Survival of first instar larvae of Danaus plexippus (Lepidoptera: Danainae) in relation to cardiac glycoside and latex content of Asclepias humistrata (Asclepiadaceae

Chemoecology 3 (1992) 81-93 81

Survival of first instar larvae of Danaus plexippus (Lepidoptera: Danainae) in relation to cardiac glycoside and latex content of Asclepias humistrata (Asclepiadaceae)

Myron P. Zalucki I and Lincoln P. Brower 2 1 Department of Entomology, The University of Queensland, Brisbane, Queensland, 4072, Australia 2 Department of Zoology, University of Florida, Gainesville, Florida 32611, USA

Received September 2, 1991 / Accepted October 30, 1991

Summary

Our paper addresses field survivorship of first instar monarch butterfly larvae (Danaus plexippus L., Lep.: Danainae) in relation to the dual cardenolide and latex chemical defenses of the sand hill milkweed plant, Asclepias humistrata (Asclepiadaceae) growing naturally in north central Florida. Survival of first instar larvae in the field was 11.5°70 in the first experiment (15-20 April 1990), and dropped to 3.4°7o in the second experiment (20-30 April). About 30°7o of the larvae were found glued to the leaf surface by the milkweed latex. Predator exclusion of non-flying inverte- brates by applying "tanglefoot" to the plant sterns suggested that the balance of the mortality was due to volant inverte- brates, or to falling and/or moving off the plants. Regres- sion analyses to isolate some of the other variables affecting survivorship indicated that first instar mortality was correlated with (1) increasing cardiac glycoside concentration of the leaves, (2) increasing age of the plants, and (3) the temporal increase in concentration of cardiae glycosides in the leaves. The study also provided confirmatory data of previous studies that wild monarch females tend to oviposit on A. humistrata plants containing intermediate concentrations of cardiac glycosides. Cardiac glycoside concentration in the leaves was not correlated with that in the latex. The concentration of cardenolide in the latex is extremely high, constituting an average of 1.2 and 9.5°70 of the mass of the wet and dry latex, respectively. The data suggest that an increase

in water content of the latex is compensated for by an influx of cardenolide with the result that the cardenolide concentration remains constant in the latex systems of plants that are growing naturally. We also observed first instar larvae taking their first bite of milkweed leaves in the field. In addition to confirming other workers findings that monarch larvae possess elaborate "sabotaging" behaviour of the milkweed's latex system, we discovered that several larvae on their first bite involuntarily imbided a small globule of latex and in- stantly became cataleptic. This catalepsis, lasting up to 10 min, may have been in response to the high concentration of cardenolide present in the latex of A. humistrata, more than 10 times that in the leaves. The results of the present study suggest that more attention should be directed to plant chemical defenses upon initial attack by first instar insect larvae, rather than attempting correlations of plant chemistry with older larvae that have already passed the early instar gauntlet. The first bite of neonate insects may be the most critical moment for coping with the chemical defenses of many plants and may play a much more important role in the evolution of insect herbivory than has previously been recognized.

Key words

first instar survival, latex, cardiac glycosides, cardenolides, Lepidoptera, Danainae, Danaus plexippus, Asclepiadaceae, Aselepias humistrata, milkweed

Introduction

The monarch butterfly, Danaus plexippus L., is a specialist on milkweeds. Females oviposit on, and larvae feed exclusively on plants in the Asclepiadaceae, and predominantly on species in the genus Asclepias (Ackery & Vane- Wright 1984). These plants are notable for their toxicity to vertebrates, because they contain a group of plant secondary compounds known as cardiac glycosides (CG) (Brower 1984).

Due in part to nerve insensitivity (Vaughan & Jungreis 1977; Jungreis & Vaughan 1977), an ability of larvae to concentrate cardiac glycosides above their gross concentra-

© Georg Thieme Verlag Stuttgart • New York

tion in the plants (Malcolm & Brower 1989; Nelson 1992), and an ability to conserve the compounds through to the adult stage (probably a form of storage excretion, Brower et al. 1988) monarchs are considered to have evolved the ability to exploit the chemical defences of the plants. This relation of monarchs has attracted substantial attention particularly with respect to adult CG content, aposematic coloration and pre- dation by birds (Brower et al. 1968, 1988; Brower 1969; Brow- er & Fink 1985). In fact the monarch-milkweed story has served as the paradigm for ecological chemistry (Brower & Brower 1964; Brower 1984; Harbone 1988; Malcolm & Za- lucki 1992).

Recently Oyeyele & Zalucki (1990), and Za- lucki et al. (1990) found that females in both Australia and

82 C h e m o e c o l o g y 3 (1992) Za luck i & Brower

North America select plants of intermediate CG levels, reject- ing plants having either very low or very high concentrations of CG. Their data also indicated that early stage survival was negatively correlated with plant CG level. Cohen & Brower (1982) found a similar (but non-significant) relationship. This is the only well documented indication that there may be a physiological cost in dealing with cardenolids in milkweed host plants, although the melanism induced by the principal cardenolide in the sterns of A . curassavica (uzarigenin) is possibly another example (Selber et al. 1980). In contrast to these findings, laboratory studies using fifth instar larvae in bioas- says found no effect of CG levels in diet, or different Ascle- pias species, on growth of IV and V instar larvae and fecundi- ty of subsequent adults (Erickson 1973; Cohen 1983). Varia- tion of adult size in the field in relation to CG concentration has been interpreted as indicating a physiological cost (Brower et al. 1972), but this has been criticised on various grounds (Dixon et al. 1978).

More generally, the survival of phytophagous insects on their host plants has attracted much attention in the ecological literature (Dempster 1983). Although extensive work on host plant resistance in agriculture indicates that various aspects of "plant quality" are related to the mortality of associated herbivores (e.g. Todd et al. 1971; Fisk 1980; Hut- chins et al. 1984), most ecologists consider biotic factors such as parasites and predators (Hassell 1985), and/or the physical environment to play the major role in survival of specialist herbivores. Since females of speciaiist herbivorous insect species supposedly select hosts to maximise immature survival (Singer 1984), and since immatures are considered to be highly adapted to feed and survive on these plants, the secondary chemicals in these "hosts" are considered to be of minimal consequence in mortality of these immatures.

Milkweeds are noted, not only for CG levels, but also for their high latex content (Van Emon & Seiber 1985) which operates as an additional chemical defense (Dus- sourd 1986, 1990; Dussourd & Eisner 1987; Compton 1987, 1989). As these authors showed, milkweed plant specialists (including first instar larvae of the queen butterfly, Danaus gilippus) and some generalists belonging to several insect or- ders exhibit various forms of vein cutting behaviour, that sa- botages the latex defense by preventing its flow into the tissue that the insects then eat with impunity. In seeming conflict with these studies, Rothschild (1977) and Dixon et al. (1978) stated that fifth instar monarch larvae actively imbibe latex from sterns of A. curassavica. This observation is also at odds with a report by Brewer (1977) that 5th instar monarch larvae cut off the flow of latex by chewing partly through the basal petioles of the leaves they subsequently eat. We, too, have repeatedly observed this behaviour, and we suggest that Roth- schild and Dixon et al. "s observation is a misinterpretation of the attempt by the larvae to cut off the flow of latex from the stem into the leaves.

Here we report on field experiments that measured survival of first instar monarch caterpillars on A. humistrata in north central Florida, in relation to cardiac glycoside concentration in the leaves and latex of individual plants. We also report observations of catalepsis and mortality of neonate first instar larvae caused by the latex of this milkweed in the field, and suggest that the high CG of the latex may facilitate this mortality.

