman ( 1976) Effect of photoperiod and temperature on
reproduction of the monarch butterfly, Danaus plexippus. Journal of Insect
Physiology. 22:1565-1568.

Beck, S.D. (1962) Photoperiodic induction of diapause in an insect. Biology
Bulletin. 122:1-12.

(1980) Insect Photoperiodism (2nd edition). Academic Press. New York.

Brakefield, P. M. & T. B. Larsen ( 1984) The evolutionary significance of dry and
wet season forms in some tropical butterflies. Biological Journal of the
Linnean Society. 22:1-12.

Clark, A. H. (1932) Butterflies of the District of Columbia & vicinity. United
States National Museum Bulletin. 157, 327pp.

Ferris, C. D. & F. M. Brown (1980) Butterflies of the Rocky Mountain States.

University of Oklahoma Press. Norman.

Howe, W. H. (1975) The Butterflies of North America. Doubleday & Co., Inc.
Garden City.

James, D. G. ( 1987) Effects of temperature and photoperiod on the development
of Vanessa kershawi McCoy and Junonia villida Godart (Lepidoptera:
Nymphalidae). Journal of the Australian Entomological Society. 26:289-92.

236

J. Res. Lepid.

Kidd, N. A. C. (1979) The control of seasonal changes in the pigmentation of the
lime aphid nymphs, Eucallipterus tiliae. Entomological Experimental
Applications. 25:31-38.

Kingsolver, J. G. (1987) Evolution and coadaptation of thermoregulatory
behavior and wing pigmentation in pierid butterflies. Evolution. 41(3):472-
490.

Klots, A. B. (1951) A Field Guide to the Butterflies of North America, East of
the Great Plains. Houghton Miflin Co. Boston.

Leigh, T. F. & R. F. Smith (1959) Flight activity of Colias philodice eurytheme
Boisduval in response to its physical environment. Hilgardia. 28:569-624.

Mather, B. (1967) Variation in Junonia coenia in Mississippi (Nymphalidae).
Journal of the Lepidopterist Society. 21(l):59-70.

McLeod, L. (1968) Controlled environmental experiments with Precis octavia
cram. Journal of Research on the Lepididoptera. 7(1):1-18.

Opler, P. A. & Krizek, G. O. (1984) Butterflies East of the Great Plains. Johns
Hopkins University Press. Baltimore.

Shapiro, A. M. (1968) Photoperiodic induction of vernal phenotype in Pieris
protodice Boisduval and Le Conte (Lepidoptera: Nymphalidae). The Wasmann
Journal of Biology. 26(1):137-149.

(1976) Seasonal polyphenism. in Evolutionary Biology. M. Hecht, W. C.

Steere and B. Wallace (eds.) Vol. 9. Plenum Press. New York.

(1978) The evolutionary significance of redundancy and variability in

phenotypic-induction mechanisms of pierid butterflies (Lepidoptera). Psyche.
85(2-3):275-283.

(1982) Redundancy in pierid polyphenisms: pupal chilling induces

vernal phenotype in Pieris occidentalis (Pieridae). Journal of the
Lepidopterists’ Society. 36(3): 174-77.

(1984) Experimental studies of the evolution of polyphenism. in The

Biology of Butterflies. Academic Press. London.

Smith, K. C. (1991) The effects of temperature and photoperiod on seasonal
polyphenism in Precis coenia (Lepidoptera: Nymphalidae). A thesis submitted
in partial fulfillment of the requirements for the degree of Master of Science
in the Department of Zoology in the Graduate School of Duke University.

Tauber, M. T. , C. A. Tauber and S. Masaki (1986) Seasonal Adaptations of
Insects. Oxford University Press. New York.

Walker, T. J. & A. J. Riordan (1981) Butterfly migration: are synoptic-scale
wind systems important? Ecological Entomology. 6:433-440.

Walker, T. J. (1991) Butterfly migration from and to peninsular Florida.
Ecological Entomology. 16:433-440.

Watt, W. B. (1968) Adaptive significance of pigment polyphenisms in Colias
butterflies. I. Variation of melanin pigment in relation to thermoregulation.
Evolution. 22(3):437-458.

Journal of Research on the Lepidoptera

30(3-4):237-244, 1991

Three New Taxa of Calephelis from Costa Rica
(Lycaenidae: Riodininae)

George T. Austin

Nevada State Museum and Historical Society, 700 Twin Lakes Drive, Las Vegas, Nevada
89107

Abstract. Two new species and one new subspecies of Calephelis
(Lycaenidae: Riodininae) are named from Costa Rica.

Introduction

Calephelis Grote and Robinson, 1869 is a large riodinine genus
(Lycaenidae); its species are typically difficult to determine. The subtle
differences between taxa in size, wing shape, pattern, and genitalia are
recognized only after considerable study and frustration. The revision of
the group by McAlpine (1971) was a start in elucidating their diversity
but has little comparative information. The species limits of the nearly
forty putative species are debatable (e.g., Scott 1986) and may require
detailed studies of the early stages or genetic data to resolve. Nonethe-
less, local faunas, at least, seem to have constant genitalia and, taking
into account possible seasonal variation, pattern and color on the wings.

While examining the Calephelis for a treatment of the Costa Rica
riodinine fauna (DeVries, in prep.), I encountered three phenotypes
which have not been previously recognized. Because of a general disin-
clination of workers to deal with the genus and the need of names for the
above mentioned work, I describe these herein to bring them to the
attention of future taxonomists.

Calephelis sodalis new species
Figs. 1, 2 (male), 3, 4 (female), 13 (male genitalia)

Description. Male. Forewing length = 11.6 mm (11.0-12.3, N = 6;
holotype = 11.2). Dorsum dark grayish red-brown; basal 2/3 of both wings
with five relatively indistinct, broken, black lines from costa to inner
margin, the outer broader than others and shaded slightly proximally
with darker brown; marginal and submarginal lines distinct, iridescent
blue-gray, the outer thin and more-or-less parallel to outer margin,
narrowly broken at veins, the inner broader, composed of irregular
scrawls or crescents, broken and often disjunct at veins, produced
distally between veins M2 and CuA} on both wings; row of rather large
black dots between iridescent lines; area distal to black dots red-brown,
lacking gray cast of remainder of wing; fringes checkered about equally
with whitish and dark gray.

Ventral surface bright orange-brown with markings of dorsum re-
peated; basal black lines very fine, segments often obsolete in some cells;
outer black line usually single (occasionally trace of doubling on more

238

J. Res. Lepid.

8

30(3-4):237-244, 1991

239

Facing Page:

Figure 1 . Calephelis sodalis Austin - holotype male, dorsal surface. Data in text.
Figure 2. Calephelis sodalis - holotype male, ventral surface.

Figure 3. Calephelis sodalis - paratype female, dorsal surface. COSTA RICA:

Puntarenas Prov.; Las Alturas, Tajo Rd. & pasture top, 8 Sept. 1990.
Figure 4. Calephelis sodalis - paratype female, ventral surface. Same specimen as
Fig. 3.

Figure 5. Calephelis exiguus Austin - holotype male, dorsal surface. Data in text.
Figure 6. Calephelis exiguus - holotype male, ventral surface.

Figure 7. Calephelis exiguus - paratype female, dorsal surface. COSTA RICA: Limon
Prov.; Sixaola Rd., 14 km SE Bribri, 3 Sept. 1987.

Figure 8. Calephelis exiguus - paratype female, ventral surface. Same specimen as
Fig. 7.

Figure 9. Calephelis laverna parva Austin - holotype male, dorsal surface. Data in
text.

Figure 10. Calephelis laverna parva - holotype male, ventral surface.

Figure 1 1 . Calephelis laverna parva- paratype female, dorsal surface. COSTA RICA:

Puntarenas Prov.; Osa Peninsula, 2.5 mi SW of Rincon, 1-7 Mar. 1967.
Figure 12. Calephelis laverna parva - paratype female, ventral surface. Same
specimen as Fig. 1 1 .

240

J. Res. Lepid .

Figure 13. Calephelis sodalis - holotype male genitalia, ventral view.
Figure 14. Calephelis exiguus - holotype male genitalia, ventral view.

heavily marked individuals); iridescent lines broader and greener than
on dorsum, inner less irregular and in form of bars especially on forewing.
No seasonal variation among individuals seen.

Head, thorax, and abdomen concolorous with wings above, paler than
ground color beneath; antennae black with distinct white annular rings,
club often whitish on outside edge, tip orange.

Genitalia: valvae short and relatively broad in lateral view, evenly, but
not broadly, curved in ventral view; transtilla exceeding valva length,
narrow and curved dorsad at the posterior end, both pairs of lateral
processes usually well developed.

Female. Forewing length = 12.5 mm (11.9-12.8, N = 7). Similar to male;
dorsum less dark especially basally; median dark band more prominent;
venter more ochraceous with all lines more prominent.

Types. Holotype male with the following labels: white, printed and
handprinted - Costa Rica Puntarenas / Las Alturas 1400m / Tajo Rd &
Bella Vista / 26 May 1991 / time 10:30 / P. J. DeVries; white, printed and
handprinted - Genitalic Vial / GTA - 1880; red, printed - HOLOTYPE /
Calephelis sodalis / Austin. Eleven paratypes (all COSTA RICA:
Puntarenas Province, leg. P. J. DeVries): Las Alturas, 1600 m, Tajo Bella
vista, 25 Aug. 1990 (3 males), 17 Aug. 1990 (1 female), 8 Sept. 1990 (1
male), 5 June 1991 (1 male); Tajo Rd. & pasture top, 1500 m, 8 Sept. 1990
(2 males, 3 females).

Deposition of types. The holotype and one female paratype will be
deposited at the Allyn Museum of Entomology. One pair of paratypes will
be retained by the author and the remaining paratypes will be returned
to Philip J. DeVries.

30(3-4):237-244, 1991

241

Type locality. COSTA RICA: Puntarenas Province; Las Alturas,
elevation 1400-1600 m. Las Alturas is on the Pacific slope of southern
Costa Rica, northwest of San Vito at about 8 km west of the Panama
border.

Distribution and phenology. This species is known only from the
type locality on the Pacific slope of southeastern Costa Rica at moderate
elevations. There evidently are at least two broods with specimens
known from May (fresh), June (worn), August (worn, fresh), and Septem-
ber (worn, fresh). No other Calephelis are known from the vicinity of the
type locality of C. sodalis.

Etymology. The species is known from the Las Alturas Biological
Station at the border of Parque Nacional la Amistad, a joint venture of
Costa Rica and Panama. The name refers to comradship or friendship in
Latin.

Diagnosis and discussion. This species is nearly impossible to
distinguish from Calephelis browni McAlpine superficially. Calephelis
sodalis is slightly larger than C. browni and the basal markings on the
ventral wings are less well defined on the average. The genitalia are,
however, distinctive. The valvae of C. browni are longer and proportion-
ally narrower in lateral view than those of C. sodalis , are less evenly
curved in ventral view, and there is less space between them anteriorly.
The transtilla of C. browni is shorter than the valvae, broader at the
posterior end, and the anterior processes are often poorly developed. The
male genitalia are similar to those of Calephelis dreisbachi McAlpine,
Calephelis montezuma McAlpine, and Calephelis azteca McAlpine, all
known only from Mexico, differing in minor details. The pattern and color
of the wings are different. Other small to medium-sized species of the
genus in Costa Rica with the transtilla exceeding the valvae in length are
Calephelis schausi McAlpine on which the transtilla greatly exceeds the
valvae, the wings are more rounded, and the maculation is quite differ-
ent; Calephelis inca McAlpine which is very small and has a paler and
brighter dorsal coloration; and Calephelis laverna (Godman and Salvin)
which again is smaller and has very different short and stout valvae.

Calephelis exiguus new species
Figs. 5, 6 (male), 7, 8 (female), 14 (male genitalia)

Description. Male. Forewing length = 10.0 mm (9.7-10.1, N = 4;
holotype = 10.1). Dorsum dark red-brown; basal 2/3 of both wings with
five indistinct, broken, black lines from costa to inner margin, the outer
shaded very slightly proximally with darker red-brown; marginal and
submarginal lines indistinct, iridescent blue-gray, both thin, the outer
more-or-less parallel to outer margin, narrowly broken at veins particu-
larly on hindwing, the inner sinuous, produced distally between veins M.;
and CuAj on both wings, broken at veins; row of rather large black dots
between iridescent lines; fringes dark gray checkered with white.

Ventral surface dull orange-brown with markings of dorsum repeated;
basal black lines very fine, broken at veins; outer black line single;
iridescent lines broader and more prominent than on dorsum.

242

J. Res. Lepid.

Head, thorax, and abdomen concolorous with wings above, paler than
ground color beneath; antennae black with narrow white annular rings,
club with orange tip.

Genitalia: valvae in lateral view relatively narrow and curving ven-
trally throughout their length, rather square anteriorly before curving
inward posteriorly in ventral view, broadly rectangular space between;
transtilla equal to valva length, very narrow and evenly tapered poste-
riorly, very slightly curved dorsad, posterior lateral processes well
developed, anterior small.

Female. Fore wing length = 9.7, 11.2. Similar to male; wings much more
rounded; ventral iridescent lines broader.

Types. Holotype male with the following labels: white, printed -
COSTA RICA / Limon Province / Sixaola Road / Paraiso / 3 Sept. 1987 /
leg. G&A Austin; white, printed and handprinted - Genitalic Vial / GTA
-1523; red, printed - HOLOTYPE / Calephelis exiguus / Austin. Five
paratypes (all COSTA RICA: Limon Province, leg. G&A Austin): same
data as the holotype (1 male); La Bomba, 2 Sept. 1987 (1 male, 1 female);
Rio Blanco Road, Rio Victoria, 6 Oct. 1987 (1 male); Sixaola Rd., 14 km
SE Bribri, 3 Sept. 1987 (1 female).

Deposition of types. The holotype and one paratype female will be
deposited at the Allyn Museum of Entomology. The remaining paratypes
will be retained for now by the author.

Type locality. COSTA RICA: Limon Province; Sixaola Road, Paraiso,
elevation ca. 50 m. This locality is on the Atlantic slope of easternmost
Costa Rica, less than 2 km north of the Panama border about half way
between Sixaola and the main road to Limon (Ruta 36). The types were
taken from roadside flowers.

Distribution and phenology. The species is known from the types,
all taken below 100 m on the Atlantic slope of eastern Costa Rica
(southeast of Limon) in September and October and a male from Madre
de Dios, Limon Province, taken in July. Two males from Heredia
Province taken in May and June are slightly darker on the dorsum with
more distinct silvered lines. These may prove to be a seasonal phenotype
of C. exiguus.

Two other Calephelis are known from the region inhabited by C.
exiguus: Calephelis sixaola McAlpine and Calephelis hrowni McAlpine.

Etymology. The name means small and refers to the small size of this
species.

Diagnosis and discussion. Superficially, this species resembles a
small Calephelis costaricicola Strand but the median dark band is less
well developed, the ventral ground color is more orange (more ochraceous
on C. costaricicola ), the lines are more poorly developed on the ventral
surface without the doubling and shading of the outer basal line, and the
forewings are less produced. The male valvae are not as broadly rounded
as on C. costaricicola and the transtilla is narrower; C. costaricicola is not
known from the Atlantic slope. Calephelis hrowni is larger, deeper
orange-brown beneath and tends to have a grayish cast distal to the more

30(3-4):237-244, 1991

243

prominent dark median band on the dorsal wings. The male genitalia of
C. browni are more robust than those of C. exiguus and curved ventrad
more abruptly at the posterior end. Of the other small species of the genus
in Costa Rica, C. inca is brighter, more orange-brown on the dorsal
surface and C. laverna is a richer red-brown on the dorsum with more
prominent iridescent lines and has a brighter ventral ground color with
the various lines much better defined. The male genitalia of both C. inca
and C. laverna are very different with the transtilla greatly exceeding the
length of the valvae.

Calephelis laverna parva new subspecies
Figs. 9, 10 (male), 11, 12 (female)

Description. Male. Forewing length = 9.9 mm (9.7-10.2, N = 6;
holotype = 9.9). Dorsum dark red-brown, tending towards brighter red-
brown distally; basal 2/3 of both wings with five relatively distinct,
broken, black lines from costa to inner margin, the outer broader than
others and shaded slightly (or not) proximally with darker brown;
marginal and submarginal lines relatively distinct, iridescent blue-gray,
the outer thin and more-or-less parallel to outer margin, narrowly
broken at veins, the inner broader, broken and somewhat disjunct at
veins, produced distally between veins M2 and CuAx on both wings; row
of rather large black dots between iridescent lines, distinct on hindwing,
distinct or not on forewing; fringes dark gray, checkered narrowly with
whitish.

Ventral surface bright, dark orange-brown with markings of dorsum
repeated; basal black lines distinct; outer black line single; iridescent
lines slightly broader and more blue-green than on dorsum, inner less
irregular and in form of bars especially on forewing. Material from March
is somewhat darker than that from July.

Head, thorax, and abdomen concolorous with wings above, paler than
ground color beneath; antennae black with distinct white annular rings,
club with orange tip.

Genitalia: valvae very short and relatively broad in lateral view; short,
stout, and straight in ventral view; transtilla greatly exceeding length of
valvae, very narrow at posterior end, anterior pair of lateral processes
well developed.

Female. Forewing length = 9.9 mm (9.2-10.6, N = 3). Similar to male;
wings much broader and rounder; dorsum slightly less dark especially
basally; median dark band more prominent; venter ochraceous to
ochraceous orange with all markings more prominent.

Types. Holotype male with the following labels: white, printed and
handprinted - COSTA RICA / Puntarenas Prov. / Osa Peninsula / 2.5 mi.
SW. Rincon / 08° x42’ N. 83° 29’ W. / III - 1 to 7 - 1967 / OTS Adv. Zoo.
Course; white, printed - J. M. Nelson / Collector; white, printed - Downey
colln. / Allyn Museum / Acc. 1985 - 14; white, printed and handprinted -
Genitalic Vial / GTA - 1863; red, printed - HOLOTYPE / Calephelis
laverna / parva Austin. Eight paratypes (all COSTA RICA: Puntarenas

244

J. Res. Lepid.

Province): same data as holotype (1 male, 1 female), Paso Canoa, 160 m,
19 July 1963, leg . L. D. Miller (1 male, 1 female), 20 km N of Palmar Sur,
140 m, 22 July 1963, leg. L. D. Miller (1 female), Parque Nacional
Corcovado, Sirena, 20 m, 6 March 1989, leg. N. Greig(l male), Corcovado,
Sirena, 7 March 1989, leg. N. Greig (1 male), La Vacita, 24 March 1990,
leg. P. J. DeVries (1 male).

Deposition of types. The holotype and a pair of paratypes will be
deposited at the Allyn Museum of Entomology. One male and two female
paratypes will be deposited at the Carnegie Museum of Natural History.
Two male paratypes are in the care of P. J. DeVries and one male
paratype will be retained by the author.

Type locality. COSTA RICA: Puntarenas Province; Osa Peninsula,
2.5 miles southwest of Rincon. This is near sea level in the northeastern
portion of the Osa Peninsula on the Pacific Coast of southern Costa Rica.

Distribution and phenology. This species is known only from the
Pacific slope of southern Costa Rica at low elevations near the coast and
centered on the Osa Peninsula. There evidently are at least two broods
with fresh material from March and July. One other species of the genus,
Calephelis schausi McAlpine, is known within the distribution of C. /.
parva.

Etymology. The name refers to the small size of this subspecies of C.
laverna.

Diagnosis and discussion. The male genitalia of this phenotype are
virtually identical to those of Calephelis laverna laverna (Godman and
Salvin) described from “V. de Chiriqui” (Panama). McAlpine (1971)
apparently viewed the type, additional specimens from the type locality
and nearby, and from South America. The only variation he noted was of
material from Trinidad and adjacent Venezuela which had somewhat
different genitalia and was described as Calephelis laverna trinidadensis .
Specimens of C. laverna which I have seen from Panama and Columbia
are considerably larger (the forewing length averaging about 12 mm), are
distinctly more reddish above, and thus the various markings are much
more distinct.