Materials and methods

Field site - All field experiments on survival were conducted on Asclepias humistrata at a site near Cross Creek in North Central Florida, described in Zalucki et al. (1990). The site consists of 5.3 ha of disturbed Florida sand- hill habitat that contalns ca 500 A. humistrata plants.

Field method - We conducted two major field trials aimed at assessing first instar monarch survival in relation to plant CG variation.

On April 9, 1990 we selected 44 A. humistrata plants distributed throughout the study site. These were chosen as rar as possible to be of similar phenological stage (age), and size (stem number and length). Plant age categories were: (1) very young buds, (2) buds, (3) buds and flowers, (4) flowers, (5) late flowers, young pods, (6) pods, and (7) old pods (seed dehiscence). Plant age was reassessed on 20 April. The cardiac gtycoside content of each plant was assayed from samples gathered on 9 April 1990, by taking a similar sized leaf from the same position on each plant (the third leaf from the top of similar sized, northerly facing sterns). Each plant was also initially assessed for the degree of herbivore damage, and the numbers of wild monarch eggs and larvae on each plant were recorded. To avoid confusion with our experimental larvae, all these initial natural larvae were removed (and placed on nearby plants). On 15 April twenty-nine of the 44 plants were chosen at random to be experimental plants. To these we added eggs laid by 9 females (see below) that had been held in captivity until about to hatch (usually 10 eggs/plant). These eggs, and any existing or newly laid eggs were carefully moni- tored, and the survival of each newly hatched first instar through to the second instar was recorded for each plant. Plants were inspected on the 16, 17, 19 and 20 of April, when all eggs had either died or hatched and developed to the second instar. We included in this experiment 5 extra plants that had at least 5 naturally lald eggs upon them.

A second experiment was conducted from 20 April using 24 plants chosen at random from the initial 44 (14 plants used in the first experiment were also included in the second). Eggs were added to these as before, and the plants were checked on the 23, 25, 27 and 30th of April. On a further 6 plants we also conducted a small predator exclusion trial (see below).

Egg sources - During the first experiment, wild spring remigrant monarchs were laying eggs in our field. Nine females were collected on April 9 and 10, red a sucrose solution and confined on stems of potted A. curassavica using organza bags. Eggs (ca 300 over 2 days) were cut from leaves using a hole punch (5 mm diameter), placed onto moistened filter paper and kept at 25-28 °C. When the black head cap- sule was evident, eggs were ready to be attached to experimental plants. As too few remigrants or their offspring (the first postmigration adults would not yet have emerged) would be avallable at Cross Creek for the second field experiment, we obtained 9 females on 13 April from the site at Miami used by Malcolm & Brower (1986). Eggs were collected and stored as for the first experiment.

Egg placement, life table construction, and predator exclusion - Eggs that were about to hatch were at-

Survival of first instar larvae of D a n a u s p l e x i p p u s Chemoecology 3 (1992) 83

Fig. 1 Method of attaching experimental monarch eggs to the leaf surface of Asclepias humistrata. A One egg on its leaf disk (right) has been glued to the surface and the position of the second egg has been prepared by pricking the leaf to produce a drop of latex (left) B The second egg on its disk is pressed onto the latex that will glue it tothe Eeaf. (Approx. lX)

tached to the selected A . hum&trata plants in positions similar to naturally laid eggs, i.e. predominantly on the underside of leaves towards the top of a stem (Zalucki & Kitching 1982a; Zalucki et al. 1990). We attached 10 eggs to leaves on 2 sterns of each experimental plant, 5 per stem, and one egg per leaf. (Occasionally we attached 2 eggs to the same leaf, orte on each surface.)

Each egg on its 5 mm leaf disc was attached to the leaf by means of a novel technique (Fig. 1). We first gently pricked the leaf surface of the experimental plant so as to produce a small drop of latex. The disk was then trans- ferred with forceps onto the drop of latex and gently pressed against the leaf surface. As the latex dried, the disks became glued in place and remained attached for several days (even after a rain shower, they could only be removed by scrap- ping).

All eggs hatched within 12 h of attachment, and mosthad hatched within 4 h. Spot checks of eggs placed on the first experimental plants, after the last eggs had been put out, showed that eggs had hatched and neonate larvae had moved onto the A . humis trata leaves and started their characteristic leaf biting and feeding behaviour (see below).

Li f e table construct ion - Life tables were con- structed for each experimental plant. Plants were systematically searched at least every 1-3 days. All eggs and larvae were either accounted for, or recorded as missing and presumed dead. Any dead eggs, larval cadavers, and larvae that had moulted to the second instar were removed so as to avoid dou- ble counting. For each plant we determined the proportion of hatched eggs that survived to the second instar. The first field experiment included naturally laid eggs. These were circled with a ballpoint pen and their rate followed. Monarchs on hatching normally eat their egg shell, before sampling leaf

tissue nearby. First instars rarely wander from the leaf on which they hatched, at least for the first 24 h. If an egg were missing, and there were no signs of nearby leaf feeding, then we assumed the egg had been eaten or dropped to the ground. If there were signs of leaf feeding, but no larva was found, then we assumed the eggs had hatched, but the neonate larva had died, dropped oft, or had been eaten. No adult monarchs were seen laying eggs dnring the second field experiment.

During the second field trial we also conducted a small experiment that effective!y excluded ground dwelling predators. To the base, and up to the first leaf pair of one stem we applied "Tanglefoot" (Great Lakes IPM, Ves- taburg, Minnesota). Any grass touching this stem was removed. The other stem was left nntreated. On 6 plants, not used in the life table trial, we attached 5 egg-bearing leaf discs to leaves on each of 2 sterns and the fate of hatched larvae was followed as described above.

Stat&tical analys& - The proportion of hatched first instar larvae surviving to the second instar for each plant was log transformed before regression analysis against CG concentrations and other plant measures as inde- pendent variables (see below).

Observat ions o f f eed ing behaviour and latex dabbing - While checking the A . hum&trata plants for life table construction we frequently observed and noted neonate larval feeding behaviour. On two days we specifically observed and compared larval behaviour in dealing with the copious latex produced by A . humistrata. On the first occasion we placed first instar larvae of various ages and some third instars on separate plants. These were observed for ca 8 man hours. On the second occasion we observed wild IV and V instar larvae feeding naturally on the A . humis trata plants.


10-

5-

i m

15- A I

Leaves: 9 Apri l

N= 44

0 0 100 200 300 400 500 600 700 800 900 1000

0 t-

"-i O"

LL. !t ~ I~~ Leaves: 2 May

N= 3 0

0 100 200 300 400 500 600 700 800 900 1000

it I ~ c Latex: 2 May

N=51

I I I I . I I 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000

Cardeno l ide concen t ra t i on !ag /O. lg dry we igh t

Fig. 2 Frequency distribution of CG in pg/0.1 g DW in early (A) and ]ate (B) leaf samples, and in late latex (C) samples. Arrows indicate the mean, and bar equals the standard deviation

After finding numerous larvae apparently mired in latex (see Results), we attempted to simulate this effect. Laboratory reared first and third instars were placed on 10 mm leaf disks and then dabbed with fresh latex from either A . curassavica or A . hum&trata. The latex was collected from freshly damaged sterns and leaves with either a small paint brush or a pipette. The larvae, on their leaf discs of A . hum&- trata were kept on moistened filter paper in a Petri dish.

Plan t cardiac glycoside analysis - Plant leaf material collected from the field for CG analysis was placed in individual plastic bags, and stored on ice for transport back to the laboratory. Here leaves were dried for 16 h at 60°C in a forced draft oven, ground to fine powder, weighed and extracted in alcohol for analysis of CG concentration as digi- toxin equivalents, using the standard spectrophotometric tech- niques of Brower et al. (1975).