Acknowledgments. I thank the curators and others at various museums for the
opportunity to examine specimens in their care: L. D. and J. Y. Miller (Allyn
Museum of Entomology), J. Rawlins (Carnegie Museum of Natural History), F.
Rindge and G. Martinez (American Museum of Natural History), and S. Borkin
(Milwaukee Public Museum). P. J. DeVries kindly loaned material in his care
and made numerous suggestions for improvement of the manuscript.

Literature Cited

McAlpine, W. S. 1971. A revision of the butterfly genus Calephelis (Riodinidae).
Jour. Res. Lepid. 10: 1-125.

Scott, J. A. 1986. The butterflies of North America, a natural history and field
guide. Stanford University Press, Stanford, CA.

Journal of Research on the Lepidoptera

30(3-4):245-247, 1991

A New Species of Calephelis from Guatemala
(Lycaenidae: Riodininae)

George T. Austin

Nevada State Museum and Historical Society, 700 Twin Lakes Drive, Las Vegas, Nevada
89107

Abstract. A new species of Calephelis (Riodininae) is named from
northern Guatemala.

Introduction.

McAlpine (1971) suggested that Guatemala may be the center of
distribution of the genus Calephelis (Lycaenidae: Riodininae). During
field work at Parque Nacional Tikal in Peten, northern Guatemala, in
February 1992, a striking new species was taken. This is named and
described below.

Calephelis tikal new species
Figs. 1, 2 (male), 3 (male genitalia)

Description. Male. Fore wing length = 11.6 mm (holotype), 11.8 mm
(paratype). Dorsum dark brown with prominent gray cast especially on
outer third of both wings; wing bases with five indistinct, concentric,
black lines, more-or-less continuous from costa to inner margin, the outer
very broad on both wings, extending basad nearly to next line; marginal
and submarginal lines faint (especially on forewing), iridescent blue-
gray, the outer thin, parallel to outer margin, broken slightly at veins, the
inner somewhat irregular, broken and disjunct (on forewing) at veins;
row of rather large, but indistinct, black dots between iridescent lines;
fringes gray-brown, paler than ground color.

Ventral surface dark red-brown with black overscaling, especially
distally; markings of dorsum repeated, more prominent; outer basal line
doubled throughout length on both wings, area between dusted with
black anteriorly on forewing; iridescent lines broader, prominent.

Head, thorax, and abdomen concolorous with wings above and (includ-
ing legs) below; antennae black with white annular rings, tip of club
yellow-orange.

Genitalia: valvae of moderate length, broad in lateral view, robust and
broad in ventral view; transtilla slightly longer than valvae, relatively
broad, curved dorsad to a slight hook.

Female. Unknown.

Types. Holotype male with the following labels: white, printed -
GUATEMALA / Peten, Parque / Nacional Tikal / #6, 12:15-12:45 / 2
February 1992 / leg. G. T. Austin; white, printed and handprinted -
Genitalia Vial / GTA - 2385; red, printed - HOLOTYPE / Calephelis tikal
/ Austin. One paratype male with same location and collector as holotype,
6 February 1992.

246

J. Res. Lepid.

Figure 1 . Calephelis tikal Austin - holotype male, dorsal surface. Data in text.
Figure 2. Calephelis tikal - holotype male, ventral surface.

Deposition of types. The holotype will be deposited at the Allyn
Museum of Entomology, Sarasota, Florida. The paratype will be re-
tained for now by the author.

Type locality. GUATEMALA: Peten; Parque Nacional Tikal, 200 m.
The types were taken along the side of the main road, just south of the
ruins at Tikal.

Distribution and phenology. This species is known at present only
from the two types taken in February.

Etymology. The species is named after its type locality, the ruins of
the Mayan city of Tikal.

Diagnosis and discussion. This is the darkest of the known Calephelis,
so dark that the initial impression is that of a Charts. The male genitalia

30(3-4):245-247, 1991

247

most closely resemble those of Calephelis acapulcoensis McAlpine with
the transtilla slightly exceeding the length of the valvae. The valvae of
C. tikal , however, are broader and less curved. The valvae are similarly
broader than those of Calephelis azteca McAlpine and Calephelis
guatemala McAlpine and less curved than those of the latter and
Calephelis yucatana McAlpine.

Acknowledgments. My studies at Tikal were made possible through a MacArthur
Foundation grant to P. R. Ehrlich via the Center for Conservation Biology at
Stanford University, Palo Alto, California. I thank the center’s staff biologists
and field companions, N. Haddad, A. Launer, D. Murphy, and T. Sisk, for their
assistance. I also acknowledge the various agencies of the Guatemalan
government which graciously issued the necessary permits.

Literature Cited

McAlpine, W. S. 1971. A revision of the butterfly genus Calephelis (Riodinidae).
Jour. Res. Lepid. 10:1-125.

Journal of Research on the Lepidoptera

30(3-4):248-256, 1991

Host specialization of satyrine butterflies, and their
responses to habitat fragmentation in Trinidad

M.C. Singer1 and P.R. Ehrlich

Department of Biological Sciences, Stanford University, Stanford, California 94305

Introduction

Butterflies may be among the most useful indicators of habitat change
(Ehrlich & Murphy 1987, Kremen 1992). In this paper we describe the
communities of grass- and sedge-feeding Satyrinae (Nymphalidae) but-
terflies from ten study sites in Northern Trinidad in 1970-74, and assess
the extent to which species may have been lost as a result of human
modification of the landscape. In undisturbed habitats such as still exist
in some neotropical countries, satyrines move through the forest, feeding
as adults on fallen fruit, and ovipositing on patches of grass or sedge that
grow at treefalls or beside streams (DeVries 1985;1987). Although the
patches of larval habitat may be widely scattered, there are few obstacles
to the movement of adults. In contrast, human disturbance increases
host plant abundance and decreases patchiness, but creates obstacles to
the movement of adults, most of which will not fly through open areas.
Thus, the clearing of tropical forests in Northern Trinidad has resulted
in considerable fragmentation of habitat for these shade-loving species.
The resultant reduction in population sizes could cause local extinction
of specialist insects from the habitats that still contain their hosts. We
seek evidence for such local extinction by comparing the distributions of
these insects with the fragmented distributions of their hosts. We also
compare our data with historical records gathered by Barcant (1970) to
ask whether there is evidence for recent extinction of either specialist or
generalist species from the Island of Trinidad.

As a prerequisite for this work we needed to classify the study species
as either host specialists or host generalists, since no information on
their diets was previously available. We also obtained information on
host plant species richness and abundance, and asked whether either of
these traits was correlated across habitats with either butterfly species
richness or butterfly abundance.

Study Species

The generic status of many of the neotropical Satyrinae is unclear (see
discussion in DeVries 1987). For the purposes of this paper we have
placed all satyrines studied here in the genus Cissia since many of them,
including palladia , terrestris, myncea, libye, themis and penelope, are
very likely to belong in this genus (Singer et al. 1983; DeVries 1987).

Satyrine butterflies are known to feed on various monocotyledonous
plants including palms and Marantaceae as well as grasses and sedges

Present address: Department of Zoology, University of Texas, Austin, Texas 78712

30(3-4):248-256, 1991

249

(DeVries 1987). A small group (the genus Euptychia) has colonized lower
plants, feeding on Selaginella (Singer et al. 1983) or on epiphytic mosses
(Singer & Mallet 1975). However, the Trinidadian species we studied
were restricted to grasses and sedges in their diets. Observations of
oviposition on other plants, including Selaginella , turned out to be
examples of oviposition away from the host plant by insects whose larvae
were unable to feed on the plants that actually received eggs (Singer et
al. 1971).

Satyrine butterflies are, to varying degrees, shade-loving insects.
Among our study insects, arnaea, myncea and junta were the most shade-
loving, while hermes, themis and penelope were the least restricted to
deep shade. This trait is important, since the most shade-loving species
should be the least able to colonize small habitat patches. In the
fragmented landscape of Northern Trinidad, such colonization would
usually require crossing open areas inimical to these insects.

Study Sites

Our study sites were all situated in Northern Trinidad. All but two,
“Trace” and “Trace Plantation,” were close to sea level. “Trace” was an ill-
kept trail along a ridge-top through montane rain forest at about 800 m
elevation. “Trace Plantation” was an abandoned, heavily-shaded and
overgrown cacao plantation at the same elevation. “Guanapo” was a
cultivated flat with bananas, cocoa, and coffee. “Dump 1” and “Dump 2”
were adjacent sites close to Guanapo town dump. “Dumpl” was entirely
second growth, with few shrubs more than ten feet high, while “Dump 2”
was less recently disturbed and quite heavily shaded. “Cave” was
another abandoned and overgrown Cacao plantation, at much lower
elevation than “Trace Plantation”. The remaining four sites were all in
the city of Port of Spain in a well-preserved patch of dry forest just North
of the Zoo. The sites coded as “POS 1, POS 2, and POS 3” were very close
together, separated from each other by uncensused strips only about 50
m wide that also contained butterflies.

With the exception of “Trace”, which was an elongated study site
bordering a trail, all of our study sites were approximately rectangular.
Most were on the order of 10,000m2 in area, but “Cave”, and the three
“POS” sites were about half this size.

Habitat fragmentation in Northern Trinidad was already extensive at
the time of this study (1970-74), and no substantial patches of forest
remained in our study area. The two “dump” sites were each separately
cut off from suitable habitat, as were the three “POS” sites as a group.
Only the least shade-restricted species would now have easy access to
these sites. The “Guanapo” site was surrounded by habitat that was not
totally unsuitable but sufficiently open that the more shade-loving
species were reluctant to travel through it. All other sites were at least
partially connected to shaded habitat through which adult insects could
travel.

250

J. Res. Lepid.

Methods

Standard counts of adult butterflies were performed in timed searches of the
habitats (cf Pollard (1977). No habitat was sampled twice in any morning or
afternoon, but some were sampled twice in the same day. No habitat was sampled
less than three times in all, and each habitat was sampled at least once in both
wet and dry seasons.

The census of the grasses was done by counting the number of species present
and assigning each habitat to a category of overall grass abundance ranging from
1 (grasses averaging less than one ramet per 10 m2 and nowhere abundant) to 5
(grasses providing almost complete ground cover).

Host specificity of the insects was investigated by several methods. For
abundant species, we obtained direct evidence by observing ovipositions in the
field, and by finding eggs and larvae and recording their hosts. For three of the
rare species, we were not able to find early stages, so had to resort to indirect
methods. These were:

1) Oviposition preference trials on captive adults. Insects were deprived of
opportunity to oviposit for 24 hours. They were then offered sequential trials with
each species of grass or sedge from their natural habitats. At each trial, the insect
was placed gently on the plant and its response (oviposition or not) recorded. The
range of plants accepted was classified as the host range of the butterfly. These
crude trials are likely to overestimate the natural host range because of the high
oviposition motivation of tested insects (Singer 1986). One species (arnaea) was
not amenable to the trials, failing to duplicate normal oviposition behavior after
manipulation. We obtained data from this species by holding females captive
until oviposition motivation was high, then releasing them and observing
acceptances and rejections of plants they encountered.

2) Feeding-preference trials on captive larvae. Larvae were placed in Petri
dishes with three or four potential hosts from their natural habitats. Each plant
species that was fed on was classified as part of the host range.

Figure 1 : Cissia census data and host-plant distribution among ten study sites.

Sites are listed from top to bottom, while butterfly species are listed across the figure
in the upper section (a) and plant species in the lower section (b). Figuire
1 a shows the numbers of each butterfly species censused at each site. For
example, the census at “Trace” comprised 2 individual palladia, 6 myncea,
196 hermes, etc. Generalist feeders are shown at the left of the figure,
separated by vertical lines from the three host specialists at the right.

Figure 1b shows the presence or absence of each plant species. Each species is
indicated by a number; the names, where known, are given below. The
figure is arranged as far as possible with the plant species vertically beneath
the insects which feed on them. The host specialists at the right of Figure
1 a, arnaea, junia and erichto , are each arranged above their respective
hosts, grass species 13, 14 and 15. The generalists, palladia,

terrestris themis, penelope, are arranged above those hosts that are

edible to them, species 1 through 13 inclusive. Tentative plant identifica-
tions are as follows: 1) Lasiacis sloanei 2) Panicum sp. 3) Trypsacum
sp. 4) Setariapaniculifera 5) Paspalum sp. 6) Paspalum conjugatum 7)
Paspalum sp. 8) Paspalum decumbens 9) Panicum pilosum 1 0)
Cyperus sp. 11) Scleria sp. 12) Panicum polygonatum. 13) Ichnanthus
pallens 14) probably Panicum maximum 15) Unidentified sedge. 16)
Unidentified grass 17) Unidentified grass 18) Bambusa vulgaris

BUTTERFLY SPECIES

30(3-4):248-256, 1991

251

Vj

-i

V

%*

<V"

&

%%

^o

y(SL

<SK

% +o

% -

V

to CM

LO t-

O CM O

CD LO LO O 00 T- 00 O 1^'

05 t- LO ■«- T_ t-

N 05 00 LO

CM i-

N CM LO i —

00 CM CM

CD CM CM 00

i- CM

CO CM CM

r^~ -i— cm

CM LO CD
-M-

LO CM LO

CM 05 t CM

co

CL

O

Q.

CO

c

CM

—

CM

CO

CO

t

LU

O

05

05

o=

CL

05

>

CO

CO

CD

CO

c

111

2

§

CO

13

E

33

E

3

b

6

o

<

CL

(/)

H

1—

CD

o

o

O

CL

CL

CL

CO

1-

BUS

(/)

• • •

• •

03

c

_co

CL

05 05

O O

co co

0

Q- 1- CM

1 g- g- 05

co E E >

3 3 D (C

CD Q Cl O

i- CM CO
CO CO W

boo

CL CL CL

BUS

m

Acceptable to 1 1 Species Inedible to Inedible

palladia ► penelope Generalists to All

252

J. Res. Lepid.

Results

Data on host specificity, butterfly censuses and grass species richness
are summarized in Figure 1. We identified eleven generalist butterflies
(at left of figure) and three specialists. Thirteen grass and sedge species
were edible to the generalists, and one of these, Ichnanthus pallens, (No.
13) was also edible to a specialist. Two grasses (nos. 14, 15) were classified
as edible only to specialists.

Field searches of the three grass species (16,17,18) at the right of the
figure revealed no eggs or larvae, and these plants were inedible to larvae
of arnaea , hermes, junia, myncea and hesione. It was not possible to
obtain sufficient larvae of the rarer insects to test them on all available
plants, since satyrine larvae are cryptic and secretive, some of them
extremely so. In consequence, our classification of these plants as
inedible to all our study insects is tentative, though in accord with all
data we were able to gather. Because we were not able to classify these
plants as hosts of satyrines, we did not include them in any of our
analyses of host diversity and abundance.

Table 1 shows the total number of butterfly and host species recorded
at each site, and an index of butterfly abundance derived by simply
dividing the total number of individuals by the length of time we spent
in our census. Table 2 shows the same data for the community comprising
only the generalist species and plants edible to generalists.

Spearman correlation coefficients and significance levels of correlation
for all the data are shown in Table 3. When specialists and generalists are
lumped there is no significant correlation between any of the insect
parameters and any of the plant parameters. After removal of the
specialists from the data (along with the plants which are edible only to
these specialists) the correlation between plant abundance and insect
abundance rises to a (just) significant level.

Discussion

ACCURACY OF HOST SPECIFICITY DATA

The evidence for classifying as such the generalists and their hosts is
of variable quality. Classification of specificity was made from field
records of oviposition and larval distribution for ten of the fifteen species
we found. For four of the rare species palladia, ocypete, alcinoe and
erichto , we had to rely on data from captive insects. One rare species,
cephus , found at St. Ann’s as a single specimen, gave no diet information,
and may not even be grass-feeding. We accordingly omitted cephus from
our analyses. In general, field data are better than laboratory data as
indicators of specificity, since insects often show greater diet breadth in
captivity than in their natural habitats (Singer 1986).

We found hermes larvae feeding in the field on 8 plant species, renata
on 6, penelope on 5 and hesione on 5. Their larvae will eat all the other
species if offered them though one of the sedges (No. 11) is only edible to
most larvae when young. Some of these plants are sufficiently rare that

30(3-4):248-256, 1991

253

Table 1 . Characteristics of plant and butterfly communities at ten sites.

Butterfly Butterfly Plant Plant

Site Species Abundance Species Abundance

Trace

9

37.6

8

4

Trace Plantation

8

15.3

2

5

Guanapo

9

38.3

11

5

Dump 1

8

13.3

4

4

Dump 2

4

7.33

4

2

Cave

3

4.33

1

2

POS 1

4

8.67

2

3

POS 2

10

39.0

2

1

POS 3

7

7.00

4

2

St. Ann’s

10

15.7

4

4

Table 2. Characteristics of generalist insects and their hosts.

Butterfly

Butterfly

Plant

Plant

Site

Species

Abundance

Species

Abundance

Trace

8

35.9

8

4

Trace Plantation

7

13.6

2

5

Guanapo

8

38.0

11

5

Dump 1

7

11.3

3

4

Dump 2

2

3.00

1

2

Cave

2

3.00

1

2

POS 1

4

8.67

2

3

POS 2

8

27.3

1

1

POS 3

7

7.00

4

2

St Ann’s

8

9.67

3

4

we would be unlikely to have found larvae on them, even if they were high
quality hosts. Among the rare insects, alcinoe and themis oviposited
readily on 5 grass species offered to free-flying insects in an insectary. We
were unable to obtain natural oviposition behavior from captive terrestris,
ocypete, or palladia. Evidence for classifying these three species as
generalists is poor, and consists only of recording which plant species
were readily consumed by captive larvae.

We have more than 30 observations of arnaea feeding on plant species
13, Ichnanthus pollens. Our evidence for the edibility of this plant to
generalists is also good: we have field observations of hermes , renata ,
libye, and hesione feeding on this species, and we have raised larvae of
myncea, alcinoe and ocypete on it. Evidence that junia is host-specific
comes from oviposition tests on a single captive adult, plus the finding of

254

J. Res. Lepid.

11 eggs (eggs are laid singly and independently in this species) on plant
14. Our evidence for categorizin gerichto is less firm, and consists only of
oviposition trials on three captive adults. All were tested on plants 1
through 18 with the exception of Nos. 3 and 4. No acceptances of any
species other than No. 15 were recorded, and we obtained 16 such
acceptances.

For two of the three species classified as specialists, arnaea and junia,
data obtained from the field are concordant with those obtained from
captive or semi-captive insects. This concordance between field and lab
data reassures us that our laboratory data do seem to have meaning. The
third specialist, erichto , was classified from behavior of captive adults
and larvae. Three adult erichto that were preference-tested would accept
only an epiphytic sedge that hung from trees (16 acceptances observed),
and the larvae likewise refused to feed on plants other than this sedge.

Among those insects classified as generalists, we have adequate field
data for myncea, hesione, libye, renata, hermes, themis and penelope
(although themis was rare in our study sites, we obtained data from
nearby populations of this insect). For alcinoe we were able to perform
preference trials on both captive adults and larvae, but for palladia and
ocypete our only data are from larval feeding preferences. For these two
species, we regard our classification of diet breadth as extremely tenta-
tive.

RELATIONSHIPS BETWEEN INSECT AND PLANT TRAITS

The lack of significant correlation between rank orders of insect species
richness and host species richness indicates the relatively minor role
played by specialists and their hosts in our data set. The only significant
trend is a positive correlation between insect abundance and host
abundance (Table 3), and this appears only after “specialist removal”.
Even this is not a conspicuous trend. There are clear exceptions to it, as
the example of site POS 2 shows. At this site, species 13 served as the sole
host of nine butterfly species, even though it was rare (Table 1). Our
results contrast with those obtained from heliconiine butterflies and
Passiflora , which have shown correlations between species richness of
the insects and that of their host plants (Gilbert & Smiley 1978).

EFFECTS OF HABITAT FRAGMENTATION

The concordance between the distribution of specialists and that of the
habitats available to them indicates that, at least for this set of species,
habitat fragmentation had not resulted in significant local extinction of
these species at the time of our study (1970-74). The presence of any of
the three monophagous species in a habitat could be confidently pre-
dicted from the presence of its host species. The data of Figure 1 show
erichto and junia distributed among sites in precisely the same manner
as their respective hosts, while arnaea was recorded from six of the seven
sites where its host grew. This phenomenon was especially striking since

30(3-4):248-256, 1991

255

Table 3. Spearman correlation coefficients.

Correlation

coefficient

Significance
of correlation

Correlation between:

All Species Generalists

All Species

Generalists

Butterfly/ plant species

0.426

0.514

0.110

0.064

Butterfly species/
plant abundance

0.290

0.356

0.208

0.157

Plant/ butterfly abundance

0.349

0.562

0.162

0.046

Plant species/

Butterfly abundance

0.393

0.474

0.131

0.084

junia was absent from POS 1 and POS 3, which were both only about 50
m from POS 2 where junia and its host were common.

There was no trend in our data for the more isolated habitat patches
(POS, Dump and Guanapo) to contain fewer species. However, much
larger sample sizes would be needed for such a test to have sufficient
power.

On a larger scale, we found no evidence for loss of species from the
island of Trinidad. We recorded fifteen species, while Barcant ( 1970) lists
seventeen satyrines recorded from Trinidad since lepidopterists first
collected there in the 1890’s. One of these seventeen (calpurnia) was
recorded from a single site only, in 1932, while a second (brixiola) had
been recorded only twice. We did not find either of these very rare species.
Although Barcant recorded palladia as present in Trinidad, he was
describing misidentified myncea (Singer et al 1983). Palladia does,
indeed, occur there, but was apparently not found by Barcant (1970) or
by other butterfly -collectors.

Although human activities must have removed many formerly suitable
habitats for these insects, at the time of this study they had not yet
resulted either in extinction of species from Trinidad, or in significant
reduction of the ability of host specialists to maintain populations in
patches of habitat where their hosts still occurred.

Summary

Eleven species of Trinidadian satyrine butterflies were found to be
generalist feeders as larvae, while three were monophagous. The pres-
ence of a monophagous satyrine at a site was accurately predicted from
the presence of its host. This relationship, coupled with a comparison of
our data with historical records, shows that, at the time of our study
(1970-74), habitat fragmentation had not resulted in significant loss of

256

J. Res . Lepid.

these insects either from local habitats suitable for them or from the
Island of Trinidad as a whole.

Comparison between sites showed no significant correlation between
insect species richness or abundance and host plant species richness or
abundance. If generalist feeders and their hosts were considered sepa-
rately, they showed a significant relationship between insect abundance
and plant abundance. This trend, however, was not a very clear one, and
there were striking exceptions to it. There was considerable overlap in
resource utilization, with nine insect species using the same host species
at one of our study sites.

Acknowledgements: We thank P. J. DeVries for pestering us over the years to
finish this paper, for providing critiques, and for retyping the final draft.

Literature Cited

Barcant, M (1970). Butterflies of Trinidad and Tobago. Collins, London
DeVries, P.J. (1985). Hostplant records and natural history notes on Costa
Rican butterflies (Papilionidae, Pieridae & Nymphalidae. J. Res. Lep. 24:
290-333.

. (1987). Butterflies of Costa Rica and their Natural History. Princeton

University Press.

Ehrlich, P. R. & D.D. Murphy. (1987). Monitoring populations on remnants of
native vegetation, in: Nature conservation: the role of remnants of native
vegetation, D.A. Saunders, G. W. Arnold, A. A. Burbidge and A.J.M. Hopkins
(eds.). pp 201-210. Surrey Beatty & sons, Canberra.

Gilbert, L.E. & J. T. Smiley. 1978. Determinants of local diversity in phytophagous
insects: host specialists in tropical environments. In: Diversity of Insect
Faunas. Symp. R. Ent. Soc. Lond. No. 9, pp 89-104.

Kremen, C. 1992. Assessing the indicator properties of species assemblages for
natural areas monitoring. Ecological applications 2: 203-217.

Pollard, E. ( 1977). A method for assessing change in the abundance of butterflies.
Biol. Cons. 12: 115-132

Singer, M.C. (1986) The definition and measurement of oviposition preference
in plant-feeding insects. Ch 3 in: Insect-plant interactions, edited by J.
Miller and T.A. Miller. Springer- V erlag, pp 65-94.

Singer, M.C., P. J. DeVries & P.R. Ehrlich. (1983). The Cissa confusa species
group in Costa Rica & Trinidad. Zool. J. Linn. Soc. 79: 101-119.

Singer, M.C., P.R. Ehrlich & L.E. Gilbert. (1971). Butterfly feeding on lycopsid.
Science 172: 1341-1342

Singer, M.C. & J.L.B. Mallet. (1986). Moss-feeding by a satyrine butterfly. J.
Res. Lep. 24: 392.

Journal of Research on the Lepidoptera

30(3-4):257-260, 1991

Celastrina nigra and its synonym C. ebenina
(Lepidoptera: Lycaenidae)

James A. Scott

60 Estes St., Lakewood, Colo. 80226

and

David M. Wright

124 Heartwood Drive, Lansdale, Penn. 19446

Abstract. Celastrina nigra Forbes 1960 is the valid name, not C.
ebenina Clench 1972, because Forbes raised the infrasubspecific nigra
Edwards 1884a to subspecies rank. C. nigra is not a homonym of
Scolitantides orion nigra Gerhard 1882.

Both nigra and ebenina have been used as the name of the same
Celastrina species, nigra by Scott (1984, 1986), ebenina by Wagner and
Mellichamp (1978), Miller and Brown (1981), and Ferris (1989). This
paper shows that C. nigra Forbes 1960 is the correct name and C. ebenina
Clench 1972 is a synonym; it summarizes a lengthy paper on the case,
which is available upon request.

The oldest names for C. nigra are by Strecker ( 1878): uLycaena lucia ab.
a female nig .” (actually a male), and “ab. b female intermedia”. Both were
named as aberrations and are unavailable infrasubspecific names (ICZN
1985 art. l[b][5]); Strecker used nig. to abbreviate nigra, niger, and
nigrum, but evidently intended this nig. to abbreviate the feminine nigra
because he erroneously thought the insect was female. Edwards (1884a)
described Lycaena pseudargiolus infrasubspecific form nigra as a “mela-
nic dimorphic male form” — a melanic form limited to males — of C.
argiolus form uiolacea (Edwards); thus it is an infrasubspecific name, a
minority element within a population, certainly not a subspecies. Miller
and Brown (1981) erred in stating that nigra was named as an aberra-
tion, a word not mentioned by Edwards. The name nigra Edwards was
actually described in W. H. Edwards’ (1884a, June) Butterflies of North
America 11:315-319 (p. “1-5”), which appeared before Edwards’ (1884b, p.
306, Dec.) “Revised Catalogue”, mistakenly cited by Miller and Brown
(1981) as the original description.

Forbes (1960) described Plebeius argiolus nigra: “Race or rather local
variety nigra Edwards (H 31:4 — female intermedia Strecker) is solid
brown above in male, and dominantly blackish in female. It is limited to
a small area in West Virginia and western Pennsylvania so far as I
know.” Forbes’ word “race” clearly describes a subspecies; his words
“rather local variety” are less clear but signify a geographically restricted
type of animal of at least a population (“local variety” in sloppy American
English generally means a subspecies with very restricted range); his
description of the limited range in W.Va.-Pa. is a clear indication of a
subspecies (ICZN 1985 article 45[f][ii], 1961 art. 45[d][ii]).

258

J. Res. Lepid.

Rules. The 1985 ICZN rules apply precisely to the case. Article 10(c)
states that an infrasubspecific name that satisfies the other criteria of
availability becomes available when the name is used for a species or
subspecies. Edwards’ nigra satisfies all criteria of availability except
that it is infrasubspecific (art. l[b][5], 45[e]); Forbes’ (1960) nigra satis-
fies all criteria including treating it as a subspecies and giving a
diagnosis. The 1985 rules regarding “forms” and “varieties” also indicate
that Plebeius argiolus nigra Forbes 1960 is an available name, while
Lycaena nigra Edwards 1884 is unavailable. ICZN 1985 article 45(g)
states that: before 1961, the use of either of the terms “variety” or “form”
is to be interpreted as subspecific rank, but if the intention of the work
reveals that infrasubspecific status is meant (as in nigra Edwards), the
name is infrasubspecific, UNLESS before 1985 it has been treated as an
available name and either adopted as the name of a species or subspecies
(which Forbes [1960] and Scott [1984] did), or treated as a senior
homonym in which case the name is deemed to be subspecific from the
date of its establishment (“establishment” — according to the glossary
definition — was in 1960 as Plebeius argiolus nigra Forbes).

The 1948 to 1985 rules mandated more rigorous standards for publish-
ing subspecies after a cutoff date: before 1951 in the 1948 rules in effect
until 1960, before 1961 in the 1961-1985 rules. Before and after this
cutoff, a name is a subspecies if the original description clearly indicated
the taxon is a subspecies, and a name is infrasubspecific if the author
expressly indicated that he regarded the taxon as infrasubspecific. But
if the author did not clearly state in the original publication whether he
regarded it as being a subspecies or an infrasubspecific form, a name
established before the cutoff is a subspecies, but after the cutoff the name
is infrasubspecific. The rollback of the date to the start of 1961 in the 1961
Code which Clench (1972) had to follow, means that pre-1961 names are
(1) subspecies if described as such, (2) subspecies if not clearly described
as subspecies or infrasubspecific (thus even if Clench thought — or any-
one now thinks — that F orbes’ description was ambiguous), (3) subspecies
if merely described as “varieties” or “forms” without accompanying
description clarifying status. C. ebenina would be valid only if Forbes had
unambiguously described nigra as an infrasubspecific name, but Forbes
did the opposite by describing it as a “race” — a subspecies.

Readers are warned that the example at the end of 1985 Article 45(g)
is grossly misleading in implying that infrasubspecific names raised to
subspecies rank take authorship and date from the original publication
of the infrasubspecific name. 1985 article 87(b) states that “Examples do
NOT form part of the legislative text of the Code” so the example must be
ignored because the implication contradicts articles 23(j), 50(c)(i), 10(c),
and 45(g)(ii)(l); and the example is ambiguous because the author of the
name in the example may have published a second paper in the same year
treating the varieties as subspecies, in which case the example could also
be (properly) interpreted as meaning that the name should take the

30(3-4):257-260, 1991

259

authorship and date of the second paper raising the names to subspecies
rank. We have informed the ICZN that this example must be replaced.

Ferris ( 1989) was mistaken when he stated that “ ebenina is valid for the
edition of the code [1961/1964] under which Clench worked”; C. nigra
Forbes (1960) was a valid available subspecies under the 1961/1964 rules
(articles 10[b], 45[d][i, ii]), and ebenina became a synonym the day it was
named in 1972. And nigra Edwards 1884 is unavailable under the 1961/
1964 rules (art. 1, 45[c], 45[d][iii]). The 1961/1964 rules were written
imprecisely because 196 1 article 45(e)(i) states that before 1961 “variety”
or “form” are not to be interpreted as either infrasubspecific or subspe-
cific rank, and article 17(9) states that a name remains available even
though before 1961 it was proposed as a “variety” or “form”; a strict
interpretation of these two rules in isolation from others might suggest
that they override 45(d)(iii) (but not 45[c]) in cases in which the word
“form” or “variety” is used, possibly making nigra an available species
group name even from 1884 to 1960. But logic clearly shows that the only
way all parts of rules 17 and 45 can be applied simultaneously without
contradiction is if the intent of the 1961/1964 rules was that a name
merely described and ranked only as “form” or “variety”, without
infrasubspecific modifiers such as “melanic dimorphic” etc., be treated as
an available subspecific name, whereas a name described as “form” or
“variety” with qualifying modifiers signifying infrasubspecific status
was to be treated as infrasubspecific and unavailable. This is the clearly
stated intention of the 1948 rules which form the basis of the 1961 rules,
and is the clearly written intention of the 1985 rules. Obviously, C. nigra
Edwards 1884 is infrasubspecific thus unavailable.

C. nigra Forbes in fact is valid and nigra Edwards invalid and
unavailable, in all versions of the rules as far back as 1842. The unofficial
Stricklandian Code (Strickland, 1842) did not permit the use of
infrasubspecific names or varieties. The unofficial Dali Code (1878)
permitted the use of “varieties” but not infrasubspecific names such as
“melanic dimorphic male forms”. The Regies from 1905 onward did not
cover infrasubspecific names until 1948 when the current rules regard-
ing infrasubspecific names were devised.

Homonymy. Lycaena pseudargiolus infrasubspecific form nigra
Edwards 1884 was named two years after Lycaena orion variety nigra
(Gerhard 1882, p. 126). The latter is now a ssp. of Scolitantides orion
(Pallas) 1771 (1985 article 45[g] etc.), and is evidently a senior synonym
of all three other ssp. now used in the species. But neither nigra Edwards
nor nigra F orbes is a homonym, because nigra Edwards is infrasubspecific
so is not covered by the 1961-1985 Code, and nigra Forbes was named in
genus Pleheius.

Earlier rules did not make nigra Edwards a homonym either. The
unofficial Dali Code (1878) did provide for homonymy of infrasubspecific
names, but only if both names were within a single species (rules LXVI
and LVII), and nigra Gerhard and nigra Edwards were always in
separate species.

260

J. Res. Lepid.

The 1905-1947 Regies mentioned only subspecies and species and did
not cover infrasubspecific names, so nigra Edwards was not subject to
homonymy.

From 1948 until the 1961 Code, the Regies allowed homonymy within
infrasubspecific names (ICZN, 1950, p. 93, rule [9][a]), but homonymy
applied separately to (A) subspecies/species names than to (B)
infrasubspecific names, such that (A) and (B) were self-contained and
mutually independent sectors of nomenclature. Homonymy was not
possible since nigra Edwards is infrasubspecific and nigra Gerhard is a
subspecies ( nigra Gerhard was named as a “variety” but is a subspecies
by 1948 rule [3] [ICZN, 1950] and all later Regies and Codes).

Acknowledgements. We are grateful to John N. Eliot and an anonymous referee
for thorough review and very helpful comments.

Literature Cited

Clench, H. K. 1972. Celastrina ebenina, a new species of Lycaenidae (Lepidoptera)
from the eastern United States. Annals Carnegie Mus. 44:33-44.

Dall, W. H. 1878. Report of the committee on zoological nomenclature. Proc.

Amer. Assoc. Advancement Sci. 26:7-56.

Edwards, W. H. 1884a. The butterflies of North America. Vol. II (1874-1884).
Houghton, Mifflin & Co., Boston, Mass.

1884b. Revised catalogue of the diurnal Lepidoptera of America north of

Mexico. Trans. Amer. Entom. Soc. 11:245-337.

Ferris, C. E., ed. 1989. Supplement to: A catalogue/checklist of the butterflies
of America north of Mexico. Lepid. Soc. Mem. #3.

Forbes, W. T. M., 1960. Lepidoptera of New York and neighboring states. Part
IV, Agaristidae through Nymphalidae including butterflies. Memoir 371
Cornell Univ. Agric. Exp. Sta. N. Y. State College of Agric., Ithaca, New
York. 188 p. (see p. 127).

Gerhard, B. 1882. Lepidopterologisches. Berliner Entomologische Zeitschrift
26:125-128.

International Commission on Zoological Nomenclature. 1950. Ed. Francis
Hemming. Bull. Zool Nomenclature 4:i-lii, 1-760.

Miller, L. D., & F. M. Brown. 1981. A catalogue/checklist of the butterflies of
America north of Mexico. Lepid. Soc. Mem. #2, 280 p.

Scott, J. A. 1984. News of Lepid. Soc. 1984 No. 1 p. 6.

1986. The butterflies of North America, a natural history and field guide.

Stanford Univ. Press., Stanford, Calif. 583 p.

Strecker, H. H. 1878. Butterflies and moths of North America.. .a complete
synonymical Catalogue. ..Diurnes. Press of B. F. Owen, Reading, Penn.
Strickland, H. E.,etal. 1842. Rules for Zoological Nomenclature. Report of 12th
meeting of British Association held at Manchester in 1842. British Assoc.
Advancement Sci. Report 1842, 11:105-121.

Wagner, W. , & T. Mellichamp. 1978. Foodplant, habitat, and range of Celastrina
ebenina (Lycaenidae). J. Lepid. Soc. 32:20-36.

Journal of Research on the Lepidoptera

30(3-4):261-266, 1991

Foam barriers, a new defense against ants for
milkweed butterfly caterpillars (Nymphalidae:
Danainae)

P.J. DeVries

Museum of Comparative Zoology, Harvard University, Cambridge, MA 02138

Abstract. All instar caterpillars of Idea leuconoe and Euploea crameri
from Brunei, Borneo trench leaves and/or cut veins of their hostplant
Parsonsia spiralis, presumably to circumvent plant defensive chemis-
try. First through third instar caterpillars also line the outside perim-
eter of circular trenches with a regurgitated foam that repels ants.
These observations are discussed with respect to chemistry of danaine
hostplants, and future questions are raised regarding this system.

Introduction

Cutting the leaves or stems of plants in the families Caricaceae,
Moraceae, Apocynaceae, or Asclepidaceae typically causes copious milky
sap to emanate from the wounded tissues. To an insect herbivore this is
a graphic example of rapidly mobilized plant defenses produced in
response to tissue damage; the rapidly oozing sap may hinder its ability
to feed. In addition to a physical defense, plant sap commonly contains
secondary chemicals that can either deter insect herbivores, or be
outright toxic to them (Dussourd & Denno, 1991). To overcome this type
of plant defense some insects have evolved the habit of severing major
leaf veins or cutting a circular trench into leaf tissues prior to feeding —
behaviors termed vein cutting, or trenching. From the insect’s point of
view, trenching and vein cutting behavior may impede or stop mobiliza-
tion of plant defenses, and make leaf tissues edible (Carroll & Hoffman,
1980; Dussourd & Denno, 1991).

Milkweed butterfly caterpillars (Danainae: Nymphalidae) typically
feed on plants containing abundant laticifers (Ackery, 1988), and some
species are known to exhibit trenching and/or leaf cutting behavior
(Ackery & Vane-Wright, 1984; Compton, 1987; DeVries, 1987; Dussourd
& Eisner, 1987). The purpose of this paper is three-fold. First, is to
describe trenching behavior in the early instar caterpillars of the danaines
Idea leuconoe and Euploea crameri from Borneo. Second, is to describe
a new defense where caterpillars produce a foam barrier that repels ants.
Finally, the observations are discussed in light of chemical defenses
found in danaines and other insect herbivores, and several questions are
raised regarding this system.

Observations

During June 1983 I made observations on the life cycles of Idea
leuconoe nigriana Grose-Smith, 1895, and Euploea crameri Lucas, 1853
in the coastal mangrove forests near Bandar Seri Begawan, Brunei on
the island of Borneo. Both of these butterflies are part of the monophyl-

262

J. Res. Lepid.

Figure 1: Second instar caterpillar of Idea leuconoe nigriana (Danainae) beginning
to create a circular trench in a Parsonsia spiralis leaf. The caterpillar has
excavated two areas in the leaf epidermis (curved arrows). Note the six
areas of foam that have been deposited on the outside perimeter of the
circular trench (straight arrows). Eventually this caterpillar completely
surrounded the trench with foam.