The leaves from 44 plants were assayed for CG levels on 9 April, and for one large plant with 8 stems, we took a leaf from the same position from each stem for CG analysis. Thirty of the same plants analysed on 9 April (early sample) were reassayed on 2 May (late sample), taking leaves from similar positions as before, but from adjacent stems. On this day we also took samples of latex from 31 of our plants for CG analysis. Latex was taken up with a pipette from leaf wounds inflicted by a pin, and placed into preweighed 5 ml volumetric flasks. These were reweighed wet, and then dried, weighed again to obtain percentage moisture, extracted in alcohol and assayed for CG levels as above. As CG concentrations in latex extracts were very high, we had to dilute extracts before spectrophotometry (0.5 ml of extract was made up to 1 ml with alcohol). Changes in CG levels from the first to the second sampling times were analysed using a paired t-test and the Wilcoxon signed rank test. Relationships between CG levels and other plant measures were assessed using regression analysis.

25 ¸

~" 2 0 -o ò

o} 15 d

o. IO-

N

Fig. 3

O Dry: y=0.933x-71.362, r 2=0.56,P<0,001 o

Wet: y=O.OO3x+0.919, r2=.001,P>0.84

(N=31) o

o o 0 0 ~ ; oO

t

8o 85 ~o ~s Per cent water in latex

Regression of the concentration of cardenolide in the [atex before and after drying against the percentage of watet in the latex, based on the same 31 latex samples. The "Dry" relationship is highly significant, whereas the "Wet" is not. These data suggest that an increase in water content of the latex is compensated for by an infiux of cardenolide with the result that the concentration remains constant in plants growing in the wild

Resu l t s

Plan t cardiac glycoside levels - While leaves taken from similar sized stems had similar CG levels, there was significant variation in CG concentrations between leaves from different sized sterns of the same plant. Thus the leaves from the six large stems (36-50 cm) of a single plant had a mean CG concentration of 235 gg/0.1 g DW (range 203-306), while leaves from the two smaller stems of the same plant (20 and 31 cm) had CG levels of 459 and 558 gg/0.1 g DW, respectively. Thus, by choosing leaves from similar positions and sized sterns for our life table data, we reduced the number of variables affecting interplant CG variation.

The 44 plants varied greatly in CG levels on 9 April, ranging from 166 to 744 gg/0.1 g DW, although most plants were distributed around the mean (419 ~tg/0.1 g DW, SD = 141, Fig. 2A). Plant CG levels generally increased during the season. The mean CG concentration in the 30 plant leaf samples taken on 2 May was 478 gg/0.1 g DW (SD= 181, range 235-978, Fig. 2B). This increase approached significance (paired t-test, t=1.549, d f=29 , p=0.0661, 1 tailed test), with 20 of the 30 plants showing an increase in CG levels compared to only 10 that declined (Wilcoxon signed rank test, Z -- 1.748, p < 0.05).

Survival o f f irst ins tar la rvae o f D a n a u s p l e x i p p u s C h e m o e c o l o g y 3 (1992) 85

Fig. 4 Dead first instar larvae of the monarch butterfly mired in A. humistrata latex that bted from the Leaf when the larvae took the[r first bite (A approx. 2X; B approx. 4X)

The concentration of cardiac glycosides in the latex was an order of magnitude higher than that recorded in leaf samples. In the undesiccated latex the CG concentration ranged from 666 to 1640 ~tg/0.1 g (X=1160, SD=260, N--31) and, in the same samples when desiccated, from 4405 to 23814 ]xg/0.1 g (X=9472, SD=4227, N=31 , Fig. 2C). Cardenolids therefore constitute an average of 1.2 to 9.5% of the wet and dry weight latex mass, respectively. There was no relationship between CG levels in latex, and the corresponding leaf samples gathered from the same plants on 2 May (r2=0.001, p=0.86). Figure 3 shows the linear regression of the concentration of CG in the latex before and after drying against the percentage of water in the latex. The slope of the wet line does not differ significantly from 0 (F(1,29)= 0.039, p>0.84), whereas the dry line is highly significant (F(1,29) = 36.9, p < 0.0001). (In fact a highly significant curvil- inear relationship for the dry line is indicated: Y=1073- 25 .5 .X+0.152 .X 2, F(2,28)=50.9, p<0.0001). These data suggest that an increase in water content of the latex is compensated for by an influx of cardenolide, with the result that the natural CG concentration remains constant.

Early stage survival - Monarch survival through the first instar on A . humistrata was very poor. In our initial count of wild monarchs on 44 plants, we recorded 108 eggs, and 20, 19, 23, 6 and 2, I, II, III, IV, and V instars respectively. If this were a stable and stationary age distribution, then one estimate of survival from egg to instar II would be 17.6°70 (19/108). Our field experiments allowed direct estimates of early stage survival.

For the first field experiment (15 April to 20 April), we put out 372 eggs, and also recorded the survival of a further 103 naturally laid eggs (giving an overall mean eggs/ plant of 13.9, SD=4.83, N=34) . These 475 eggs gave rise to 435 neonate larvae (92%). All except three of the experimental eggs hatched. Natural eggs were of course exposed to possible sources of mortality for much longer than our experimental eggs. Of the 103 natural eggs, 36 did not give rise to first instars. Of these, 10 turned grey to black and may have been parasitised, hut this was not ascertained. The other 26 were missing with no sign of larval feeding. We assumed these had either been eaten, or had been dislodged from the plant. Of the 435 neonate larvae, 50 moulted to the second instar; a survival rate of 11.5°70. During our checking of plants we found 124 dead first instar larvae. It is unusual to find cadavers in life table studies (Kyi et al. 1991). The high rate of re-

covery (29%) of dead firsts, can in part be attributed to the frequency and care with which we checked the plants. The other contributing factor was the condition of the cadavers; all were found mired, head first in latex, usually dried and desiccated (Fig. 4). The remaining 261 first instars could not be found, and we presume had died, been killed and/or (less likely) wandered or dropped off the plants.

For the second field experiment (23-30 April), we put out 234 eggs on 22 plants. We found no naturally laid eggs during this experiment. Again only 3 eggs failed to hatch. Of the 231 neonate larvae, only 8 made it to the second instar, a survival rate of 3.4°70 compared to 11.5070 in the first experiment (test of proportions, Z=3.49, p<0.05). We recovered 58 of the 231 larvae as mired cadavers in experiment 2 (25%), which did not differ significantly from experiment 1 (test of proportions, Z -- 1.103, P > 0.05).

Mortality of first instar larvae on A. hum&- trata was not related to larval density. In the first experiment, initial larval densities per plant ranged from 1 to 22 (if we in- clude all plants which had at least one natural egg). We re- gressed the number of surviving second instars against the initial number of eggs per plant, each expressed as a density per cm of plant stem length (Y=0.11.X-0.0004, F(1,36)= 18.29, P = 0.0001, Fig. 5A). The regression of the residuals from this regression against the initial number was not significant (r 2= 0, Fig. 5B), indicating no density dependence in our data. This procedure was not followed for the second experiment, because the experimental protocol determined the initial densities per plant with minimal variation (range I0-12).

There was also no difference in survival between ground predator excluded and non-excluded plants. Out of 30 first instars placed on to each treatment over 6 plants (i.e. 5 per treatment per plant), 2 survived to the second instar on each. Similar numbers of mired dead firsts were found in each treatment (12 on the excluded and 13 on the non-excluded, a mean of 42%). There was no indication that the exclusion of ground dwelling potential predators, such as ants and beetles, markedly improved first instar survival. We explore possible causes of first instar mortality below.