Figure 2: A trenched area abandoned by a second instar Idea leuconoe nigriana
caterpillar. The caterpillar has eaten away the circular area in the epidermis
of the Parsonsia leaf. Note the broken ring of residue on the outside
perimeter of the trench left by the dried foam barriers.

etic subtribe Euploeina, and range widely throughout the island of
Borneo, and elsewhere in southeast Asia (Ackery & Vane-Wright, 1984).
In the Brunei mangrove habitat, both species fed on Parsonsia spiralis
(Apocynaceae) as caterpillars, a common, woody vine in the coastal
mangrove forests, and in second growth mangrove habitats (see Ackery
and Vane-Wright, 1984).

In the field and in an ambient temperature laboratory, first through
early third instar caterpillars of both/, leuconoe , and#. crameri typically
cut a circular trench in the leaf, and fed on the tissues within this circular
trench (Figs. 1 & 2). Fourth and fifth instar caterpillars typically cut
large leaf veins or the petioles, and then fed on tissues distal to these cuts.
The sap of P. spiralis is slightly sticky when exposed to air, but unlike
many other members of the Apocynaceae the sap is clear, not milky.
Although trenching and vein cutting occurs in other danaine species
(DeVries, 1987; Dussourd and Eisner, 1987), this is apparently the first
report specifically for I. leuconoe and E. crameri.

In addition to trenching, the first through third instar caterpillars of
both species also lined the outside perimeter of their circular trench with
a yellowish foam barrier. Depending on the individual caterpillar, the
foam barriers ranged from a closed circle to a broken circle of foam. The
foam barriers were erected as follows. After hatching and eating the
chorion of the egg, a caterpillar would chew into the leaf, extend the body

30(3-4):261-266, 1991

263

away from the chewed area, regurgitate a small amount yellowish foam,
and deposit it on the leaf surface distal to the chewed area. This process
was repeated until the caterpillar remained inside a ring-trench ringed
on the outside perimeter with yellowish foam (Fig. 1). Although forth and
fifth instar caterpillars cut the veins of leaves prior to feeding, they were
never observed to produce foam or erect barriers.

One of the most abundant ant species in Brunei mangrove habitats was
the weaver ant, Oecophylla smarigdina (Formicinae). As elsewhere in its
range, a single colony of this ant may form numerous, interconnected,
arboreal nests that include large areas within its foraging territory.
Several features of weaver ant biology are pertinent here: they have
acute vision, are extremely sensitive to movement, aggressively territo-
rial, they recruit quickly to any disturbance or resource within their
territory, and they possess well-developed mandibles and chemical spray
defenses used in prey capture and defending the colony. Although they
harvest secretions produced by Homoptera and lycaenid butterfly cater-
pillars, the diet of weaver ants also includes substantial proportions of
arthropods. Thus, where weaver ants are ubiquitous they can exert
considerable predator pressure on the local arthropod community (re-
viewed in Holldobler & Wilson, 1990).

A simple experiment suggested that the foam placed by Idea and
Euploea caterpillars on the outside of circular trenches functioned to
repel ants. The experiment consisted of bridging a captive portion of a
weaver ant colony to a cut portion of Parsonsia vine (placed in a bottle of
water) that had either a first or second instar 7. leuconoe or E. crameri
caterpillar on it complete with circular trench and foam barrier. Once set
up, the interactions between caterpillars and ants were noted.

The observations are summarized as follows. As individual ants moved
onto the Parsonsia cutting they investigated the stem and all leaves.
When an ant encountered the foam barrier with the caterpillar inside it,
there were typically two reactions. First, if the caterpillar was resting
inside the ring, upon contacting the foam the ant would immediately
back away, groom its antennae, and then move to another area of the
plant. Second, when an ant found a foam ring with a moving caterpillar
within, it attempted to attack the caterpillar (presumably because the
ant could see it), but was repelled by contact with the foam. In this case
an ant would make 2-5 attempts at attacking the caterpillar, then back
away, groom its antennae and legs, then move to another area of the
plant. This experiment was repeated with six individual caterpillars and
plants, and in all instances I found that foam barriers that were between
0.25 and 2 hours old repelled the attempts of 10-17 individual ants to get
at the caterpillars within foam barriers.

Approximately 6 hours after the foam barriers were deposited they
degraded, and eventually dried to a noticeable yellowish scum (Fig 2).
Repeating the experiment three times with degraded barriers, I found
that the dried foam had no observable deterrent effect on the ants, which
attacked and killed caterpillars.

264

J. Res. Lepid.

Discussion

The caterpillars of I. leuconoe and E. crameri excavate trenches or cut
veins in their hostplant tissue prior to feeding (Fig. 1 & 2). As in other
insects (including some species of Danainae), this behavior may circum-
vent plant chemical defenses, thereby making the leaf tissues more
palatable to the caterpillars (e.g., Dussourd & Denno, 1991).

Sawfly larvae in the genus Stauronema (Tenthredinidae) secrete anti-
predator chemical defenses through specialized epidermal glands, and
also surround themselves and the leaf area they feed on with a foam
regurgitation. The foam regurgitated by Stauronema larvae has a repel-
lent effect upon ants (Boeve & Pasteels, 1985). The production of
defensive foam by sawfly larvae and danaine caterpillars indicates this
trait has evolved independently at least twice among insects, and points
to the possibility that the use of foam to repel ants may occur in other
groups as well.

Some butterflies and Homoptera incur strong survival benefits by
forming symbioses with ants (e.g., DeVries, 1991, DeVries, 1992; Pierce
et al., 1987; Way, 1963). However, predation by ants can strongly affect
the distribution and abundance of those insects that cannot form such
symbiotic associations (reviewed by Whittaker, 1991). The abundance
and predatory nature of O. smarigdina suggests that foam barriers
probably increase the chance of survival for early instar Idea and
Euploea caterpillars by acting as a deterrent to foraging weaver ants.

Nothing is known about the chemical composition of the foam produced
by Idea and Euploea caterpillars (Fig. 1). We also do not know whether
it is derived from plant tissues, synthesized by the caterpillars directly,
or results from an interaction of both. The genus Parsonsia contains a
suite of secondary chemicals, including large fractions of pyrrolizidine
alkaloids (Edgar, 1984; Edgar &Culvenor, 1975). Moreover, pyrrolizidine
alkaloids are integral to all life stages of many danaine species (Ackery
& Vane-Wright, 1984), and act as a repellent to a variety of predators
(Brown, 1984; Dussourd et al., 1988; Boppre, 1990). Consequently, the
foam produced by Idea and Euploea caterpillars may contain defensive
chemicals derived directly from Parsonsia leaf tissue, including
pyrrolizidine alkaloids.

A long history of systematic and ecological interest in the Danainae
makes them one of the best understood groups of all butterflies (see
Ackery & Vane-Wright, 1984). The observations here raise four ques-
tions pertinent to danaine evolution and biology. First, are foam barriers
confined to the subtribe Euploeina, or does this trait occur elsewhere
within the Danainae ? Second, are foam barriers built only by/, leuconoe
caterpillars that feed on Parsonsia , or does this trait also occur in those
that feed on members of the Asclepidaceae? Third, do the traits of
trenching and erecting foam barriers occur independently within the
danaines? Finally, are foam barriers specific to ants, or do they repel a
suite of arthropods? Detailed experiments and analyses will be required

30(3-4):261-266, 1991

265

to identify the exact function of foam barriers and their chemical
composition, but simple field observations and manipulations can quickly
provide more information relevant to danaine phylogeny and behavioral
ecology.

Acknowledgments: The caterpillar and ant behaviors described here were filmed by R.
Foster, and the 16mm footage is in the film vaults of Partridge Films LTD, London.
Professors P. R. Ackery, B. Bolton, and C. J. Humphries, and R. I. Vane-Wright who
were at the BMNH at the time of this study (now abbreviated the NHM) kindly
identified the butterflies, ants, and plants. Insect voucher specimens are in the author’s
collection, and a specimen of hostplant has been deposited in the collection of the NHM,
London. Thanks to R. Dudley and N. Greig for commenting on a draft of this paper, M.
Rosenberg for the travel grant, and the entire staff of the then BMNH for their help.
This work was supported by a Fullbright Fellowship, Partridge Films Ltd., the
MacArthur Foundation, and is dedicated to Sonny Stitt, Sonny Criss, and Joe Farrell.

Literature Cited

Ackery, P. R. 1988. Hostplants and classification: a review of nymphalid
butterflies. Biol. J. Linn. Soc. 33: 95-203.

Ackery, P. R. & R. I. Vane-Wright 1984. Milkweed butterflies. London, British
Museum (Natural History).

Boeve, J. & J. M. Pasteels. 1985. Modes of defense in nematine sawfly larvae,
efficiency against ants and birds. J. Chem. Ecol. 11: 1019-1036.

Boppre, M. 1990. Lepidoptera and pyrrolizidine alkaloids: exemplification of
complexity in chemical ecology. J. Chem. Ecol. 16: 165-185.

Brown, K. S. 1984. Adult-obtained pyrrolizidine alkaloids defend ithomiine
butterflies against a spider predator. Nature. 309: 707-709.

Carroll, C. R. & C. A. Hoffman. 1980. Chemical feeding deterrent mobilized in
response to insect herbivory and counteradaptation by Epilachna
tredecimnotata. Science. 209: 414-416.

Compton, S.G. 1987. Aganais speciosa and Danaus chrysippus (Lepidoptera)
sabotage the latex defences of their host plants. Ecol. Ent. 12: 115-118.
DeVries, P. J. 1987. The Butterflies of Costa Rica. Princeton, Princeton University
Press.

. 1991. The mutualism between Thisbe irenea and ants, and the role of ant

ecology in the evolution of larval-ant associations. Biol. J. Linn. Soc. 43: 179-
195.

. 1992. Singing caterpillars, ants and symbiosis. Scientific American. 267:

76-82.

Dussourd, D. E. & T. Eisner. 1987. Vein-cutting behavior: insect counter ploy to
the latex defense of plants. Science. 237: 898-901.

Dussourd, D. E., K. Ubik, K. Harvis, J. F. Resch, J. Meinwald, & T. Eisner. 1988.
Biparental defense endowment of eggs with acquired plant alkaloid in the
moth Uthesia ornatrix. Proc. Natl. Acad. Sci. U.S.A. 85: 5992-5996.
Dussourd, D. E. &. R. F. Denno. 1991. Deactivation of plant defense:
correspondence between insect behavior and secretory canal architecture.
Ecology. 72: 1383-1396.

Edgar, J. A. 1984. Parsonsieae: ancestral larval foodplants of the Danainae and
Ithomiinae. Symp. Roy. Ent. Soc. 11: 91-93.

Edgar, J. A. & C. C. Culvenor. 1975. Pyrrolizidine alkaloids in Parsonsia
(Apocynaceae) which attract danaid butterflies. Experientia. 31: 393-394.

266

J. Res. Lepid.

Holldobler, B. & E. O. Wilson. 1990. The Ants. Cambridge., Harvard University
Press.

Pierce, N. E, R. L. Hitching, R.C Buckley, M.F.J. Taylor, & K.F. Benbow. 1987.
The costs and benefits of cooperation between the Australian lycaenid
butterfly, J almenus evagoras, and its attendant ants. Behav. Ecol. &
Sociobiol. 21: 237-248.

Way, M. J. 1963. Mutualism between ants and honeydew producing Homoptera.
Ann. Rev. Entom. 8: 307-344.

Whittaker, J. B. 1991. Effects of ants on temperate woodland trees. IN:
Interaction Between Ants and Plants. D. Cutler & C. Huxley (eds.). 67-79.
Oxford, Oxford Univ. Press.

Journal of Research on the Lepidoptera

30(3-4):267-271, 1991

Temporal and spatial overlap in the mate-locating
behavior of two species of Junonia (Nymphalidae)

Ronald L. Rutowski

Department of Zoology, Arizona State University, Tempe 85287-1501, USA

Abstract. In northeastern Australia, males of two species of Junonia ,

J. villida and J. orithya, use a sit-and-wait tactic to detect females at
encounter sites. In a brief study of the behavior of both species, I
documented considerable overlap in the locations of the encounter sites
used by males and in the time of day at which males were at encounter
sites. Because of this overlap interspecific interactions are frequent,
but males of J. villida at least respond much more strongly to conspe-
cific males than to males of J. orithya. The discussion focuses on the
lack of divergence in the mating behavior of these sympatric species.

Introduction

Males of many insects use landmarks such as hilltops or open areas
along paths as places to encounter receptive females (Thornhill and
Alcock, 1983). These landmarks typically do not contain any resources of
material benefit to females but are instead places where receptive
females appear primarily to mate. Some landmarks are used at the same
time by a variety of species many of which are highly territorial. For
example, Alcock (1987) found convergence in the spatial and temporal
structure of male mate-locating tactics of a diverse set of hilltopping
insects including a pompilid wasp, a bot fly, and representative of all
families of butterflies (Alcock 1984, 1987; Alcock and O’Neill, 1987).
However, there have been few studies that deal with the issue of what
happens when closely related species, similar in overall morphology and
behavior, share encounter site preferences. The general expectation is
that selection will favor male behavior patterns that reduce the costs in
terms of energy and distraction arising from interactions with other
species. Brown and Alcock (1990), Hafernik (1982), Turner (1990), and
Pinheiro (1990) all reported spatial segregation of butterfly species that
concurrently utilize similar encounter sites. Moreover, Callaghan ( 1982)
indicates that hilltopping Neotropical butterflies in 10 riodinid genera
are temporally as well as spatially segregated.

In the Indo-Australian region, two species of Junonia Hubner, J.
villida Godart and J. orithya Butler, have similar distributions and
habitat preferences (Common and Waterhouse, 1982). Moreover, pub-
lished reports (Heinrich 1972) and my preliminary observations indi-
cated that males of these species utilize mate-locating tactics similar to
those seen in J. coenia Hubner in North America (Hafernick 1982, Scott
1975b). Males perch in open areas of bare ground and fly up at passing
conspecifics and heterospecifics. This brief study was designed to deter-
mine if the males of these two sympatric species are spatially and

268

J. Res. Lepid.

temporally separated and if males respond more vigorously to conspecif-
ics than they do to heterospecifics. A stronger response to conspecifics is
expected for two potential reasons. First, males might be expected to
make an effort to remove potential competitors from their perching
locations. Heterospecifics are not mating competitors, so a stronger
response should be directed at conspecifics (Brown and Alcock, 1990).
Second, conspecific males may be more likely than heterospecific males
to be misidentified as females when approached (Scott, 1975a) and then
be chased and courted in error, especially given the low level of sexual
dimorphism in these species (Common and Waterhouse, 1982).

Methods

The data were collected between 14 June and 6 July 1989 in an open eucalypt
woodland on the campus of James Cook University of North Queensland,
Townsville, Queensland, Australia. A census route, that was 600 m in length and
followed a dirt road through the bush, was walked hourly to establish the
distribution of perched males along the road and the daily timing of male activity.

In addition, observations were made on interactions between perched J. villida
males and butterflies flying by their perch. The durations of these interactions
were measured with a handheld stopwatch from when the males came within 2-
3cm of one another until they separated. The form and outcome of the interac-
tions was recorded immediately following the observation.

The census route sloped very gently uphill to the north except for a dip near the
middle of the route where a small stream crossed the road. The only other major
features were a small road that branched off to the west from the census route at
about 200 m from the north end and a few large trees overhanging trees at 60, 240,
and 330 m from the north end. The overhanging vegetation cast partial shade
and was at least 15 m above the surface of the road.

All parametric summary statistics are given as the mean ± one standard error
of the mean (SEM). Statistical comparisons were evaluated at the 0.05 level of
significance.

Results

Males of the two species overlapped both in terms of the timing of their
perching activity and the spatial distribution of preferred perch sites. In
both species males were found perched on the road between 10:00 AM
and 15:00 PM and the largest number of males of both species was found
on the road between noon and 15:00 PM (Fig. 1).

Neither species was uniformly distributed along the census route (Fig.
2). Males of J. villida were seen perched in only 25% of the 30m segments
along the census route. Similarly, J. orithya males were seen in only 40%
of the 20 segments. Also, males were not uniformly distributed among
those segments in which they were seen perched (J. villida : x2 = 13.6, 4
df, p = 0.009; J. orithya : %2 = 21, 7 df, p - 0.004). Males of the two species
were distributed differently among those segments where at least one
species was observed (%2 = 26, 9 df, p = 0.002). Only one of the species was
seen in those segments where less than 2 males were seen during the 19
censuses. Nonetheless, there was a great deal of overlap in their distri-

30(3-4):267-271, 1991

269

TIME OF DAY

Fig. 1 . The relationship between the mean number of males seen perched along the
census route and time of day (J. villida, squares and positive error bars (1
SEM); J. orithya, circles and negative error bars (1 SEM)). The number of
censuses from which the means and SEMs are calculated is shown above
each time period.

CO

LU

O

O

15

DISTANCE ALONG ROUTE (M)

Fig. 2. The relationship between the total number of males seen perched in a 30 m
stretch of the census route in 19 censuses and the distance along the
census route (open bars, J. orithya, filled bars, J. villida). The bars for each
30 m segment are above the distance at the end of the segment away from
the northern end of the route.

butions; for example, that part of the route at 150-210 m was popular
with both species and in this and other overlap areas there was no
evidence of spatial segregation with 30m segments. Males ofboth species
were often seen perched within a meter of one another.

Males ofboth species flew up from their perches and approached both
conspecifics and heterospecifics. Detailed observations were made on the
form and duration of interactions between perched J. villida males and

270

J. Res. Lepid.

congeners. An interaction with a conspecific typically began when a
perched male flew up and chased a conspecific flying overhead. The two
males then began flying rapidly around one another while gaining
altitude. At a height of 3 m or so the two males quickly separated and one
flew off while the other returned to the area and perched. Ten interac-
tions of this sort that were timed lasted 3.99 + 0.398 sec (range = 2.1- 6.4
sec).

Interactions between perched J. villida and J. orithya males flying
overhead always ended very soon (1.55 + 0.184 sec, range = 1.2 - 2.6 sec,
n = 8) after the two came into close contact. These interactions were
significantly briefer than those between J. villida males (Student’s t =
5.11, 16 df, p = 0.0001) and never led to an ascending flight. If both males
had perches nearby, both usually returned to these sites after the
interaction.

Discussion

The data suggest substantial overlap in when and where males of these
two species sit-and-wait in their efforts to locate females, at least during
the late autumn and early winter. This is in contrast to studies which
have found clear divergence in the mate-locating tactics of males in a
variety of sympatric butterfly species (Brown and Alcock, 1990; Callaghan,
1982; Pinheiro, 1990; Turner, 1990), including two New World species of
Junonia (Hafernik 1982). This lack of divergence in the behavior of the
closely related sympatric species suggests that (1) spatial and temporal
distributions of receptive females in the two species are similar and (2)
the potential costs of interspecific interference are relatively low (Rutowski
(1991)).

At this point little is known about the behavior and ecology of receptive
females in these species and the factors that might affect their distribu-
tion. In both species the larvae are polyphagus and feed on a wide range
of similar foodplants (Common and Waterhouse, 1982) which might
produce similar spatial patterns of emergence in these species. Also,
females may mate more than once in J. villida and other members of the
genus (Ehrlich and Ehrlich, 1978; Scott, 1975a). Similarities in patterns
of female receptivity may also contribute to interspecific similarities in
the distribution and abundance of receptive females.

Males may gain little by adopting distinct perch sites because the costs
of interspecific interactions appear to be small, at least under the
conditions that applied during this study. These costs would be deter-
mined by the frequency of interactions with heterospecifics as well as the
form that such interactions take. Although no measures were made of
interaction rates in this study; interspecific interactions were frequently
observed. Indeed, males indiscriminately approach just about anything
passing nearby, including birds, other insects, and thrown rocks. How-
ever, after the initial approach males of at least J. villida males respond
to males of J. orithya less intensely that they do to conspeciflcs. In this

30(3-4):267-271, 1991

271

way, males apparently minimize the time spent interacting with
heterospecific males and there is then no strong selection that favors the
differential utilization of encounter sites by these species. It would be of
interest to see if male behavior diverges when conditions such as higher
population densities of both species produce conditions that increase the
costs of interspecific interactions.

Acknowledgements. This work was done while I was on sabbatical leave in the
Department of Zoology at James Cook University of North Queensland. I thank
Professor Rhondda Jones and the Department of Zoology for the space and other
resources they provided.

Literature Cited

Alcock, J. 1984. Convergent evolution in perching and patrolling site preferences
of some hilltopping insects of the Sonoran Desert. Southwestern Nat. 29:
475-480.

. 1987. Leks and hilltopping in insects. J. Nat. Hist. 21: 319-328.

Alcock, J. & K. M. O’Neill. 1987. Territory preferences and intensity of
competition in the grey hairstreak Strymon melinus (Lepidoptera,
Lycaenidae) and the tarantula hawk wasp Hemipepsis ustulata
(Hymenoptera, Pompilidae). Amer. Midi. Nat. 118: 128-138.

Brown, W. D., & J. Alcock. 1990. Hilltopping by the red admiral butterfly: mate
searching alongside congeners. J. Res. Lep. 29: 1-10.

Callaghan, C. J. 1982(83). A study of isolating mechanisms among Neotropical
butterflies of the Subfamily Riodininae. J. Res. Lep. 21: 159-176.

Common, I. F. B., & D. F. Waterhouse. 1982. Butterflies of Australia, Field
Edition. Angus and Robertson, Sydney.

Ehrlich, A. H., & P. R. Ehrlich. 1978. Reproductive strategies in the butterflies:
I. Mating frequency, plugging, and egg number. J. Kansas Ent. Soc. 51: 666-
697.

Hafernik, J. E., Jr. 1982. Phenetics and ecology of hybridization in buckeye
butterflies (Lepidoptera: Nymphalidae). Univ. Calif. Publ. Entomol. 96: 1-
109.

Heinrich, B. 1972. Thoracic temperatures of butterflies in the field near the
equator. Comp. Biochem. Physiol. 43A: 459-467.

Pinheiro, C. E. G. 1990. Territorial hilltopping behavior of three swallowtail
butterflies (Lepidoptera, Papilionidae) in western Brazil. J. Res. Lep. 29:
134-142.

Rutowski, R. L. 1991. The evolution of male mate-locating behavior in butterflies.
Amer. Nat. 138: 1121-1139.

Scott, J. A. 1975a. Movement of Precis coenia, a “pseudoterritorial” submigrant
(Lepidoptera, Nymphalidae). J. Anim. Ecol. 44: 843-850.

. 1975b. Variability of courtship of the buckeye butterfly, Precis coenia

(Nymphalidae). J. Res. Lep. 14:142-147.

Turner, J. D. 1990. Vertical stratification of hilltopping behavior in swallowtail
butterflies (Papilionidae). J. Lep. Soc. 44: 174-179.

Journal of Research on the Lepidoptera

30(3-4):272-278, 1991

Cephalic Sclerites and Chaetotaxy of a Hairy
Caterpillar, Lymantria marginata Wlk.
(Lepidoptera: Lymantriidae)

Jasvir Singh*

Department of Zoology, Sanatan Dharm College, Muzaffarnagar (UP)-251001, India.

Abstract. The head capsule sclerites and cephalic chaetotaxy of first
instar larva of Lymantria marginata Wlk. are described. Primary setae
and punctures of first instar are constant in number and position. Each
half of the capsule is characterized by the presence of 21 primary setae
and 10 punctures. The A3 seta of the anterodorsal group is the longest
seta on the cranium. Additional secondary setae are recorded for each
successive instar. Most setae were added in second and third instars.

The punctures and number of setae of vertical, genal and clypeal groups
remained unchanged throughout larval development.

Introduction

Dyar (1896), Forbes (1910), Gardner (1938), Hinton (1946, 1947),
Bucker (1967), Goel and Kumar (1981) and Kumar and Goel (1986) have
described cephalic chaetotaxy of several lepidopteran larvae of taxo-
nomic significance. Most work deals with non-hairy caterpillars for
simplicity. The first instar larva presents a generalized plan of tubercles
and setae in Lepidoptera. However, hairy caterpillars have shown
remarkable differences from the first instar larva into their subsequent
instars. The changes for each instar need to be explored for a proper
understanding of the morphological variations between species. The
cephalic chaetotaxy of Lymantria marginata Wlk., a defoliator of mango
in North India (Singh and Goel, 1986), is described. The morphological
changes involved during successive instars are also elucidated.

Materials and Methods

The caterpillars from first to seventh instars in laboratory culture were killed
and fixed in KAAD mixture (Peterson, 1962) and stored in 70% alcohol. The head
was removed after placing the larva in heated 5% KOH for 2 minutes. The
material was washed in distilled water, dehydrated and stained with hematoxy-
lin, and mounted in Canada balsam. Exuviae from each instar were also mounted
in Canada balsam and used for confirming setal patterns and punctures.
Illustrations were done with the aide of a camera lucida on high power microscope
(Olympus). The setal numbers groupwise on the head capsule were counted from
first to seventh instars and their morphological changes at each moult were
recorded. Five head capsules representing each instar were observed for setal
counting. The distribution of setae is given in Table 1 for different instars on each
half of the cephalic region. Hinton’s (1947) terminology is followed.

^Present Address: Indian Cardamom Research Institute, Spices Board, Sakleshpur
(Karnataka)-573134, India.

30(3-4):272-278, 1991

273

Table 1 . Distribution of primary setae of first instar larva and additional secondary
setae for different instars on each half of the cephalic region of Lymantria

marginata Wlk.

Setal group

1

II

Instar stages

III IV V

VI

VII

Vertical group (V)

3

3

3

3

3

3

3

Clypeal group (C)

2

2

2

2

2

2

2

Genal group (G)

1

1

1

1

1

1

1

Posterodorsal group (P)

2

17

27

29

32

29

30

Lateral group (L)

1

2

2

3

2

3

3

Adfrontal group (AF)

2

2

7

6

7

6

6

Frontal group (F)

1

3

5

7

6

9

10

Anterodorsal group (A)

3

14

28

27

23

25

26

Ocellar group (0)

3

5

12

11

9

10

11

Subocellar group (SO)

3

5

7

9

8

9

10

Total

21

54

94

98

93

97

102

Results

The head capsule: The head capsule of L. marginata is heavily
sclerotized, dark brown, elliptical and compressed cephalocaudally in a
hypognathus orientation. The two upper region large sclerites, parietal
or vertex (PT) of the epicranium are demarcated by an inverted Y-shaped
coronal adfrontal sulcus (Fig.l). The coronal stem(CS) splits ventrally
into two adfrontal (AF) cleavage lines. Each adfrontal sulcus terminates
anteroventrally at the epicondyle (EC), a place for dorsal articulation of
the mandible of that side (Fig.l). The triangular sclerite enclosed be-
tween lateral arms of frontoclypeal suture is the frons (FR). The tentorial
arms (TA) arise from the lateral adfrontal sutures to internally strengthen
the head capsule. The clypeus (CL) lies in front of the frons and a thin
cuticular anteclypeus (AC) is attached to it for articulation of the labrum
(LB). Because of the weak adfrontal cleavage line, the narrow strip
between the frontoclypeus and parietals denotes the adfrontal area (AF)
which appears inflected into the adfrontal sulcus (ADR) (Fig.2). The
parietal of either side posteriorly forms the genal area (GA) which
supports the triangular hypostomal area (HA). Both areas are separated
by hypostomal sutures (HS) and extend mesially into hypostomal process
(HP) to connect to a similar process of the other side forming a membra-
nous hypostomal bridge (HB) (Fig.3). There are six ocelli (OC1-OC6) on
either side of the cranium posterolateral near the antennal socket (AS)
(Fig.4). The cranium posterodorsally is perforated by a large occipital
foramen (OF) (Fig.3).

Cephalic chaetotaxy: The head capsule of the first instar L. marginata
is characterized by the presence of 21 primary setae of which 17 are long

274

J. Res. Lepid.

Figs. 1-4. Head capsule of first instar caterpillar of Lymantria marginata
Fig. 1 . Anterior view

Fig, 2. Frontoclypeus with labrum (a) Frontal view (b) Inner view

Fig. 3. Posterior view

Fig. 4. Parietal (lateral view)

Aa Anterodorsal puncture
A1-3 Anterodorsal setae
AC Anteclypeus
ADR Adfrontal sulcus
AF Adfrontal
AFa Adfrontal puncture
AF1-2 Adfrontal setae
AS Antennal socket
Cl -2 Clypeal setae
CL Clypeus
CS Coronal stem
EC Epicondyle
Fa Frontal puncture

Abbreviations

FI Frontal seta
FR Frons

FS Frontoclypeus sulcus
Ga Genal puncture
G1 Genal seta
GA Genal area
HA Hypostomal area
HB Hypostomal bridge
HP Hypostomal process
HS Hypostomal socket
La Lateral puncture
LI Lateral seta
LB Labrum

Oa Ocellar puncture
01-3 Ocellar setae
0C1-6 First to six ocelli
OF Occipital foramen
Pa,b Posterodorsal punctures
PI -2 Posterodorsal setae
PT Parietal sclerite or vertex
SOa Subocellar puncture
SOI -3 Subocellar setae
TA Tentorial arm
Va Vertex puncture
VI -3 Vertex setae

30(3-4):272-278, 1991

275

and 4 minute. On each half there are 10 punctures. The clypeus bears two
clypeal setae Cl and C2 without punctures. The frontal group single seta
(FI) lies closer to frontal suture and puncture (Fa) closer to middorsal
line passing through the coronal stem (Fig.2). The adfrontal region has
two setae, AF1 and AF2, and puncture AFa along the adfrontal suture.

The setae Al, A2, A3 and puncture Aa are present in the anterodorsal
group of the head capsule and lie between the adfrontal and ocellar
groups. Seta A3 is the largest seta on the head capsule and lies closer to
OC2 and slightly posterolaterad of A2. The setae organized by length are
A3>A2>A1. The ocellar group bears two almost equal sized setae 01 and
02 and puncture Oa. The seta 01 lies within the ocellar area and seta 02
is posteroventrad to 01, while 03 is ventrad of 0C6. The puncture Oa is
closer and anterior to 0C4 (Fig.4). The seta SOI is close to antennal
socket, S02 is between 0C5 and 0C6, and S03 is posterior to SOI (all
subocellar group). Relative lengths are S02>S03>S01. The puncture
SOa is close to 03 anteriorly. The genal group possesses a small minute
seta G1 and a puncture Ga, both posteroventrad to 0C6.

The lateral group has a single seta LI and puncture La. The seta LI is
anterolaterad to PI, and puncture La is posterodorsad to LI. The two
setae (PI and P2) and two punctures (Pa and Pb) are part of the
posterodorsal group. The seta PI is the second largest on the head
capsule and lies posterolaterad to AF2, and P2 is posterodorsad to PI.
The puncture Pa lies ventrad to PI, whereas Pb is anteroventrad to P2.
The vertical group is characterized by the presence of three minute setae
VI, V2, V3 and puncture Va. There is a small gap between the three setae
and puncture Va lies equidistant between V2 and V3.

Distribution of secondary setae in instars I-VTI : Besides the
primary setae, numerous additional setae are observed on each half of
the head capsule from the second through final instars, giving a tufted
look to the caterpillar. However, the position of primary setae and
punctures remains unchanged; primary setae have more prominent
setal sockets compared with the additional setae. There are no additional
punctures observed on the head capsule of any larval instar. An asym-
metrical condition has been observed for secondary setae. Many addi-
tional setae occur in second and third instar larvae whereas the number
of additional setae during the remaining instars increase slightly from
third instar (Table 1).

The first instar larva has 21 primary setae but due to the regular
increases of secondary setae at each instar, the seventh instar has 102
setae. The vertical (V), clypeal (C) and genal (G) are the only groups that
do not deviate in number from first to last instar. It is notable that
wherever there are increases in the AF group during the third and fifth
instars and decreases in the fourth, sixth and seventh instars, than a
corresponding decrease in L group setae occurs during the third and fifth
instars and vice-versa in the fourth, sixth and seventh instars. The
posterior region of the frons, however, has no additional setae. The

276

J. Res. Lepid.

maximum setal numbers were added to the posterodorsal (P) and
anterodorsal (A) groups, i.e. 56% of the total number of setae (559) from
first to the seventh instar.

Discussion

The sclerites are like those of most ditrysian Lepidoptera. Snodgrass
(1935) described the triangular frontal part as the clypeus while Hinton
(1947) called it the frontoclypeal apotome. Walker (1931) applied the
names anteclypeus and postclypeus. In the present study the frons,
clypeus and anteclypeus, by their clear demarcations are considered
separate. DuPorte (1946) described the epicranial suture as a line of
weakness which normally splits open during ecdysis (Hinton, 1947). InL.
marginata and most Lepidoptera, especially butterflies, the coronal stem
splits only at the time of pupation and not at each stadial ecdysis.

Number of setae can be used to identify the instar ofL. marginata. The
size of the frontoclypeus relative to the head capsule was used to identify
instars by Singh ( 1956) and Mathur and Singh ( 1963), but the proportion
of the frons and coronal stem is almost equal in first and second instar of
L. marginata. The adfrontal area is less distinct in L. marginata, as
described by Crampton (1921) than in some other Lepidoptera.

The number and position of setae and punctures on head capsules ofL.
marginata appears typical of Lepidoptera as described by Heinrich
(1916), Hinton (1946), Mukerji and Singh (1951), Mathur and Singh
(1963), Bucker ( 1967) and Goel and Kumar (1981). A few other studies of
lymantriids are available. The same position of seta FI and puncture Fa
was observed by F orbes ( 19 10) in a lymantriid, Hemerocampa leucostigma.
Seta A3 is the longest seta on the cranium in L. marginata whereas
Hinton (1946) described PI as generally the longest seta of the cranium.
Bucker (1967) reported A2 as the longest seta in his study of four
lymantriid species.

The ocellar and subocellar groups of L. marginata are like those
reported by Mathur and Singh (1963) in pyralid larvae and by Bucker
(1967) in lymantriids. The genal group of L. marginata has a single
minute seta and puncture. The same condition agrees with Heinrich
(1916), Ripley (1923), and Downey and Allyn (1979). Hinton (1946) has
described two setae G1 and G2 and one puncture (Ga), instead of one seta,
in Lepidoptera. One seta (Gl) and two punctures (Ga and Gb) have been
reported by Goel and Kumar (1981) in a arctiid, Diacrisia obliqua (Wlk.).

Hinton (1946) differentiated the Lymantriidae as a specialized family
of the ditrysian group by the presence of secondary setae or punctures
following the first or succeeding moults. Being prominent and smooth,
the primary setae hardly lose their identity from the surrounding
secondary setae even after first instar. The punctures and setae of the
vertical, genal and clypeal groups remained unchanged throughout the
larval period of L. marginata in position and number. However, no setae
were detected in the upper part of epicranium and frons in this study.

30(3-4):272-278, 1991

277

Forbes (1910) described the upper portion of the epicrania and frons as
having no additional setae in the lymantriid Euproctis chrysorrhoea. The
part of the vertex and gena which retracts into the prothorax possessed
three minute setae on the vertex and on the caudal part of each side of the
gena. According to Hinton ( 1946), the minute setae were positioned in the
overlapping parts and functioned as proprioceptors as those on the
thorax and abdomen. Likewise, the presence of the greatest number of
secondary setae are in the posterodorsal and anterodorsal groups in L.
marginata. The long setae, which undoubtedly function for tactile pur-
poses, are rather evenly distributed over the anterior and exposed parts
of the head.

Acknowledgements. The Department of Science and Technology, New Delhi
provided financial support and Dr. S.C. Goel gave valuable guidance and
constant encouragement.

Literature Cited

Bucker, A.H.A. 1967. Studies on the chaetotaxy of a few lymantriids. Bull. Ent.
8: 36-42.

Crampton, G.C. 1921. The sclerites of the head and the mouth parts of certain
immature and adult insects. Ann. Ent. Soc. Am. 14: 65-110.

Downey, J.C. & A.C. Allyn. 1979. Morphology and biology of the immature
stages of Leptotes cassius theonus (Lucas) (Lepidoptera:Lycaenidae). Bull.
Allyn. Mus. 55:1-27.

Duporte, E.M. 1946. Observations on the morphology of the face in insects. J.
Morph. 79: 371-418.

Dyar, H.G. 1896. Notes on the head setae of lepidopterous larvae, with special
reference to the appendages of Perophora melsheimerii. Jr. N.Y. Ent. Soc. 4:
92-93.

Forbes, W.T.M. 1910. A structural study of some caterpillars. Ann. Ent. Soc.
Am. 3: 94-143.

Gardner, J.C.M. 1938. Immature stages of Indian Lepidoptera (I) Lymantriidae.
Indian For. Rec. 3: 187-212.

Goel, S.C. & A. Kumar. 1981. Cephalic demarcations and chaetotaxy of a larvae
Diacrisia obliqua (Wlk.) (Arctiidae). Uttar Pradesh J. Zool. 1: 6-15.
Heinrich, C. 1916. On the taxonomic value of some larval charcters in the
Lepidoptera. Proc. Ent. Soc. Wash. 18: 154-164.

Hinton, H.E. 1946. On the homology and nomenclature of the setae of
lepidopterous larvae with some notes on the phylogeny of the Lepidoptera.
Trans. R. Ent. Soc. Lond. 97: 1-37.

1947. The dorsocranial areas of caterpillars. Ann. Mag. Nat. Hist. 14: 843-

852.

Kumar, V. & S.C. Goel. 1986. Cephalic chaetotaxy of last instar caterpillar of
Plusia orichalcea Fabr. (Noctuidae). Uttar Pradesh J. Zool. 6: 40-44.
Mathur, R.N. & P. Singh. 1963. Immature stages of Indian Lepidoptera No. 13,
Pyralidae, subfamily Pyraustinae. Indian For. Rec. 10: 117-148.

Mukerji, S. & H. Singh. 1951. Studies on the chaetotaxy of larvae of Plusia sp.
(Lepidoptera:Phalanidae). Proc. R. Ent. Soc. Lond. 20(B): 15-24.

278

J. Res. Lepid.

Peterson, A. 1962. Larvae of insects. Part I. Columbus, Ohio. 315pp.

Ripley, L.B. 1923. The external morphology and postembryology of noctuid
larvae. 111. Illinois Biol. Monogr. 8: 1-169.

Singh, B. 1956. Some more Indian geometrid larvae (Lepidoptera) with a note
on the identity of components of various groups of setae. Indian For. Rec. 9:
136-163.

Singh, J. & S.C. Goel. 1986. Biology of Lymantria marginata Wlk.
(Lyman triidae:Lepidoptera), a mango defoliator in Western Uttar Pradesh.
Entomon 11: 265-267.

Snodgrass, R.E. 1935. Principles of Insect Morphology. McGraw Hill, New York.
ii+667pp.

Walker, E.M. 1931. On the clypeus and labium of primitive insects. Canadian
ent. 43: 72-81.

Journal of Research on the Lepidoptera

30(3-4):279-288, 1991

Distribution and Flight Behaviour of the
Junglequeen Butterfly, Stichophthalma louisa
(Lepidoptera: Nymphalidae), in an Indochinese
Montane Rainforest

Vojtech Novotny, Martin Tonner and Karel Spitzer

Institute of Entomology, Czech Academy of Sciences, Branisovska 31, 370 05 Ceske
Budejovice, Czech Republic

Abstract. A population of Stichophthalma louisa (Lepidoptera,
Nymphalidae) was studied in a montane climax rain forest in Vietnam.