Monarch-milkweed interactions: relationship between plant variables and monarchs - Although we did not follow ovipositing females, frequent checking of plants ena- bled us to record 171 naturally laid eggs on 33 of our 44. exper-


.03F1 ~;.,-3.986E-4, ,~ =.334/ ~ .6 A A

.025 • • / .5 õ LnY = - 0 . 0 0 2 x - l . 461

ù02 7 ~ . • ~ . o

• 015 o " "

,,r-- • • • • ~. • ~ . ' " " " . " ' ~ . " .-..---:....~ o~t -~ ~ . - . . . . . .

.00 =~ 200 ' 460 ' 660 ' 800

"~ •.~ P lant c a r d e n o l i d e c o n c e n t r a t i o n = E ~g/o.lg DW O,I "T .6 i i « I

0 .02 .04 .06 .68 'I • ,2 .,~ ,~ .,~ .~ ~ o

Eggs per cm of plant

03

" t

" 0 . m 03

n-

ù027

ù015 -

ù01'

ù005.

0

ù 0 0 5

- . 01 -

- . 015

y=Ox-9.418E-22, r 2 = 0

• 0 .

B

I 1 • •

%

.1~ .02 .64 •06 .08 ,I .12 .14 . '6 .18 .2

E g g s p e r c m o f p l a n t

Fig. 5 Number of second instars per cm of plant plotted against the initial number of eggs per cm of plant (N=38) (A). The residuals from the regression (A) plotted against the initial number of eggs per cm of plant (B), indicates no significant density dependent mortality. See text for details

imental plants. These we related to CG levels in plants classif- ied as having low, medium, and high CG levels (Table 1). More eggs were found on plants in the intermediate CG fange than at lower or higher CG levels, a pattern consistent with out previous finding (Zalucki et aL 1990).

For our survival experiments only plants of similar stage (flowers/bud to flowers/young pods) were included. Even so there were significant relationships between stage (X) catagorised on a 4 point scale (1 = buds and flowers; 2 = flowers only; 3 = flowers and young pods; 4 = young pods only) and CG level (Y); plant CG increased with increasing plant stage, at least for plants with flowers/buds and above (Y = 46.9-X + 206.9, r 2 = 0.11, F(1,40) = 4.854, P = 0.0334).

,>

õ E .£

0 C). o B.

.5

, 4 "

, 3 '

. 2 '

o l

0 0

L n Y = - O . 2 8 5 x - O • 9 3 7

ii

B

3 4 5

Plant age. c l ass

Fig• 6 Proportion of hatched firsts surviving to the second instar plotted against plant CG (A) and plant age (B). Only plants (N =33) with at least 5 eggs were included when calculating proportion surviving

The proportion of larvae surviving from the first to second instar in experiment 1 (Y, log transformed) was negatively correlated with both plant CG (Fig. 6A) Y=0.002.X-1.461, r2=0.132, F(1,31)=4.716, P=0.0377), and plant age (Fig. 6B; Y=0.283.X-0.937, r2=0.111, F(1,31) = 3.886, P = 0•0577)•

In experiment 2, when survival was very poor, there was no relationship between survival and CG levels measured on 9 April (Y=0.00027.X-2.874, r2=0.018, F(1,22) = 0.378, P = 0.5454). The poorer survival during the second trial was associated with the general increase in both plant CG levels (see above), and plant stage (mean plant stage on 9 April equals 4.3 (SD = 1.08) compared to 5.5 (0.96) on 20 April). Because the second lot of leaf CG levels and latex CG levels were measured after the second survival trial, we did not regress survival against these.

There was no relationship between the percentage of dead first instars found mired in the latex and plant CG (r2=0.032, P=0.88) , nor between the percentage firsts missing and plant CG (r 2= 0.01, P = 0.97)• Even the relationship between survival to the second instar and plant CG was weak, with a large scatter around the regression line (Fig. 6A). Many factors other than CG levels no doubt influence survival of larvae on plants, such as temperature, variable cloud cover, relative humidity, leaf toughness, latex flow rate (see below), leaf quality (nutrient levels), rainfall and predators, as well as interactions between these. All these could not be dis- tinguished in out life table study, and would tend to obscure

Survival o f f irst ins tar l a rvae o f Danaus p l ex ippus C h e m o e c o l o g y 3 ( t992) 87

. 5 '

I

t~ > >

(b

<3

.2 '

.I

0

- 4 0 0

Y = - 1 , 9 1 X l O "4 X - 0 , 0 4 6

Table 1 Relationship between number of wild monarch eggs per plant for plents with eggs present grouped into 3 CG leveis (number of ptants without eggs in brackets), Means followed by the same Ietter ere not significantly different (paired t-test, p>O,05)

• Q • ® •

""""'"~~~" . 100-299 • . - ~ 300-499

o 500-800 Q

i i

-2~o o 200 4öo

a CG concentration

Fig. 7 Chenge in percentage first instar larval survival per plant plotted against change in CG concentration over the same tirne period, for the 14 individual plants for which both measures were aveilable

2G levels :pg/0.1 g DW)

No. of plants

12 (1) 12 (7)

9 (3)

Mean nümber of eggs per plant

3.7a 6.8b 4.9ab

SE

0.71 1,65 1.03

Table 2 Relationship between survival ,(P) and p[ants grouped into CG levels, for plants of similar age*, (N = number of plants in CG category). P is estimated from the ln[tial number of first Instars establis- hed summed over all plants in a particu!ar CG level. Variance estimate on P, VAR(P), calcutated after Seber (1982). Values of P fo/Iowed by the same letter are not significantly different (P > 0.05, test of proportions)

exctudes the youngest category {see text)

SG levels

any relationship between CG and survival. Thus to improve 100-299 both the life table estimates of survival, and to clarify any re- 300-399

¢00-499 lationship to CG we combined survival data for plants over 500-599 both experiments grouped into similar CG ranges. 300-800

The proportion of larvae surviving from the first to second instar (arcsin square-root transformed) generally declined with increasing CG ( Y = - 0 . 0 3 1 . X + 0 . 5 5 , r2=0.701, F(1,3)-7.038, P=0.0768, Table 2). The larger number of first instars at each CG level increased the preci- sion of survival estimates to the second instar (Table 2), relative to the individual plant data. Survival was generally higher at lower CG levels, than at higher CG levels (Table 2).

There were 14 plants between the two field trials for which we had two measures of survival and two measures of CG. We calculated the change in survival and CG concentratiõn for each plant (Fig. 7). The decline in survival with increased CG was significant (Y=1.91-10-4.X-0.Õ46, r z = 0.29, F(1,12) = 4.911, P = 0.0468), although a polynomial described the relationship better (Y=0.0006+ 1 .8 .10-4 .X - 8.4"10-7"X2,R2adj=0.487, F(1,11)=7.18, P=0.01) . The general decline in survival would also have been correlated with increased plant age (= phenological stage).

Monarch larval feeding behaviour in relation to latex flow

Because we found so many first instars in the field mired in latex (a total of 207 out of 691, including the predator exclusion experiment, or ca 30%), we attempted to simulate this effect in the laboratory. Of 12 first instars dabbed with a small drop of latex from A. curassavica, 10 were alive and apparently feeding normally after 24 h. Only two were dead, mired in latex. All third instars were alive and feeding 24 h after dabbing with latex from either A. curassavica or A. humistrata. Of 12 firsts dabbed with latex from A. humistrata 11 were feeding, and only 1 was mired after 24 h. All larvae showed characteristic behaviour after treatment with latex. They backed away, moved their head from side to side thereby transferring the latex to the leaf. They then either

12 8 8 5 4

Initial No. First P Var(P

138 120 83 52 41

0.123a 0,167b 0.048c 0,058c 0,049c

0,007 0,007 0,014 0,024 0,034

tried to feed or remained still for some time (2-10 min) before moving and feeding. These experiments were conducted in a closed Petrie dish, with larvae feeding on leaf discs lying on moistened filter paper so that the humidity was high. This may have prevented the latex from drying rapidly. Under field conditions at Cross Creek, the latex appears to solidify within 1-2 min. We therefore closely observed larvae of all stages (but mainly first instars) feeding in the field.