The species was confined to the closed forest; open sites represented a
barrier to its dispersal. The distribution within the forest was primarily
determined by adult food sources (sap exuding from wounded trees of
various species), which varied unpredictably in location and quality. As
S. louisa tracked these resources, it was equally variable in its spatial
distribution. Most of the variability in distribution was associated with
feeding places and the network of gullies, “flyways” along which the
butterflies flew. The turnover of individuals within any area (either a
food place or the surrounding forest) was high. By the irregular and
wide movements of individuals within the population, S. louisa is
dissimilar to other tropical forest butterflies which have more predict-
able spatial structure of populations.

She came flying out of the wood over yonder — How fast those Queens can
run! ... I’ll make a memorandum about her, if you like — She’s a dear good
creature.

Lewis Carroll

Through the Looking Glass

Introduction

For pragmatic reasons most population studies on Lepidoptera have
been on small populations of sedentary species confined to distinct and
small “islands” of suitable habitat. Structurally simple and open temper-
ate zone habitats, meadows in particular, are most frequently studied. In
this study, the population structure of a forest butterfly Stichophthalma
louisa , which is confined to a large area of a structurally complex
montane monsoon forest, is examined.

Stichophthalma louisa Wood-Manson 1877 is a montane species with
the distribution from Upper Burma, northern and central Thailand to
North Vietnam. The isolated, little known and highly variable popula-
tions of northern Vietnam appear to form several local geographical races
(Fruhstorfer, 1927). Most of specimens from the study site, the Tam Dao
Mts., can be probably referred to S. louisa sparta de Niceville 1894
(Niceville, 1894; Okano, 1985; P. Ackery pers. com.), but typical “louisa”
and intermediates are also present. The life history is unknown. Most

280

J. Res. Lepid.

species of the genus Stichophthalma are probably univoltine with larvae
associated with monocotyledonous plants (Fruhstorfer, 1927). According
to Ackery (1988), larvae of a related Chinese species S. howqua Westw.
feed on Phyllostachys spp. (Poaceae).

Methods

STUDY SITE

The study site was a large remnant of a montane rain forest (19,000 ha) in the
Tam Dao Mts. (N. Vietnam; 21° 30’ N, 105° 40' E; 800 - 1000 m alt). It is
floristically very rich, without any conspicuously dominant tree species. The tree
layer formed a dense continuous canopy, approximately 10 m high, without
emergent trees, but with epiphytes and climbers (Figs. 1 and 2). Bamboo thickets
dominated on ridge tops and disturbed sites, the shrub and ground layers were
fragmented (Spitzer et al. 1993).

As steep slopes and dense vegetation made movement through the forest
difficult, observations on the population of S. louisa were confined to a 5-meter
wide belt on either side of a narrow path tracking a contour at 800 m asl. The path
was only about 1 meter wide and did not affect the composition of the surrounding
vegetation. It intersected an extremely steep slope (30° - 40°) so the canopies of
the trees growing down-hill could be observed. A transect of 2 km was demarcated
along this path, arbitrarily divided into 200 10-meters long sections numbered 1
- 200. The transect intersected three habitat types: closed forest (sections 1-112,
1120 meters), ecotone with forest on the down-slope and a deforested area on the
up-slope (113- 167, 550 meters), and ruderal ( 168 - 200, 330 meters). The ruderal
zone was a mosaic of small cultivated fields and abandoned land overgrown by

Fig. 1 The Tam Dao mountains covered by rain forest.

30(3-4):279-288, 1991

281

tall grasses. Butterfly communities along the transect were described by Leps
and Spitzer (1990) and Spitzer et al. (1993).

Along the forest transect, five habitat types were distinguished among the 112
forest sections: closed forest, gullies, bamboo thickets, glades, and feeding trees.
Shallow rocky gullies with either permanent or temporary brooks were numer-
ous within the forest. As the transect intersected the slope horizontally, the
gullies were perpendicular to it. They represented “tunnels” within the vegeta-
tion, with closed canopy above. Dense bamboo thickets covered small-scale
disturbed patches, usually the result of selective logging. Glades were open
patches ranging from tens to hundreds of square meters, either rocky or
overgrown by grasses. They were either of natural origin or from human
disturbance. Feeding trees were trees which exude sap and that are used as the
major nutrient resource by adult butterflies.

Our observations were made daily, between 10.00 and 18.00 hour, from 5 June
to 7 July 1991. This was the first half of a rainy season (rainy season starts in June
and ends in October).

BUTTERFLY DENSITY, DIURNAL ACTIVITY, AND MOVEMENTS

The number of *S. louisa was estimated by direct counts of individuals flying or
resting on the vegetation within the 10-meter wide belt along the forest transect
(sections 1 - 112). Altogether, 65 transect counts were conducted. Our observa-
tions depended on butterfly activity recognizing that some of the resting, but
none of the flying, individuals were overlooked. Accordingly only results under
optimum weather conditions provided reliable density estimations.

The number of butterflies recorded flying per unit time during the forest
transect counts was an index of activity used to evaluate diurnal activity of S.

Fig. 2 Montane monsoon forest in the Tam Dao Mts., the habitat of Stichophthalma
louisa.

282

J. Res. Lepid.

louisa. Altogether, data on 389 specimens recorded between 10.00 and 18.00 hour
over 32 hours of observation were analysed.

Movement of S. louisa was examined by a capture-mark-recapture method
between 8 June and 7 July 1991 (see Fig. 3). Because of the dense forest
vegetation, butterflies were captured and marked only along the transect, which
was effectively a one-dimensional linear structure across a two-dimensional
population continuum, not permitting home range, population size and other
demographic estimations.

Results

BUTTERFLY BEHAVIOUR

Stichophthalma louisa is cryptic, the upper and undersides of their
wings resemble dead leaves. When resting on vegetation, butterflies
close their wings. When flying, they slowly glide in a zigzag course within
the vegetation, resembling falling leaves. Encounters between two but-
terflies frequently resulted in a “contest” during which the couple
repeatedly spiralled upwards with the butterflies clinging together.

As other amathusiines, S. louisa was never seen visiting flowers. The
only observed source of food for butterflies was sap bleeding from injured
trees. When feeding, butterflies ignored one another, frequently sitting
side by side in a vicinity of the sap flow.

SEASONALITY

The seasonal dynamics of the Tam Dao population can only be inferred
from our short-term observations made at various times over several

Fig. 3 Marked specimen of Stichophthalma louisa.

30(3-4):279-288, 1991

283

years. This information indicates a distinct seasonality with the occur-
rence of adults restricted to the June to September rainy season, with the
highest abundance during June - July. No adult butterflies were ob-
served in May, October, and December. The present study coincides with
the period of high butterfly abundance following peak emergence. Freshly
emerged specimens were observed throughout the course of the study,
but the proportion steadily decreased from 87 % at the very beginning (5
- 10 June) to 36 % between 23 and 28 June.

DIURNAL ACTIVITY

The activity varied widely from 5 to 34 butterflies observed per hour
and did not reveal a consistent daily pattern. S. louisa was a typical
diurnal species, not crepuscular like some other amathusiines. The
periodicity observed on any given day possibly followed fluctuations in
weather.

VERTICAL DISTRIBUTION

Stichophthalma louisa had a distinct vertical distribution, being con-
fined mostly to the canopy and top of the shrub layer with a flight height
range 2 - 8 m. The population density of S. louisa observed in the shrub
understorey vegetation (0 - 2 m) close to the transect was approximately
1.4 butterflies per hectare. This value is ten times lower than that for the
canopy (see below). S. louisa was never observed to fly above the canopy
layer.

HABITAT PREFERENCE

The number of butterflies observed along the transect is summarized
in Table 1. Altogether, 65 transect counts were made and each butterfly
seen was assigned to its transect section. S. louisa appeared to be
restricted to the forest, its edge represented a barrier that was rarely
crossed. Glades within the forest habitat, although small, were avoided.
Butterflies were also rarely encountered in patches of very dense vegeta-
tion, especially bamboo thickets. On the other hand, butterflies were
observed flying more frequently along gullies and brooks than in the
surrounding dense forest.

The proportion of the transect sections classified as a particular habitat
type gives a rough idea of its representation within the forest (Table 1).
Closed canopy forest covered about 77 % of the transect, 13 % was covered
with bamboo thickets and the remaining 10 % were open glades. Along
the 1120 m of the transect there were 15 gullies (spanning over 19
sections), five of which were bisected by a permanent brook. In the forest
natural glades were scarce and small. Such glades were common along
the transect due to man-made clearings. These together spanned over 10
transect sections. Naturally produced feeding trees had a density of
about 1.2 per hectare, with two out of five feeding trees bleeding sap due
to man-made injuries.

284

J. Res. Lepid.

Of 812 butterflies recorded, 38 % were found within the closed forest,
25 % flying along gullies, and 29 % in sectors with a feeding tree. In terms
of butterfly distribution, glades and bamboo thickets were not important
as each of these habitats harboured less than 5 % of the population.

The variance/mean ratio for the closed forest, s2/x = 2.1, is fairly close
to unity, although significantly different from it (x2 test, P>0.05). This
indicates that the distribution of S. louisa in the closed forest is close to
random, not clumped (Poisson distribution has the variance/mean ratio
equal to unity). Accordingly, no further differentiation of the closed forest
habitat type into subhabitats is probably needed.

Dispersion of S. louisa within each habitat is expressed by the Lloyd’s
index of mean patchiness (Pielou, 1976) as given in Table 1. Distribution
of butterflies among 19 gullies and brooks was highly heterogeneous.
There were five sections used as flyways with densities ranging from 16
to 22 butterflies. Another 14 sections showed densities not different from
the closed forest habitat (1-9 butterflies per section, mean 4.5). The
occurrence through time of butterflies in the high-density sections was
uneven and fluctuated independently in the various sections. This is
shown in Fig. 4, which depicts the numbers of butterflies observed each
day over one month sampling period across three adjacent flyways
(sections 77, 74, 70).

Feeding trees differed in their attractiveness for S. louisa , indicated by
a high value of Lloyd’s index. The attractiveness of any given feeding tree

Table 1 . Habitat and microhabitat preferences of Stichophthalma louisa.

Habitat

n

N

X

s2

Forest

112

812

7.2 a

158.2

Ecotone

55

31

0.6 b

1.1

Ruderal

33

4

0.1 b

0.3

Forest microhabitat

n

N

X

s2

L

Feeding trees

5

236

47.2 a

1279.7

1.6

Gullies and brooks

19

203

10.7 b

141.8

2.2

Closed forest

62

293

4.7 c

10.1

1.2

Bamboo thickets

14

34

2.4 d

2.6

1.0

Glades

12

28

2.3 d

4.4

1.4

All habitat types are characterized in the text. The forest habitat is further subdivided
into 5 microhabitats in the lower part of the table, n = number of 1 0-meters section
in the transect. Total number of butterflies recorded (N), arithmetic mean (x), and
variance (s2) of butterfly numbers per section, and Lloyd’s index of mean patchiness
(L) are given for each habitat. Values in lines followed by the same letter are not
significantly different (P>0.05; Kruskal-Wallis test).

30(3-4):279-288, 1991

285

as a food source varied over time. This variability was not correlated with
changes in butterfly numbers on other feeding trees or in the surround-
ing forest and neighbouring gullies used as flyways (Spearman correla-
tion coefficient between number of butterflies observed per day on a
feeding tree and in surrounding forest were not significant, P»0.05, for
all feeding trees).

Stichophthalma louisa did not create large aggregations in any forest
habitat type. On feeding trees, no more than 6 specimens were ever
present simultaneously. No hilltopping or response to other obvious
landmarks were observed, although all neighbouring prominent sum-
mits were inspected. The butterfly was never observed to indulge in
puddling behaviour. At night, the butterflies roosted singly in the
vegetation. During the 65 sampling periods of the forest transect sec-
tions, 8 specimens was the maximum recorded at any one time.

POPULATION DENSITY AND MOVEMENTS

The number of S. louisa counted during any transect sample ranged
from 1 to 26 (mean 18.4). Since the transect represents an area of 11,200
sq. m. (1120 m x 10 m), the mean population density of S. louisa in the
forest was 16.4 individuals per hectare. The observed density remained
constant over the study period.

A total of 235 butterflies were marked and released, with 524 butter-
flies checked for marks from 8 June though 7 July. Only 23 previously
marked butterflies were recaptured. The low recapture frequency indi-
cates that butterfly mortality was high and/or they were very transient,

<n

0)

CD

13

CO

CD

.Q

E

June

Fig. 4 The number of Stichophthalma louisa observed flying along gullies. No records
were made on 1 6 and 22 June due to bad weather. A section 77, gully with
a brook; B section 74, dry gully; C section 70, dry gully.

286

J. Res. Lepid.

moving widely across the forest. The possibility that butterflies react to
marking by leaving the study area cannot be discounted, but no such
escape behaviour was observed in freshly marked butterflies. High
mortality cannot alone be an explanation as a high percentage of
specimens had worn wings and recaptured individuals showed a low
decay rate. Our data do not permit discriminating between these pro-
cesses.

Stichopthalma louisa is able to move over considerable distances in a
short time. The average distance travelled by recaptured butterflies was
122 m, with the longest travelled in a single day of 460 m. Another
butterfly was seen visiting two feeding trees 270 m apart during a 15
minute interval. The longest time elapsed between a capture and recap-
ture of a butterfly was 17 days.

Thirty three butterflies were marked on the feeding tree in section 31,
but none was recorded again among the total of 107 specimens observed
at the section. Between 17 and 21 June, 39 butterflies were recorded on
the feeding tree in section 7. They all were captured and marked. Among
them, only one butterfly visited this tree repeatedly. This suggests a low
frequency of repeated visits by a particular butterfly with a high turnover
rate of individuals on particular feeding tree.

Discussion

The few tropical forest butterflies studied thus far have a rigid popu-
lation structure with fixed home ranges and/or a predictable pattern of
movements ( Heliconius spp.: Gilbert & Singer, 1975; Morpho spp.:
Young, 1971, and references therein). The data obtained in this study
indicate that S. louisa individuals move widely and irregularly.

A general picture of movement of S. louisa within the study area
emerges. The occurrence of butterflies was primarily determined by the
availability of food sources, which were unpredictable and varied in
location and quality. Consequently, the S. louisa population, which
tracked these resources, was equally variable in its spatial distribution.
Most of this variability was associated with the distribution of feeding
places and the network of gullies along which flight paths of the
butterflies were concentrated. In contrast, within the remaining area of
the closed forest the dispersion of the population was fairly close to
random and did not vary over time. The turnover of individuals within
any area, either food sites or the forest, was high.

A wide spectrum of spatial patterns and flight behaviour has been
reported in butterflies (e.g. Gilbert & Singer, 1975). The two principal
factors that determine flight behaviour are the species’ reaction to
patches of unsuitable habitat and mobility of individuals within a
suitable habitat. Areas of unsuitable habitat either do not represent
barriers to movement so that any given population may be distributed
across several “habitat islands” (e.g. Erehia epipsodea : Brussard &
Ehrlich, 1970a, b, c), or else the unsuitable habitat may be barrier to the

30(3-4):279-288, 1991

287

dispersal of populations (e.g. Heodes uirgaureae : Douwes, 1975;
Euphydryas editha : Ehrlich, 1965; Mellicta athalia : Warren, 1987).
Within a suitable habitat, individuals are restricted by their mobility to
a particular area that may be either fixed or changing over time. A home
range may be fixed by some recognition behavior ( Philotes sonorensis:
Keller et al., 1966), by use of a visual landmark as a meeting point for
mating ( Oeneis chryxus : Daily et al. 1991), as a site for gregarious
nocturnal roosting ( Marpesia berania : Benson & Emmel, 1973), or by a
particular distribution of feeding sources which a butterfly exploits
(. Heliconius ethilla : Ehrlich & Gilbert, 1973).

Of butterflies studied so far, S. louisa belongs to those species in which
the population is kept within an area by strict avoidance of unsuitable
habitats, but individuals move widely and irregularly within favourable
habitats. There were no constant features in the spatial pattern resem-
bling the regular repeated flights of individuals along particular routes
as in Heliconius spp. (Gilbert & Singer, 1975). Ehrlich & Gilbert (1973)
speculate that the “scheduled” flights of Heliconius represent optimum
foraging routes. Clearly, there are no feeding routes that are perma-
nently optimum for S. louisa due to temporal variability in its primary
food source. Because of this, the distribution patterns of Heliconius and
S. louisa populations differed greatly although both were determined
mainly by the pattern of adult food resources, a pattern found in other
species of butterflies (e.g. Brakefield, 1982; Gilbert & Singer, 1973;
Douwes, 1975).

Acknowledgements. This work was supported by the International Council for
Bird Preservation and Flora & Fauna Preservation Society grant. We thank R.
H. T. Mattoni, A. F. G. Dixon, J. Leps, P. R. Ackery, and an anonymous referee
for their invaluable comments and advice on this manuscript.

Literature Cited

Ackery, P.R. 1988. Hostplants and classification: a review of nymphalid
butterflies. Biological Journal of the Linnean Society 33:95-203.

Benson, W.W. & T.C. Emmel. 1973. Demography of gregariously roosting
populations of the nymphaline butterfly Marpesia berania in Costa Rica.
Ecology 54:326-335.

Brakefield, P.M. 1982. Ecological studies on the butterfly Maniola jurtina in
Britain. I. Adult behaviour, microdistribution and dispersal. Journal of
Animal Ecology 51:713-726.

Brussard, P.F. & P.R. Ehrlich. 1970a. The population structure of Erebia
epipsodea (Lepidoptera: Satyrinae). Ecology 51:119-129.

. 1970b. Contrasting population biology of two species of butterfly. Nature

227:91-92.

. 1970c. Adult behavior and poulation structure in Erebia epipsodea

(Lepidoptera: Satyrinae). Ecology 51:880-885.

Daily, G.C., P.R. Ehrlich & D. Wheye. 1991. Determinants of spatial distribution
in a population of the subalpine butterfly Oeneis chryxus. Oecologia 88:587-
596.

288

J. Res. Lepid.

Douwes, P. 1975. Distribution of a population of the butterfly Heodes virgaureae.
Oikos 26:332-340.

Ehrlich, P.R. 1965. The population biology of the butterfly, Euphydryas editha.
II. The structure of the Jasper Ridge colony. Evolution 19:327-336.

Ehrlich, P.R. & L.E. Gilbert. 1973. Population structure and dynamics of the
tropical butterfly Heliconius ethilla. Biotropica 5:69-82.

Fruhstorfer, H. 1927. Amathusiidae. In: A. Seitz (ed.): Die indo-australischen
Tagfalter - Die Grossschmetterlinge der Erde. Bd. 9. A. Kernen, Stutgart, pp
403-452. butterfly species. American Naturalist 107:58-72.

Gilbert, L.E. & M.C. Singer. 1973. Dispersal and gene flow in a butterfly
species. American Naturalist 107: 58-72.

. 1975. Butterfly ecology. Annual Review of Ecology and Systematics 6:365-

397.

Keller, E.C., R.H.T. Mattoni & M.S.B. Sieger. 1966. Preferential return of
artificially displaced butterflies. Animal Behavior 14:197-200.

Leps, J. & K. Spitzer. 1990. Ecological determinants of butterfly communities
(Lepidoptera, Papilionoidea) in the Tam Dao Mountains, Vietnam. Acta
Entomologica Bohemoslovaca, 87:182-194.

Niceville, de L. 1894. New or little-known butterflies from the Indo-Malayan
Region. Journal of the Asiatic Society of Bengal 63:1-59.

Okano, K. 1985. Descriptions of four new butterflies on Amathusiidae
Nymphalidae and Papilionidae (Lepidoptera). Tokurana 10:1-17.

Pielou, E.C. 1976. Mathematical Ecology. Wiley & Sons, New York.

Spitzer, K., V. Novotny, M. Tonner & J. Leps. 1993. Habitat preferences,
distribution and seasonality of butterflies (Lepidoptera, Papilionoidea) in a
montane monsoon forest. Journal of Biogeography, 20: xxx [in press]

Warren, M.S. 1987. The ecology and conservation of the heath fritillary
butterfly, Mellicta athalia. II. Adult population structure and mobility.
Journal of Applied Ecology 24:483-498.

Young, A.M. 1971. Wing coloration and reflectance in Morpho butterflies as
related to reproductive behavior and escape from avian predators. Oecologia
7:209-222.

Journal of Research on the Lepidoptera

30(3-4):289-296, 1991

On the functional foreleg tarsus in Caerulea males
(Lepidoptera: Lycaenidae: Polyommatini)

Rudolf H. T. Mattoni1 and Konrad Fiedler2

Abstract. With very few exceptions, males in the butterfly family
Lycaenidae have fused foreleg tarsomeres lacking distal claws. Func-
tional five-segmented male foretarsi with claws are for the first time
reported from the species-rich Lycaenidae tribe Polyommatini. In the
Glaucopsyche section, both species of the Sino-Tibetian endemic subge-
nus Caerulea ( coeligena Oberthiir 1876, coelestis Alpheraky 1897)
uniquely share this unusual trait. The phylogenetic significance of that
character within the Lycaenidae is discussed.

Leg structures of arthropods often yield valuable characters for phylo-
genetic inferences. Among the Lepidoptera, leg morphology has been
amply used in attempts to resolve questions concerning the higher
classification. Recent examples for the Papilionoidea, or true butterflies,
were published by Robbins (1988a & b, 1989). A five-segmented tarsus
with one distal pair of claws on all legs constitutes the groundplan of
butterfly tarsal structure. This groundplan is preserved in the families
Papilionidae and Pieridae, whereas in the Riodinidae and Nymphalidae
(sensu Harvey 1987, 1991) foretarsi of both sexes show various degrees
of reduction (females) or fusion (males) of tarsomeres and bear no claws.
In particular, forelegs are not used for walking in these two families.

In the family Lycaenidae a different pattern prevails. Female lycaenids
always possess functional and fully segmented foretarsi, whereas males
of almost all species have entirely fused tarsomeres and foretarsal claws
are absent (so-called ankylosed tarsi). However, in both sexes the
foretarsi are used in walking. As a rare exception (ca. 70 out of roughly
4500 described species, cf. Bridges 1988), segmented foretarsi with claws
occur in males of a few Lycaenidae genera. The taxa recorded thus far are:
Liphyra, Euliphyra, Aslauga, Paraslauga, Egumbia, Lachnocnema and
Thestor (together constituting the predominantly Ethiopian tribe
Liphyrini of the subfamily Miletinae); the East Asian genera Artopoetes,
Japonica, Ussuriana, Coreana and Protantigius (all in the Theclini-
Thecliti); the New Guinean Titea (Theclini-Luciiti); one species of the
Oriental genus Pratapa and the Bornean endemic Sukidion inores
(Eumaeini-Iolaiti); and the Neotropical Theclopsis (Eumaeini-Eumaeiti;
see Eliot 1973, Robbins 1988b; higher classification largely following
Scott & Wright 1990 and Fiedler 1991). We here report on the first
observation of males with functional foretarsi in the large cosmopolitan
tribe Polyommatini, representing more than 1000 described species.

1 The Lepidoptera Research Foundation, 9620 Heather Road, Beverly Hills, CA 90210,
USA

2Theodor-Boveri-Biozentrum der Universitat, Zoologie 2, Am Hubland, D-97074 Wurzburg,
Germany

290

J. Res. Lepid.

Fig. 1. Adults of the Glaucopsyche group of related species. Column 1, male
upperside; 2, female upperside; 3, underside. Row 1 , Glaucopsyche alexis
, Spain; Maculinea arion, France; lolana iolas, Spain; Caerulea coelestis,
Tibet; C. coeligena, Tibet; Phengaris atroguttata, Taiwan (female, India).

In the course of morphological studies on the Glaucopsyche section
sensu Eliot (1973) (this grouping is also referred to as Scolitantidini,
Mattoni 1977), the senior author discovered functional foretarsi in both
species of the taxon Caerulea Forster 1938. The Holarctic Glaucopsyche
section comprises about 15 recognized nominal genera with some 50

30(3-4):289-296, 1991

291

species and several hundred so-called subspecies (Mattoni 1977, Bridges
1988). The phylogenetic relationships between the “genera” are not yet
established, nor have all species or subspecies been revised. Caerulea
contains only two species: C. coeligena (Oberthiir 1876) with ssp. pratti
(Hemming 1931), and C. coelestis (Alpheraky 1897) with ssp. dubernardi
(Hemming 1931). Both are poorly known from a limited geographical
area in southern China: Szechuan, Hupeh, and the eastern border of
Tibet. Sometimes considered to constitute a single species only (Bridges
1988), they appear to be partly sympatric supporting their status as two
valid species with two subspecies each (Hemming 1931, Mattoni unpubl. ).
Their life-histories and ecology are totally unknown, except that they
occur at relatively high altitudes and are early spring fliers (implying a
probable pupal hibernation, as usual in the Glaucopsyche section except
in Maculinea).

The phylogenetic relationships of Caerulea are uncertain, although the
male genitalia are quite similar to Maculinea. F orster ( 1938), in erecting
Caerulea , classified it as a subgenus of Glaucopsyche together with
Maculinea. The legs of Caerulea are quite dissimilar to other members of
the Glaucopsyche complex. In addition to the possession of distinct male
tarsomeres, femur and tibia are robust in Caerulea and the fore tibia has
a well developed distal process (much smaller in Glaucopsyche s. str.,
absent in Maculinea). Furthermore, Caerulea males do not possess
androconia. Figure 1 shows adult specimens, both sexes and undersides,
of representatives of the presumptive closely related genera or subgen-
era involved in this discussion. Figure 2 illustrates the male genetalia of
the same set of species. These figures present an overview of the very rich
set of character states available for study in this group of insects.

Figs. 3-6 illustrate both male and female forelegs, with a detailed
ventral view of the tarsomeres in both sexes. The chaetotaxy is different
between the two sexes, particularly conspicuous is the greater spine
number (A-type trichoid sensilla) of the female. The limited sample
comprising two different subspecies of C. coeligena precludes a definite
statement about sexual dimorphism in Caerulea leg chaetotaxy, but such
dimorphism is quite usual among the Lycaenidae (Robbins 1988b,
Mattoni unpubl.).

The ankylosed condition of the foretarsus, obviously an apomorphic
character state relative to the Papilionoidea groundplan, is male-limited
in lycaenid butterflies. In any population, therefore, genetic competence
exists for developing a functional tarsus, since such competence is
present, and the character invariably expressed, in the females. Second-
ary reversion of the functional male foretarsus would consequently
represent a form of “decoupling” as discussed by Vane-Wright (1979).

We suggest that the ankylosed condition originated by a unique
“macromutational” event. This event occurred in the ancestral stem
lineage giving rise to the Lycaenidae, Riodinidae and Nymphalidae
(these 3 families together form a well-defined monophyletic unit:

292

J. Res. Lepid.

Fig. 2. Male genetalia, lateral view and valve. Species shown in figure 1 and discussed ii

30(3-4):289-296, 1991

293

Kristensen 1976, Scott & Wright 1990). In the Lycaenidae, this charac-
ter became integrated into a gene complex whose expression is male-
limited. Subsequently, several independent reversions to the functional
tarsus occurred in a few lycaenid lineages, and these reversions became
fixed genetically. Two possible genetic mechanisms can be proposed: 1)
a female-linked genetic factor on the Y chromosome normally controls
the expression of a segmented foretarsus, and this factor could enter the
male genome via crossing over; or 2) a male-limited genetic factor
normally blocks the expression of the segmented foretarsal condition,
and this factor is affected by direct mutation (see also Eliot 1973).
Definitive analysis would require discovery and genetic study of popu-
lations dimorphic for the character, if such populations should exist.

An alternative hypothesis would postulate that functional forelegs
were always restricted to ancestral forms, and that the ankylosed
condition, once fixed, never reverted. Clearly, the bulk of other morpho-
logical as well as zoogeographical evidence strongly contradicts this
assumption. If true, Caerulea together with all other lycaenids sharing
a functional male foretarsus were either to represent the closest living
relatives of the ancestor of all the remaining Lycaenidae. Or, alterna-
tively, one would have to postulate repeated independent evolution of
ankylosed male foretarsi among the Lycaenidae. Clearly, the most
parsimonious explanation of the pattern observed is that the recurrence
of functional male foretarsi in a few, and systematically isolated,
lycaenid lineages is the result of convergent evolution.

What phylogenetic inferences, if any, can be based on the recurrence
of a presumably ancestral character state? In the case of Caerulea , the
peculiar leg morphology probably represents one autapomorphy of the
group, supporting the monophyly of the (sub-)genus. However, leg
morphology provides us with no information relevant to finding the
sister-taxon of Caerulea. Hence, the decision whether, in a strictly
phylogenetic sense, Caerulea is a monophyletic subgenus of Glaucopsyche,
or the sister-genus of Glaucopsyche , or even the sister- taxon of a larger
assemblage of members of that section, remains open.

In the case of the Liphyrini, the common occurrence of functional male
foretarsi in all included genera provides support for the monophyly of
this grouping. By implication Lachnocnema and Thestor must be re-
moved from the Miletini sensu Scott & Wright (1990). This alteration in
tribal arrangement, also suggested by Eliot (pers. comm.), well reflects
the zoogeography of the Miletinae. Liphyrini in the new sense adopted
here are then wholly Ethiopian with the single exception of Liphyra
(probably a secondary invader of the Oriental region), whereas vice
versa Miletini become Oriental with one African ( Megalopalpus ) and
one Nearctic ( Feniseca ) extension.

Four of the Thecliti genera with functional male foretarsi (. Artopoetes ,
Coreana, Ussuriana, Japonica) belong to a common clade with larval
hostplants in the Oleaceae. Thus, loss of the ankylosed male foretarsus

294

J. Res. Lepid.

Figs. 3-6Caerulea forelegs.

Fig. 3. Femur, tibia, and tarsus. C. coelestis coelestis, male, China, Tatsienlou, Tibet,
no date, ex coll. Oberthur. RHTM #128.

Fig. 4. Femur, tibia, and tarsus. C. coeligena pratti, female, China, Ichang, Flupeh,
no date. RFITM #129.

Fig. 5. Ventral view of tarsomeres. C. coeligena coeligena, male, China, Tibet, 1 887,
leg. R. Dejean. RFITM #127.

Fig. 6. Ventral view of trasomeres. C. coeligena pratti, female, same specimen as in
Fig. 2.

30(3-4):289-296, 1991

295

might be a synapomorphic trait of this assemblage. However, the Iberian
Laeosopis roboris clearly belongs to this same group (based on morpho-
logical grounds as well as hostplant association), but retains the ankylosed
condition (D. Kovac, pers. comm.). A definitive treatment requires a
phylogenetic analysis of the Thecliti as a whole.

In all other lycaenids with functional male foretarsi, this trait can at
best be used as additional autapomorphy defining small genera (like
Titea or the monotypic Sukidion ), or species-groups. Nevertheless, the
recurrence of presumed ancestral character states is in itself an interest-
ing phenomenon, and it remains to be seen whether further cases can be
discovered in the course of morphological studies of hitherto insuffi-
ciently known Lycaenidae subgroups.

Acknowledgements. We wish to express our thanks to R. I Vane-Wright
and the trustees of the Natural History Museum, London (formerly
BMNH), for their generosity in providing some specimens for study.
These specimens and preparations will be deposited in the BMNH. The
remaining specimens will be deposited at the MCZ, Harvard University.
R. I. Vane-Wright provided constructive and extremely helpful criticism
on an early manuscript version. Dr. D. Kovac (Senckenberg-Museum,
Frankfurt) kindly provided information on leg morphology of L. roboris.

Literature Cited

Bridges, C. A., 1988. Catalogue of Lycaenidae and Riodinidae (Lepidoptera:

Rhopalocera). Urbana/Illinois. vii + 816 pp.

Eliot, J. N., 1973. The higher classification of the Lycaenidae (Lepidoptera): a
tentative arrangement. Bull. Brit. Mus. Nat. Hist. (Entomol.) 28: 371-505,
6 pis.

Fiedler, K., 1991. Systematic, evolutionary, and ecological implications of
myrmecophily within the Lycaenidae (Insecta: Lepidoptera: Papilionoidea).
Bonner Zool. Monogr. 31: 1-210.

Forster, W., 1938. Das System der palaarktischen Polyommatini (Lep.

Lycaenidae). Mitt. Munch. Entomol. Ges. 28: 97-118.

Harvey , D. J., 1987. The higher classification of the Riodinidae (Lepidoptera).

Ph.D. thesis, Univ. Texas, Austin, vii + 215 pp.

Harvey, D. J., 1991. Higher classification of the Nymphalidae. Pp. 255-273 in:
H. F. Nijhout, The development and evolution of butterfly wing patterns.
Smithsonian Inst. Press, Washington/London, xvi + 297 pp.

Hemming, A. F., 1931. Revision of the genus Iolana, Bethune-Baker. Trans.
Entomol. Soc. London 79 (2): 323-333.

Kristensen, N. P., 1976. Remarks on the family-level phylogeny of butterflies
(Insecta, Lepidoptera, Rhopalocera). Z. Zool. Syst. Evolutionsforsch. 14: 25-
33.

Mattoni, R. H. T., 1977. The Solitantidini I. Two new genera and a generic
rearrangement (Lycaenidae). J. Res. Lepid. 16: 223-242.

Mattoni, R. H. T., 1981. The Scolitantidini II. The world’s smallest butterfly?
Notes on Turanana, and a new genus and species from Afghanistan
(Lycaenidae). J. Res. Lepid. 18: 256-264.

296

J. Res. Lepid.

Robbins R. K., 1988a. Comparative morphology of the butterfly foreleg coxa and
trochanter (Lepidoptera) and its systematic implications. Proc. Entomol.
Soc. Wash. 90: 133-154.

Robbins, R. K., 1988b. Male foretarsal variation in Lycaenidae and Riodinidae,
and the systematic placement of Styx infernalis (Lepidoptera). Proc. Entomol.
Soc. Wash. 90: 356-368.

Robbins, R. K., 1989. Systematic implications of butterfly leg structures that
clean the antennae. Psyche 96: 209-222.

Scott, J. A. & Wright, D. M., 1990. Butterfly phylogeny and fossils. Pp. 152-208
in: O. Kudrna (ed.), Butterflies of Europe. Vol. 2: Introduction to
lepidopterology. Wiesbaden, Aula-Verlag. 557 pp.

Vane-Wright, R. I., 1979. Towards a theory of the evolution of butterfly colour
patterns under directional and disruptive selection. Biol. J. Linn. Soc. 11:
141-152.

Journal of Research on the Lepidoptera

30(3-4):297-301, 1991

Book Reviews

THE COMMON NAMES OF NORTH AMERICAN BUTTERFLIES, edited by
Jacqueline Y. Miller (foreward by Paul A. Opler). 1992. Smithsonian Institution.
177 pp. Soft cover. Order from: Smithsonian Institution Press, Dept. 90, Blue
Ridge Summit, PA 17294, $14.95 + $2.25 mailing.

A list of common names? The first question occurring to the biologist who
happens to work on Rhopalocerans or to the post-neo lepidopterist is “Why not
use the scientific names? I have learned them and so can anyone else.” The
arguments against producing an “official” standardized list of common names
were stated by Murphy & Ehrlich (J. Res. Lepid. 22:154-56). The (effective)
rebuttal is Pyle’s (ibid. 23:89-93). What properties should names have and can the
Latinized (scientific) names suffice alone?

The properties that a name, Latinized or common, should have are these: It
must be unique, shared by no other organism. It should be descriptive of the
organism. It should show relatedness to other organisms. Most important, it
should not change. Changing names defeats the information retrieval/storage
value of the name. Witness the tendency of authors to omit names they don’t favor
from their indexes. Perhaps Occidryas shouldn’t appear in the index here, but I
wouldn’t feel that way if it were Euphydryas that was missing. I’d be ranting
about suppression and attacks on the information retrieval system. Names
should provide easy access to this information storage/retrieval system. Anyone
should feel free to pick them up and use them.

On this last point, the Latinized name does not work. Most people, including
our media link to the public, tune out the Latinized names. If we want the
attention of these people to be on butterfly biology and on butterfly conservation
issues, we have to supply common names. Government agencies also tend to
require and to use in official documents, the common name before the scientific
name. This is at least partly due to the public’s use of common names in prodding
government agencies into action. Today in my consulting work (land use permit-
ting, EIRs), I was asked about the “checkered moth”, suggesting either a need for
this book or for a list for moths. Common names help link people to the Latinized
names and thus to the scientific literature.

A standardized list of common names, once through a couple of revisions, would
tend to change less than the Latinized names. Evolutionary biologists change the
Latinized names whenever their momentary concepts of phyletic relationships
change, but they leave the common names alone. In the case of birds, the common
names are in fact more useful than the scientific names, since the common names
now have greater currency than the frequently changed Latinized names. The
common names of birds are therefore of more use than the “scientific” Latinized
names to both scientists and laymen.

The point that common names carry little information is true, but hardly
differentiates them from Latinized names. Every patronym is an assault on the
information content of a name, Smith’s Blue as much as smithi. On this point I
stand with Darwin and with Dimock (J. Res. Lepid. 23:94-101). Common names
can be chosen to be descriptive and non-patronymic, even when the Latinized
name is a patronym. Euphydryas editha luestherae can be called the Chaparral
Checkerspot or the Lousewort Checkerspot, for instance.

298

J. Res. Lepid.

Common names often fail to show relatedness to other species, but that is not
a necessary component of being common names. The common name, Checkerspot,
links seven genera of the eleven in the subfamily Melitaeinae. The Latinized
names link only the three -dryas species by similarity of name.

Common names, then, can fill the role names of organisms should fill, in some
ways better than scientific names. For communication with the public, common
names clearly serve better than do scientific names. For this purpose, common
names should be anglophone and friendly to the ear of English-speaking persons.
English names that are short and descriptive of widely understood traits will be
most effective. Names that have emotional appeal may be superior to descriptive
names for conservation purposes.

The twenty committee members who labored to generate this highly valuable
reference work have my sincere gratitude. Authors seeking appropriate common
names can now find them. Anyone with information or interest in a butterfly that
he or she knows only by a common name can now look it up and see what the
Latinized name might be. The book offers a list of common names collected from
all major sources, with a recommended name in boldface where a choice was not
too difficult. In many cases, several common names are listed and none is
designated as recommended. Publication of this work will have the instant effect
of producing new sources of previously published common names and advice on
which of several possibilities to adopt into the standard list. I would mention a
real need for separate names for the species and for the nominate subspecies. A
second edition should appear, in my opinion, in well under ten years.

Raymond R. White, 788 Mayview Avenue, Palo Alto, CA 94303.

C UIA DE LAS MARIPOSAS DIURNAS DE GALICIA. By Eliseo H. Fernandez
Vidal. 1992. 219 pp. Diputacion Provincial, La Coruna. Hardbound, 20 x 38
c n. ISBN: 84-86040-57-4. Available from the author at reduced price to lepidop-
terists for US $20.00 plus postage ($8.00 sea mail; $15.00 air mail). Address:
Flaza de las Angustias 4-2, 15403 Ferrol, La Coruna, Spain.

Words like thoroughness, quality, and excellence apply to this guide to the
butterflies of Galicia. Galicia is the region of northwestern Spain, north of
Fortugal, and is comprised of four provinces. The author, a resident lepidopterist,
has produced an outstanding contribution on Palearctic butterflies.

The book is hardbound and copiously illustracted with color photographs of
butterflies. Most species are shown as pinned specimens in color, as well as living
adults and some larvae in photographs showing them on flowers or leaves. At the
location of the text of each species, the butterflies are depicted in life-size black
and white line drawings showing the main wing pattern of upperside and
underside of both male and female. All of this work produced by the author
reveals him to be an accomplished photographer and artist as well as writer.
Introductory chapters cover morphology, nomenclature, and ecology. A clear
policy on conservation and responsible collecting is developed. Under the text of
each species we find hostplants, habitats, key characters, distributions, and
infraspecific names cited. The only thing I dislike about the book is a trivial
complaint: in many figure captions and tables, generic names are abbreviated
and require flipping to other pages to find the full name.