The behaviours associated with leaf feeding by first and third instar larvae on A. hum&trata observed are illustrated in Figure 8, and summarised in Figure 9. Before feeding, first instars would lay down a mat of silk with characteristic side-to-side movements of the head (we saw all larval stages doing this). The larvae would then position them- selves on the mat, usually with their heads towards the main midrib vein. With legs apparently anchored on the silk mat, the larvae either attempted to bite into the leaf, or struck the leaf surface with a r a n d sideways and downward motion of the head. This mandible slashing caused an immediate outflow of latex, which formed into a small globule. After mak- ing the initial incision, the larvae generally avoided the latex by positioning their heads to the side. They then usuaUy laid down more silk onto the mat, before moving back to feed at the same site as the initial incision, but now on the side distel to the midrib vein. This usually caused another outflow of latex, followed by a repetition of the above behaviours. The larvae were usually very adept at avoiding the latex outflow, hut on several occasions their sideways/or backward movement was not rapid enough to avoid a particularly large, or rapid outflow of latex. The larvae wetted with latex then rapidty backed away from the outfiow and moved their heads from side to side and dabbed the latex adhering to their body onto the leaf surface, and/or rubbed their mouthparts with their


Fig. 8 Reaction of a young third instar monarch larva to latex flow. A after having spun a silk mat to sit on while feeding, the larva ate two holes through the leaf; note the latex bleeding from the edge of the upper hole. B to avoid the latex flow, the larva pulls back its head by curving its body. C the larva then retreats to the surface of a large leaf vein where it will wait unti/the ratex begins to congeal before resuming its feeding. (Approx. 2X)

C Feed , , ~

Retreat \ \ ++/ æ

\ \ cM I \ ~ Retreat, ~ ' l / \ ~ dab latex onto leaf, ~ / \ ~ rub head with front ~ /

X legs and onto silk /

.#

Retreat and catalepsis

Fig. 9 Sequences of behaviour associated with feeding of first instar monarchs on leaves of Asclepias humistrata

forelegs and/or onto the silk mat (Fig. 9). They would then usually repeat the behaviours associated with biting the leaf. On a few occasions we also observed that a particularly rapid outflow of latex resulted in larvae actually swallowing some amount of the fluid. When this occurred, the anterior portion of the larva's body looked pale, the larva backed onto the silk mat, assumed a position with its prolegs anchored, and raised the anterior part of its body and tucked its head down. The larva would remain cataleptic in this position for up to 10 min before continuing its biting behaviour (Fig. 9).

After several sequential bites, a small hole was formed with a relatively large blob of latex on the midrib side (Fig. 8). (Often before taking a bite the larva appeared to spin silk onto the surface of the drying latex. This behaviour was particularly pronounced in fifth instars - see below). Once the larva had produced a hole, it would continue to ex- pand it by feeding on the side furtherest from the midrib vein, where the latex outflow was greatly reduced. In no case out of c a 20 first instars observed for 4 h did we see a first instar become mired, although we had expected to, based on our field observations of numerous dead larvae.

Larger larvae red in a similar way to first instars, except they seemed better at avoiding latex, and they fed more rapidly. We also observed a fifth instar beginning to

Survival of first instar larvae of Danaus plexippus Chemoecology 3 (1992) 89

Fig. 10 Latex avoidance behaviour by a fifth instar monarch larva feeding on A. humistrata. A the larva is notching the midvein of the petiole near the base of the leaf. B, C the larva pulls its head up from the vein wound and pushes the emerging latex droplets down the pJant stern. (Ap- prox. 1 X)

feed on a young leaf (Fig. 10). After laying down a silk mat on the underside of the leaf, the larva proceeded to bite into the ventral side of the petiole. This resulted in a copious outflow of latex. The larva then seemed to spin silk over the outflow, and pushed large blobs of latex to the side and off the plant. Once the flow had decreased, the larva continued to bite into the petiole, and repeated this entire sequence of behaviours. When latex flow had all but ceased, the larva ro- tated its position 180 ° and began to feed on the leaf on the end furtherest from the petiole. At no stage did we see this larva actively imbibing latex. In general, it appears that once fifth instar larvae have cut off the flow of latex to the leaf, they feed on the leaf edge and no further flow occurs. Re- peated damage to a leaf also results in cessation of latex flow (Fig. l l A vs l lB).

Discussion

The distribution of eggs on plants was as expected for females that selected plants with intermediate CG levels, as we found in our earlier study on A. humistrata at this site (Zalucki et al. 1990), and on A. fruticosa at an Aus- tralian site (Zalucki et al. 1989; Oyeyele & Zalucki 1990). Sim- ilar egg distributions were found on A. viridis (Van Hook & Zalucki 1991) at another Florida site. In our earlier work at Cross Creek (Zalucki et al. 1990) we also recorded a negative relationship between first instar survival and plant CG level. A similar but non-significant negative relationship was found by Cohen & Brower (1982) in a smaller study on A. humistrata (N = 9 plants) at a site less l:han 1.4 km from the current study area.

Survival of first instar larvae on A. hum•trata was very low (range 3.4 to 11.5°70) in our field experiments. We have summarised all similar data that are available from the literature for monarchs feeding on various milkweed hosts in the USA and Australia (Table 3). We calculated the percentage survival from egg to the second instar stage, as the estimate of first instar survival. First instars can be hard to

Fig. 11 A Two fifth [nstar larvae feeding on the edge of an A. humistrata leaf. The lack of latex flow can be because a larva previousty severed the pet[o[e, or because a larva repeatedly damaged the leaf. B cutting the edges of a leaf that has not been damaged by a larva pro- duces copious latex


Table 3 Early stage survival of monarchs recorded in the literature on various Asclepias spp. unless otherwise stated

N

qost Site Year Egg Source

4sclep&s humistrata

4. syriaca

4. fruticosa

4. curassavica

Calotropis procera

Cross Ck, Fla

Ordway, Fla

Saukville, Wi

Spencer, NSW Various, QId Moggill, QId

Mo9gil[, QId

Australia

1981 1983 1987 1990 1990 1983

1979

1979 1978/79 1985 1985

1990

1969

330 577

98 435" 231 ~

99

269

1990 1316

596 445

172

50

N %

II Surv

90 27 135 23

33 34 50 12

8 3.4 40 40

53 20

168 8 118 9 63 11 44 10

30 17

10 20

Cohen & Brower 1982 Malcolm et al. 1987 Zalucki et a/. 1990 Trial 1, this study Trial 2, this study Malcolm et al. 1987

Borkin 1982

James 1981 Zalucki & Kitching 1982 b M. P. Zalucki, ünpubl. M. P. Zalucki, unpubl.

M. P. Zalucki, unpubl.

Orrell 1970

* based on eggs hatched

find. It is usually easier to find eggs, and second instars, and results for these in the literature may be more reliable. Only our present study estimates first instar survival directly.

There appears to be considerable variation in survival to the second instar on the same host at the same site (A. humistrata 3.4-34% between years), and between sites within years (1983, Cross Creek 23% vs Ordway 40%, Table 3). Other species ( e .g .A . fruticosa) gave very constant survival values across sites and years (8-11%, Table 3), although Zalucki & Kitching (1982b) give a much fuller account of the variable mortality on this host in southeast Queensland, Aus- tralia. In the current study, we found considerable variation in survival between plants of A. humistrata at our field site in our first trial (0-30% mean = 11.5%). There was less variation in the second trial (0-20°70, mean = 3.4%), with values heavily skewed to the left. Some of this variation was statisti- cally related to CG levels in the plant, (13%), and/or plant age (11%). As in our earlier study (Zalucki et al. 1990), we found a weak negative relationship between survival and CG (Figs. 6A, 7; Table 2). The decline in survival between our two trials was also associated with a general increase in CG levels (and plant age). As in any field based study, it is difficult to disassociate plant measures that co-vary (Miles et aL 1982).

Variation in plant CG level, as measured in out leaf samples, was not the only factor contributing to early stage survival. Many factors can cause mortality of neonate larvae (Kyi et al. 1991). Dempster 's (1983) review of life table studies of phytophagous insects showed that weather variables, particularly temperature, and predators are often key mortality agents. We did not record temperatures (and it did not rain during the period of observation), but our small predator exclusion trial indicated that ground dwelling predators were not a major mortality factor. Other factors that could contribute to mortality of neonate larvae includes leaf toughness (Aide & Londono 1989) (which may be plant age related), plant latex content, the rate at which the latex congeals (which must be temperature and humidity related), rainfall, and the interactions between these.

We did not record toughness per se, but the negative effect of plant stage on survival indicates this may be a factor. Plant CG and age were positively related, and a more detailed study will be required to disentangle these effects. Out study did reveal a major effect of plant latex. An average of 30% of the first instars were found mired in latex. Our data therefore support Dussourd & Eisner's (1987) sug- gestion that latex may be a major defence mechanism of milkweeds against herbivores, even those that are specialised feed- ers.

Milkweed latex has been extensively analysed chemically, ever since milkweeds were determined to be a possible source of rubber and oll (e.g. Nielsen et al. 1977). It contains various compounds, including the isoprene derivatives c~- and 13-amyrin, and their acetates (Van Emon & Seiber 1985), and can also be particularly high in CG as found in this study, constituting on average 1.2 and 9.5% of the mass of the wet and dry latex, respectively. Interspecies differences in concentrations and kinds of cardenolids occur (Seiber et al. 1982) and as found by Nelson et al. (1981) CG levels changed sea- sonally in A. eriocarpa in all above ground plant patts as well as the latex. We also found an increase in CG levels in leaf samples during out study. Interestingly, latex CG was not correlated with leaf sample CG. This lack of a correlation could be related to changing moisture levels in various plant parts during the day as weil as with the seasonal increase in water stress (Nelson et al. 1981). Regression of cardenolide concentrations against the water content of the latex leads to the con- clusion that an increase in watet content of the latex is compensated for by an influx of cardenolide. Thus, at least under the conditions found in the natural habitat of A. humistrata in north central Florida, cardenolide concentration in the latex system appears to remain constant irrespective of its watet content. The proportion of various kinds of CG in latex also differs from that found in leaves (Nishio 1980; Nelson et al. 1981). In general low polarity CG's were found to be more concentrated in the latex (Selber et al. 1982). Monarchs do not sequester low polarity CG's very efficiently (Nelson 1991).

Survival o f f irs t ins tar l a rvae o f Danaus p l ex ippus C h e m o e c o l o g y 3 (1992) 91

The latex in A. humistrata is contained under pressure in a reticulated sealed system of vessels called laticifers (Blaser 1945; Wilson & Mahlberg 1980; Lucansky & Cloug 1986). When a leaf is punctured or its petiole cut, latex flows out and rapidly coagulates on contact with air. If leaves are extensively or repeatedly damaged, latex flow ceases. Either, (1) the latex supply is exhausted or, (2) the latex pressure drops to that of the external air pressure or, (3) the plant can restrict outflow at the midrib or petiole when the leaf is damaged distally. We cut leaves at right angles to the midrib vein with a pair of scissors. Latex outflow was rapid, followed by coagulation (Fig. l lB). After repeated cutting of the same leaf, the latex flow ceased. On leaves that were being vora- ciously consumed by fifth instars, that had not severed the petiole, we found that pin pricks into this leaf, and into leaves lower down the same stem of the plant did not result in any latex flow (Fig. 1 lA).

Our observations of all larval stages of monarchs feeding on A. humistrata indicated that they were very adept both at avoiding latex and circumventing latex flow by isolating leaf tissue from the latex source (see also Dussourd 1986, 1990; Dussourd & Eisner 1987). This was done by a form of trenching, midrib vein cutting, petiole snipping and / or by repeated, successive damage to a leaf (Figs 8, 10, 11). Dillon et al. (1983) have described similar behaviours in a sphingid larvae feeding on latex bearing and toxic plants and Compton (1987, 1989) described vein-cutting behaviours in Danaus chrysippus and other lepidopterans. Similarly, early instar larvae of Euploea tore first isolate leaf sections before feeding, by systematically snipping the veinlets of leaves of their Apocynaceae, Asclepiadaceae, and Moraceae hosts (Am- batali 1990). Later instars isolate larger areas, and whole leaves, by petiole and midrib vein snipping.

While performing these operations, monarch larvae appeared to attempt to avoid any contact with the latex. If latex did contact their mouthparts or head region, the larvae either rapidly cleaned it away or involuntarily imbibed it and went cataleptic. Thus although we found many first instar larvae mired in latex in our field work, we did not actually observe any get stuck in this fashion. So how do larvae become mired? One possible clue is contained in the observation that first instars occasionally could not avoid imbibing latex. Latex contains very high concentrations of low polarity CG, and other compounds which may adversely affect the gut, and health of larvae (e.g. o~- and [3-amyrins, rubber and various sterols). Larvae appeared traumatised following such an ex- perience, although they did resume feeding after 10 min. Per- haps repeated exposure to small amounts of latex, may weak- en larvae and their ability to avoid latex outfiow during feeding, leading to the miring effect. Another possibility may be that larger imbibed doses lead to catalepsis for a more pro- longed time, sufficient for the latex to cause adherence to the leaf. The lack of relationship between dead mired larvae and leaf CG, may be due to the absence of a relationship between leaf and latex CG, which in turn may be due to the fact that the concentration of CG in the latex so greatly exceeds (> 20 times) that in the leaves. The overiding importance of CG in the latex may also explain the weakness of the negative relationship we found between survival to the second instar and leaf CG. We need to measure latex CG levels, at the time the larva actually take their first bite and relate these to first instar survival. Leaf toughness may also have a significant ef-

fect on the ability of larvae to avoid latex. If a larva is feeding on a tough leaf, it might be more difficult to bite into the leaf, and avoid latex outflow at the same time. We predict poor survival on plants with tough leaves that have high CG levels in the latex. Only detailed field experiments will be able to test this.

Asclepiäs hum&trata in a laboratory choice experiment was not a preferred host of first instar monarchs (L. P. Brower & M. P. Zalucki, unpubl.). Perhaps the missing larvae in the field moved oft plants. If this õccurred, their survival away from plants would be negligible. Third, and larger instar monarch larvae are capable of moving between host plants (Rawlins & Lederhouse 1981; Borkin 1982; Zalucki & Kitching 1982b; M. P. Zalucki unpubl.). First instars on A. humistrata would not survive long in the dry, high temperature, and exposed conditions away from this host plant. Only careful analysis of the plant 's chemistry, and the sensory physiology of larvae will resolve host preference behaviour of monarch larvae. Most insect-host plant studies utilise large IV and V instar larvae (e.g. Mathavan & Bhaskaran 1975; Slans- ky & Feeny 1977; Scriber 1981). This may make the construction of feeding budgets easier, but orte wonders about the ecological relevance of such studies. Clearly, the first bites of neonate larvae can be crucial to survival and this aspect of larval ecology needs further study (see also Chapman & Ber- nays 1989).