The text is in Spanish, but thanks to the usage of Latin names, abundant
figures, and numerous cognates shared between Spanish and English it should
be easy to extract basic information by the all-too-many American readers who

30(3-4):297-301, 1991

299

are intimidated by foreign languages. An alert reader may observe what appear
to be misspellings of some localities, for example, Sierra del Caurel (Spanish) and
Serra do Caurel (Gallego). Gallego is the regional language of Galicia, akin to
Portuguese, and it makes sense that the author would cite localities in the
regional tongue as the signage at collecting sites is given as such.

The most noteworthy aspect of this book is the biogeography of the butterflies
of Galicia. In a concluding chapter, the title of which translates as “Origin,
establishment, and faunal composition of the Rhopalocera of Galicia,” F ernandez
Vidal demonstrates his scientific competence and his grasp of his regional fauna.
In 14 pages of text, instructively illustrated with six maps and two tables, the
geological history of the Iberian Peninsula is described in great detail. All 155
species of butterflies are cited in a table, giving their geographical origins,
postglacial dispersal routes, and time of establishment. For example, Nymphalis
antiopa is stated to have a Siberian origin and reached Galicia in postglacial
times by the Cantabrian route.

Many lepidopterists, particularly Europeans, will want to buy this beautiful
book. It would be very fortunate if every young, budding lepidopterist had a book
that gives such a basic yet detailed treatment to the butterflies of the area where
he/she is living. Fernandez Vidal offers a very good model for other regional
butterfly books. The price is very low for such a fine book.

Richard S. Piegler, Department of Zoology, Denver Museum of Natural
History, 2001 Colorado Blvd Denver, CO 80205

THE THUNDER TREE: LESSONS FROM AN URBAN WILDLAND. 1993.
Robert M. Pyle. Houghton Mifflin. New York. 220 pages. $19.95.

A boy and an irrigation ditch through Denver, Colo., and its suburbs is an
unlikely lifelong friendship.

But here it is in a beautiful reminiscence of how a boy became a naturalist in
love with his High Line Canal and its butterflies, its toads and animals. And here
is a deeper story of human civilization/suburbanization moving to the edge of the
canal and gradually but inexorably changing nature.

Robert Michael Pyle, founder of the Xerces Society, biologist and conservation-
ist was the boy who found a sense of place and self over a 40-year association with
what he called his “ditch.” And found a “Thunder Tree,” a hollowed snag along the
ditch which saved his life, providing shelter during a violent hailstorm and
forever imprinted him with the “nature” provided by the canal and its journey
through the plains surrounding Denver.

Boyhood adventures in the ditch there were, but Pyle, the boy, soon became
Pyle the butterfly collector and later the college student with a PhD in ecology
from Yale. He authored five books, including Wintergreen , which won the John
Burroughs Medal for the best natural history book of 1987, and the acclaimed
Handbook for Butterfly Watchers. He lives now in Gray’s river, Washington, but
across 40 years and many return trips his beloved High Line Canal became his
obsession and metaphor for man’s use of nature and misuse of its glories. For, as
Denver spread itself to the suburbs, the High Line Canal was very nearly
swallowed up by housing and the infrastructure which supports the lifestyle of
city folks.

Woven through this tale of city and country is a thoughtful, patient but
sometimes sorrowing coming-to-grips of the inevitable conflict between growth
and progress and a fragile environment of plains country with its grasses, birds

300

J. Res. Lepid.

and butterflies and Pyle’s struggle to preserve its natural history in the midst of
the classic western struggle over water rights, water development, and wild
species preservation and to remember the Arapaho Indians.

“The citizens of Arapahoe County will never be Arapaho” he writes. “We can’t
get back to Black Kettle’s day. But if we were to watch the wood nymphs, ponder
the pronghorns, and seek to adapt accordingly, perhaps we could still share in the
grace of the grasses. We do have our limits after all. It is the gamble of our growth
that says we have no further need for the Grass Dance.”

In this 40-year journey along the High Line Canal, Pyle, the ecologist and
naturalist, leads us through the dilemma of progress and biodiversity, through
grasslands and water creatures and their relationship to the land and in turn its
relationship to water law and water politics.

For the moment, the High Line Canal has been “saved” for future generations,
but Pyle sees its future as fragile and insecure, always threatened by change from
development and progress. He concludes his splendid collection of essays and
reminiscence with the wry observation:

“But the true significance of the High Line Canal is this: if nature can make it
here, it can make it anywhere.”

A profound challenge for all of us, who, like John Hay in The Bird of Light said:

“In a world made smaller by human dominance, the butterflies of spring keep
me in touch with the planet’s abiding distances.”

Lu Haas. 15515 Sunset Blvd, Apt. 219, Los Angeles, CA 90272.

SYSTEMATICS OF THE CHRYSOXENA GROUP OF GENERA (LEPI-
DOPTERA: TORTRICIDAE: EULIINI. 1991. John W. Brown and Jerry A.
Powell. U. Calif. Publications in Entomology, Vol. 111. 95 pages with 143 figures,
2 tables. $15.00

University of California publications in entomology are uniformly excellent as
a consequence of the outstanding board of editors and the peer review process.
The series also must be congratulated for the act of making its works available
at a reasonable price. This monograph is no exception.

The Tortricidae contains about 5,000 named species, which probably represent
half the world fauna when collection and studies are completed. The authors
follow the conservative, and I believe most rational, classification of Horak and
Brown for the (Super) family, with the Euliini recognized as having 47 genera.
The authors treat 54 species in six genera. Their analysis and phylogenetic
hypothesis for the tribe is based upon 35 characters, while for the deduced
phylogeny of the Chrysoxena group they employed 23 characters. The group is
distributed from Wyoming to southern Brasil.

The general interest section of the paper is in the first 20 pages, the systematic
treatment is given in 56 pages, with six pages presenting all known biological
information (not much), and winding up with Literature cited and the bulk of the
illustrations. Almost all species are illustrated with reasonable quality photo-
graphs of wing facies and most are represented by male and female genitalic
preparations.

In a lecture Dr. Brown presented on his work with the group he injected his wit
into the subject by showing a slide of fettucini while phonetically reviewing

30(3-4):297-301, 1991

301

relationships of the various tortricine tribes. His leavening is missing from the
paper.

We unhesitatingly point out the necessity of supporting this level of work.
Rudolf H. T. Mattoni, 9620 Heather Road, Beverly Hills, CA 90210, USA.

MOTHS OF AUSTRALIA. I. F. B. Common. 1990. E. J. Brill. Netherlands. 535
pages with 32 color plates. ISBN 90 04 09227 7. Price c. US$ 175.

Bravo to the author and press for a must have book (if you can afford it). In
preparing a regional field guide to moths and I had been searching for a thorough
well written one shot reference for a summary biology of moths. Common
answered these prayers. His introductory section of 82 pages starts with general
life history and structure including an outstanding distillation of illustrations
from line drawings by the author to SEM. This is followed by 22 pages of biology
with subsections on such topics as larval gregariousness, sensory perception,
pupal movements, thermoregulation, etc. Every thing you wanted to know about
moths but were afraid to ask. Then brief pieces on population control, economic
significance, and evolution and geographical distribution. A summary family
classification-checklist winds up this part.

Part two is a review of the Australian moth fauna organized by superfamily,
but covering groupings to subfamily. There are appendices for collecting and
study methods and a foodplant list.

Both text and illustrations are extremely efficient. Over 1000 species are given
as either black and white figures or on color plates. The former are of the highest
quality as most photographic material was prepared by the acclaimed Ederic
Slater. The color plates cover at least 400 species as adults and another 100 in
early stages. There are no balloon words or gratuitous illustrations, no sentences
like “these brilliant denizens dancing in the elfin forests against caerulian skies.”
I cannot overemphasize the scholarly quality of the book.

Although the work is clearly regional, its background material is universal. I
might have a bone it pick with Common’s adoption of some of the opinions of
Minet with regard to inflated higher categories, but such subjective criticisms are
minor in context of its overall value.

If you cannot afford the book, by all means urge your local and University
library to buy it. The book will be a valuable reference for decades to come. Thank
you Dr. Common.

Rudolf H. T. Mattoni, 9620 Heather Road, Beverly Hills, CA 90210, USA.

Journal of Research on the Lepidoptera

30(3-4):302-304

Note

The Role of Vernal Pools in the 1992 Mass Dispersal of Vanessa
cardui (Nymphalidae) with New Larval Hostplant Records

Vernal pools are small depressions, underlain by hardpan or dense clay, in the
grasslands of the Cental Valley. Winter rainfall perches above the impervious
subsoil and slowly evaporates as a result of increasing temperatures in the
spring. During mid-to-late spring, vernal pools are islands of fresh vegetation in
an otherwise dry landscape. The flowering forbs within the pools are predomi-
nantly native with a high proportion of endemics. Many of these plants require
a specific inundation period. This results in the vernal pools containing concen-
tric rings of different plants, each with their own uniquely colored blooms. The
historic extent of vernal pools has been severely diminished by agricultural
conversion and urban expansion. For all of these reasons, vernal pools have long
attracted botanists.

In western Tehama County, several areas of vernal pool formations remain.
During a visit to these pools on April 14- 16th, I observed the mass Vanessa cardui
migration through the area. The painted ladies were flying in great numbers in
a NNW direction with individuals often stopping to nectar on Navarretia
leucocephala Bentham (Polemoniaceae). In some small (<30 sq. meters) pools, as
many as 50 individuals were nectaring on this low growing, white-flowered
vernal pool endemic. The Vanessa cardui were not observed to nectar on other
species even though several were blooming.

During a subsequent visit on April 27-29th, the numbers of Vanessa cardui
migrating through the area were significantly lower. Upon close inspection of the
vernal pool plants, I observed many early instar larval nests on Psilocarphus
brevissimus Nuttall (Asteraceae: Inuleae), another vernal pool endemic. Ran-
dom transect placements in several of the small, shallow vernal pools yeilded an
average of >100 larval nests per square meter. Larger pools, which had been
partially inundated during the height of the migration, had many larval nests on
the margins and few in the center. Extremely deep pools, which had been
completely inundated during the migration had only an occasional larval nest.

Many third and fourth instar larvae were also observed during the late-April
field visit. These larvae had left their nests and were apparently in search of
fresher food. Most were eating adjacent Psilocarphus brevissimus. However,
some larvae were observed on Psilocarphus tenellus Nuttall, Micropus californicus
Fischer & Meyer and Filago gallica Linnaeus (Asteraceae: Inuleae), plus
Achyrachaena mollis Schauer (Asteraceae: Madiinae) and Plagiobothrys stipitatus
var. micranthus (Piper) Johnston (Boraginaceae). Although Eryngium (Apiaceae)
is cited as a larval host plant in The Butterflies of North America (Scott, 1986),
none were observed on the Eryngium vaseyi Coulter & Rose occurring in the
vernal pools. Several large larvae were collected from the Tehama County vernal
pool site. These were successfully reared on Psilocarphus brevissimus with
pupation beginning on May 2nd and emergence beginning on May 9th.

A final field visit was conducted on May 5-6th and many newly emerged
Vanessa cardui were observed migrating. However, the preceeding week had
been 15-20° F above local average temperature and most of the remaining early

30(3-4):302-304

303

Photo 1 : Typical vernal pool from western Tehama County.

Photo 2: Psilocarphus brevissimus, a vernal pool endemic plant and new larval
hostplant record for Vanessa cardui.

304

J. Res. Lepid.

instar larvae were found dead in their nests on the dessicated hostplant. Larger
larvae were observed moving from plant to plant.

Following the discovery of Vanessa cardui larvae on vernal pool endemic
plants, a quick reconnaissance of vernal pools in Glenn, Colusa, Sacramento and
Solano counties was made. On May 5-7th, several vernal pool formations were
visited and Vanessa cardui larvae or larval nests were observed at each site on
Psilocarphus brevissimus. At two sites in Solano County, later instar larvae were
also observed on Evax caulescens (Bentham) Gray (Asteraceae: Inuleae).

Of considerable significance is the lack of adult nectar sources at some of the
vernal pool sites. The terrace soil formation vernal pools in Tehama County are
dominated by Navarretia leucocephala and Psilocarphus brevissimus which
places an adult nectar source and the hostplant in close proximity. However, the
basin soil formation vernal pools in Solano County contain very little or no
Navarretia leucocephala and relatively small amounts of Psilocarphus brevissimus.
Nevertheless, these pools contained a similar ratio of larvae to hostplant. From
these observations we may deduce that use of Psilocarphus brevissimus as a
larval hostplant was not purely adventitious.

Carol W. Witham, 1028 Cypress Lane, Davis, California 95616.

Host specialization of satyrine butterflies, and their responses to 248

habitat fragmentation in Trinidad
M.C. Singer and P.R. Ehrlich

Celastrina nigra and its synonym C. ebenina (Lepidoptera: Lycaenidae) 257
James A. Scott and David M. Wright

Foam barriers, a new defense against ants for milkweed butterfly 261

caterpillars (Nymphalidae: Danainae)

P.J. DeVries

Temporal and spatial overlap in the mate-locating behavior of two 267

species of Junonia (Nymphalidae)

Ronald L. Rutowski

Cephalic Sclerites and Chaetotaxy of a Hairy Caterpillar, Lymantria 272
marginata Wlk. (Lepidoptera: Lymantriidae)

Jasvir Singh

Distribution and Flight Behaviour of the Junglequeen Butterfly, 279

Stichophthalma louisa (Lepidoptera: Nymphalidae), in an
Indochinese Montane Rainforest

Vojtech Novotny, Martin Tonner and Karel Spitzer

On the functional foreleg tarsus in Caerulea males (Lepidoptera: 289

Lycaenidae: Polyommatini)

Rudolf H. T. Mattoni and Konrad Fiedler

Book Reviews 297

Note 302

THE JOURNAL OF RESEARCH
ON THE LEPIDOPTERA

Volume 30 Number 3-4 Fall 1991

IN THIS ISSUE

Date of Publication: December 1, 1993

The Butterflies of Mauritius 145

P.M.H. Davis and M.J.C. Barnes

Convergent Evolution in Western North American and Patagonian 162

Skippers (Hesperiidae)

Arthur M. Shapiro

Multivariate and Phylogenetic Analyses of Larval and Adult 175

Characters of the Editha Complex of the Genus Lycaena
(Lepidoptera: Lycaenidae)

Gordon F. Pratt, David M. Wright,
and Gregory R. Ballmer

A morphological search for the sound mechanism of Hamadryas 196

butterflies (Lepidoptera: Nymphalidae)

Julian Monge-Najera and Francisco Hernandez

Extirpation and Recolonization of the Buckeye, Junonia coenia 209

(Nymphalidae) Following the Northern California Freeze of
December, 1990

Arthur M. Shapiro

Notes on the Immature Biology of Two Riodinine Butterflies: 221

Metacharis ptolomaeus and Napaea nepos orpheus
(Lycaenidae)

Curtis J. Callaghan

The Effects of Temperature and Daylength on the Rosa Polyphenism 225
in the Buckeye Butterfly, Precis coenia (Lepidoptera:

Nymphalidae)

Kelly C. Smith

Three New Taxa of Calephelis from Costa Rica (Lycaenidae: Riodininae) 237
George T. Austin

A New Species of Calephelis from Guatemala (Lycaenidae: Riodininae) 245
George T. Austin

Cover Illustration: Lateral view of thoraxes of six species of Hamadryas .
Line drawing by Monge-Najera and Hernandez.

45 1G1 SI XL
04/94 198115

03

io\

S3

d)

>

S3

[KflL

7

V?

m

CO

I ^ s ! |

HSONIAN INSTITUTION NOIIDIIISNI NVINOSH1IWS S3iavaaiT LIBRARIES S

2 \ ^ 5 CO

- «J m

Q X^OS^X 0,1 N^VASVV-^- Q " w ^ ^VASn>^ Q

2 3 Z -J 2 _

0SHJL1WS S3ldVden LIBRARIES SMITHSONIAN INSTITUTION NOIlfUllSNI N

| I^| | »

HSONIAN INSTITUTION^ NOlinillSNrNVINOSHimS^SB I HVM 8 11 L I B RAR I ES^S
^ <o 5 . . g 5 g

•?/ S3

X
CO

o
z

^ ^ s 2 ^ >

IOSHilKIS'"S3 I 8V8 a Sl^ LI B RAR I ES^SMITHSONIAN_SNSTITUTION^NOIiniIISNI_N

__ Y_‘. /^stoSA — ,• ^ •« *2 — /^SsSDA w

< (*§. Ji) = (ie. .Ji) £;!CJr = Itk.Ji) 2 fer

£>/ m OQ

q n^v.dc/ _ q

HSONIAN^ INSTITUTION NOliniliSNI^NVINOSHIlWS^SS I BVH 9 n“Yl BRAR I ES^S

2 r* „ 2 r~ _ z

2 xS£^o£x m 2 n x^oa^x o

h *1 ^

03
S3
>

o/ S3
^^IKo7 —

tVASV^X m '■ \>Sg55> 2 \^VASI^X' Fj

JOSHilWS^SB I HV8 8 I1~L I B RAR I ES SMITHSONIAN^INSTITUTION^NOimillSNJ t

CO . Z *«, CO __ 2 « CO 2

1 I XC^I^X < A S <

f) o i 2

^los^ix >* 2 'w1 > XiTAS^X -e-

2 CO 2

«\ ~

HSONIAN ^INSTITUTION N0linillSNi_NVIN0SHilWS^S3 I d VH 8 ll_L! BRAR I ES S

fc, S w w u ^ ^oa^x ^

_i — /v^WfePA h -I /^sw^3\ —

c < lsE S "

OSHimS S3I8V89H LIBRARIES SMITHSONIAN INSTITUTION NOIlfUUSNI N
2 r~ x r* ___ 2 “*

03 2 OJ \ CO

m x^vosv1^ ^ m ' ,.S^V 2 \^yAsn>^ m

S^sa I 8 V8 8 11 ~LI B R AR I ES SMITHSOMAN^INSTITUTION^NOlinillSNI NVINOSHJ
in z v> „ z y in z

K\UX

z
o

CD

x
b

2 - .v^ > i * ^ > ■' 2

'(“institution wNouniusNi NviNosHiiws^sa ! a va a nzu b rar i ES^SMiTHsor
" 5 \ £ 5 "

CO

O

Z

S

5 2

2 «J

5 SBIdVdan LIBRARIES SMITHSONIAN INSTITUTION NOIJLfUUSNI NVIN0SH1

Z r* — Z ~~ “ “

__ _ .. Pi

^“INSTITUTION^ NOIinillSNI~~NVINOSHlllNS S3 I UVU 8 H~LI B RAR I ES^SMITHSOf
z %v. CO z g 5 g __

s ^ i 1 | | |

sws3 1 avaa hzlib rar i eswsmithsonian institution noixiuiisni nvinoshj

CO ^ __ CO 5 m 5

^so«^x v^iyArx ^

yj

G£

<

CE

CD

SCOTV

GC

2 . . SSM! c

O DCj^ 2 o

N^INSTITUTION^NOlinillSNI^NVINOSHJLIlNS^Sl I 8VH 8 ll^LI B RAR I ES^SMITHSOI
z r- * z r“ z

P z © y<5?nrt//X z ©

m x^yosv^x 2 m *- )^y ^ \wy m

S* S3I8V88 n~LI BRAR I ES SMITHSONIAN~INSTITUTION%OliniLLSNI NVINOSH.
co . z , m z * co z

z

CO

c

H

O

5 S3 1 8 va an libraries Smithsonian institution NoiiniusNi nvinoshj

z r- z ~ “ “

O z X^visJ2^7X ©