Monarch larvae in the third instar and older, offen leave the plant on which they are feeding, and move onto another plant. It is not clear why they do this when there is sufficient leaf matter to complete development on the origi- nal plants. Possibly plants change in response to larval feeding, and larva move in response, as has been suggested for other plants (Edwards & Wratten 1983).

Low survival of early stage monarchs on a milkweed host, to which they are supposed to be highly adapted (Ackery & Vane-Wright 1984) may be considered odd. Others have also recorded poor early stage survival of phytophagous insects, or low herbivory in relation to plant secondary compounds (e.g. Isman & Duffey 1982; Louda & Rodman 1983), although this is not always the case (Fox & Macauley 1977). The high variability of milkweeds in CG levels and type, and latex content, within (Selber et al. 1982) and between (Malcolm & Brower 1989) host plant species, both tem- porally (Nelson et al. 1981) and spatially (Malcolm et al. 1989), combined with the monarchs migratory movements, may preclude close adaptation to any one milkweed host, at least in North America. Classic evolutionary theory would predict better performance if monarchs were restricted to a single host in a situation where they can' t migrate, such as occurs on islands. Monarchs as well as a few milkweed species colonised the South Pacific in the last century, and occur on a number of islands. This natural experiment may enable us to test this idea.

Acknowledgements

We wish to thank Alfonso Alonso M. and To- nya Van Hook for help with the field research and laboratory work, Marion Griffiths for reading an earlier version of the manuscript, Jocelyne Campbell for typing and Zane Hogan of Micanopy, Florida, for access to his land. This work was

92 Chemoecology 3 (1992) Zalucki & Brower

supported by a University of Queensland International Grant Scheme to M. P. Z. and by a University of Florida Division of Sponsored Research Grant to L. P. B.

References

Ackery PR, Vane-Wright RI (1984) Milkweed Butterflies: Their Cladistics and Biology. Ithaca: Cornell University Press

Aide TM, Londono EC (1989) The effects of rapid leaf expansion on the growth and survivorship of a lepidopteran herbivore. Oikos 55:66-70

Ambatali FA (1990) Influence of host plants on the development, feeding behaviour and parasitism of Euploea core corinna WS Macleay (Lepi- doptera: Nymphalidae). Master of Agric Studies Thesis, University of Queensland

Blaser HW (1945) Anatomy of Cryptostegia grandiflora with special refe- rence to the latex system. Am J Bot 32:135-141

Borkin SS (1982) Notes on shifting distribution patterns and survival of immature Danaus plexippus (Lepidoptera: Danaidae) on the food plant Asclepias syriaca. Great Lakes Entomol 15:199-206

Brewer J (1977) Short lived phenomena. News Lepid Soc 4:7 Brower LP (1969) Ecologieal chemistry. Sci Amer 200 (2):22-29 Brower LP (1984) Chemical defence in bntterflies. Symp R Entomol Soc

Lond 11:109-134 (= Vane-Wright RI, Ackery PR (eds) The Biology of Butterflies. London: Academic Press)

Brower LP, Brower JV (1964) Birds, butterflies, and plant poisons: a study in ecological chemistry. Zoologica 49:137-159

Brower LP, Fink LS (1985) A natural toxic defence system in butterflies versus birds. Ann NY Aead Sci 443:171-188

Brower LP, Brower JVZ, Corvino JM (1967) Plant poisons in a terrestrial food chain. Proc Natl Acad Sci USA 57:893-898

Brower LP, Ryerson WN, Coppinger LL, Glazier SC (1968) Ecological chemistry and the palatability spectrum. Science 161: 1349-1351

Brower LP, McEvoy PB, Williamson KL, Flannery MA (1972) Variation in cardiac glycoside content of monarch butterflies from natural populations in eastern north America. Science 177:426-429

Brower LP, Edmunds M, Moffitt CM (1975) Cardenolide contents and palatability of a population of Danaus chrysippus butterflies from West Africa. J Entomol 49:183-196

Brower LP, Nelson CJ, Seiber JN, Fink LS, B ondC (1988) Exaptation as an alternative to coevolution in the cardenolide-based chemical defence of monarch butterflies (Danaus plexippus L.) against avian predators. Pp 447-475 in Spencer KC (ed.) Chemical Mediation of Co- evolution. New York: Academic Press

Chapman RF, Bernays EA (1989) Insect behavionr at the leaf surface and learning as aspects of host plant selection. Experientia 45:215-222

Cohen JA (1983) Chemical interactions among milkweed plants (Asclepi- adaceae) and lepidopteran herbivores. PhD Thesis, University of Flo- rida, Gainesville

Cohen JA, Brower LP (1982) Oviposition and larval success of wild monarch butterflies (Lepidoptera: Danaidae) in relation to host plant size and cardenolide concentration. J Kansas Entomol Soc 55:343-348

Compton SG (1987) Aganais speciosa and Danaus chrysippus (Lepidopte- ra) sabotage the latex defenses of their host plants. Ecol Entomol 12:115-118

Compton SG (1989) Sabotage of latex defenses by caterpillars feeding on fig trees. S Afr J Sci 85:605-606

Dempster JP (1983) The natural control of populations of butterflies and moths. Biol Rev 58:461-481

Dillon PM, Lowrie S, McKey D (1983) Disarming the "evil woman": petiole constriction by a sphingid larva circumvents mechanical defenses of its hostplant, Cnidoscolus urens (Euphorbiaceae). Biotropica 15:112-116

Dixon CA, Erikson JM, KeUett DN, Rothschild M (1978) Some adaptations between Danaus plexippus and its food plant, with notes on Da- haus chrysippus and Euploea core (Insecta: Lepidoptera). J Zool Lond 185:437-467

Dussourd DE (1986) Adaptations of insect herbivores to plant defences. PhD Thesis, Cornell University, Ithaca NY

Dussourd DE (1990) The vein drain; or how insects outsmart plants. Nat Hist 90:44-49

Dussourd DE, Eisner T (1987) Vein-cutting behaviour: insect counterploy to the latex defence of plants. Science 237:898-901

Edwards P J, Wratten SD (1983) Wound induced defences in plants and their consequences for patterns of insect grazing. Oecologia 59:88-93

Erickson JM (1973) The utilization of various Asclepias species by larvae of the monarch butterfly, Danaus plexippus. Psyche 80:230-244

Fisk J (1980) Effects of HCN, phenolic acids and related compounds in Sorghum bicolor on the feeding behaviour of the planthopper Peregri- nus maidis. Entomol exp appl 27:211-222

Fox LR, Macauley BJ (1977) Insect grazing on Eucalyptus in response to variation in leaf tannins and nitrogen. Oecologia 29:145-162

Harborne JB (1988) Introduction to Ecological Biochemistry. 3rd ed. New York: Academic Press

Hassel MP (1985) Insect natural enemies as regnlating factors. J Anim Ecol 54:323-334

Hutchins RFN, Sutherland ORW, Gnanasunderam C, Greenfield W J, Williams EM, Wright HJ (1984) Toxicity of nitro compounds from Lotus pedunculatus to grass grub (Costelytra zealandica) (Coleoptera: Scarabeidae). J Chem Eeol 10:81-93

Isman MB, Duffey SS (1982) Toxicity of tomato phenolic compounds to the fruitworm, Heliothis zea. Entomol exp appl 31:370-376

J ames DG (1981) Studies on a winter breeding population of Danaus plexippus (L.) (Lepidoptera: Nymphalidae) at Spencer, New South Wales. Gen Appl Entomol 13:47-53

Jungreis AM, Vaughan GL (1977) Insensitivity of lepidopteran tissue to ouabain: absence of ouabain binding and Na +, K + ATPases in larval and adult midgut. J Insect Physiol 23:503-509

Kyi A, Zalucki MP, Titmarsh IJ (1991) Factors affecting the survival of the early stages of Heliothis armigera (Huebner) (Lepidoptera: Noctu- idae). Bull Entomol Res 81: 263-271

Louda SM, Rodman JE (1983) Concentration of glucosinolates in relation to habitat and insect herbivory for the native crucifer Cardamine cor- difolia. Biochem System Ecol 11: 199-207

Lucansky TW, Cloug KT (1986) Comparative anatomy and morphology of Asclepias perennis and Asclepias tuberosa subspecies rolfsii. Bot Gaz 147:290-301

Malcolm SB, Brower LP (1986) Selective oviposition by monarch butterflies (Danaus plexippus L.) in a mixed stand of Asclepias curassavica L. and A. incarnata L. in south Florida. J Lep Soc 40:255-263

Malcolm SB, Brower LP (1989) Evolutionary and ecological implications of cardenolide sequestration in the monarch butterfly. Experientia 45:284-295

Malcolm SB, Zalucki MP (eds) (1992) Biology and Conservation of the Monarch Butterfly. Los Angeles: Natural History Museum of Los An- geles County

Malcolm SB, Cockrell B J, .Brower LP (1987) Monarch butterfly voltinism: effects of temperature constraints at different latitudes. Oikos 49:77- 82

Malcolm SB, Cockrell BJ, Brower LP (1989) Cardenolide fingerprint of monarch butterflies reared on common milkweed, Asclepias syriaca L. J Chem Ecol 15:819-853

Mathavan S, Bhaskaran R (1975) Food selection and utilisation in a da- naid butterfly. Oecologia 8:55-62

Miles PW, Aspinall D, Rosenberg L (1982) Performance of the cabbage aphid, Brevicoryne brassicae (L.), on water-stressed rape plants, in relation to changes in their chemical composition. Aust J Zool 30: 337- 345

Morrow PA, Fox LR (1980) Effects of variation in Eucalyptus essential oil yields on insect growth and grazing damage. Oecologia 42:209-219

Nelson CJ (1992) Sequestration and storage of cardenolides and cardenolide glycosides by Danausplexippus L. and D. chrysippuspetilia (Stoff) when reared on Asclepias fruticosa L.; a review of some factors influ- encing sequestration process. In press in Malcolm SB, Zalucki MP (eds) Biology and Conservation of the Monarch Butterfly. Los Ange- les: Natural History Museum of Los Angeles County

Nelson CJ, Seiber JN, Brower LP (1981) Seasonal and intraplant variation of cardenolide content in the California milkweed, Asclepias eriocarpa, and implications for plant defence. J Chem Eco 7:981-1010

Nielsen PE, Nishimura H, Otvos JW, Calvin M (1977) Plant crops as a source of fuel and hydrocarbon-like materials. Science 198:942-944

Nishio S (1980) The fates and adaptive significance of cardenolides seque- stered by larvae of Danaus plexippus (L.) and Cycnia inoptinatus (Hy. Edwards). PhD Thesis, University of Georgia

Orrell J (1970) The baby is a cannibal. Wildlife 7:44-47 Oyeyele S, Zalucki MP (1990) Cardiac glycosides and oviposition by Da-

naus plexippus on Asclepias fruticosa in south-east Queensland (Au- stralia), with notes on the effects of plant nitrogen content. Ecol Ento- mol 15:177-185

Rawlins JE, Lederhouse RC (1981) Developmental influences of thermal behaviour on monarch caterpillars (Danaus plexippus): an adaptation for migration (Lepidoptera: Nymphalidae: Danainae). J Kansas Ento- mol Soc 54:387-408

Rothschild M (1977) The cat-like caterpillar. News Lepid Soc 6:9 Scriber JM (1981) Sequential diets, metabolic costs, and growth of Spo-

doptera eridiana (Lepidoptera: Noctuidae) feeding on dill, lima beans, and cabbage. Oecologia 51:175-180

Survival of first instar larvae of Danaus plexippus Chemoecology 3 (1992) 93

Seber GAF (1982) The Estimation of Animal Abundance. London: Griffin & Co

Seiber JN, Tuskes TM, Brower LP, Nelson CN (1980) Pharmacodynamics of some milkweed cardenolides fed to larvae of the monarch butterfly (Danaus plexippus L.). J Chem Ecol 6:321-339

Seiber JN, Nelson C J, Lee SM (1982) Cardenolides in the latcx and leaves of seven Asclepias species and Calotropis procera. Phytochemistry 21:2343-2348

Singer MC (1984) Butterfly-hostplant relationships: host quality, adult choice and larval success. Symp R Entomol Soc Lond 11:81-88 (= Vane-Wright RI, Ackery PR (eds) The Biology of Butterflies. London: Academic Press)

Slansky F Jr, Feeny PP (1977) Maximization of the rate of nitrogen accu- mulation by larvae of the cabbage butterfty on wild and cultivated food plants. Ecol Monogr 47:209-228

Todd GW, Getahum A, Cress DC (1971) Resistance in barley to the green- bug, Schizaphis graminum. 1. Toxicity of phenolic and flavonoid cõmpounds and related substances. Ann Entomol Soc Am 64:718- 722

Van Emon JV, Seiber JN (1985) Chemical constituents and energy content of two milkweeds, Asclepias speciosa and A. curassavica. Econ Bot 39:47-55

Van Hook T, Zalucki MP (1991) Oviposition by Danausplexippus on As- clepias viridis in northern Florida. J Lep Soc 45:215-221

Vaughan GL, Jungreis AM (1977) lnsensitivity of lepidopteran tissue to oabian: physiological mechanisrns for protection from cardiac glycosides. J Insect Physiol 23:585-589

Wilson K J, Mahlberg PG (1980) Ultrastructure of developing and mature nonarticulated laticifers in the milkweed, Asclepias syriaca L. (Ascle- piadaceae). Am J Bot 67:1160-1170

Zalucki MP, Kitching RL (1982a) Dynamics of oviposition in Danaus plexippus (Insecta: Lepidoptera) on milkweed, Asclepias spp. J Zool 198:103-116

Zalucki MP, Kitching RL (1982b) Temporal and spatial variation of mortality in field populations of Danaus plexippus L. and D. chrysippus L. larvae (Lepidoptera: Nymphalidae). Oecologia 53:201-207

Zalucki MP, Oyeyele S, Vowles P (1989) Selective oviposition by Danaus plexippus L. (Lepidoptera: Nymphalidae) in a mixed stand of Asclepi- asfruticosa and A. curassavica in southeast Queensland. J Aust Ento- mol Soc 28:141-146

Zalucki MP, Brower LP, Malcolm SB (1990) Oviposition by Danausplex- ippus in relation to cardenolide content of three Asclepias species in the southeastern U.S.A. Ecol Entomol 15:231-240

Survival of first instar larvae of Danaus plexippus (Lepidoptera: Danainae) in relation to cardiac glycoside and latex content of Asclepias humistrata (Asclepiadaceae

Documents