1

The effects of leg loss and regeneration on prey capture, growth and development time in wolf spiders

A thesis submitted to the

Division of Research and Advanced Studies of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

Master of Science

In the Department of Biological Sciences of the College of Arts and Sciences

2005

by

Kerri M. Wrinn

B.S., Berry College, 2002

Committee: Dr. George Uetz, Chair Dr. Elke Buschbeck Dr. Michal Polak

2 Abstract:

I addressed the effects of autotomy and regeneration on foraging success, body condition,

growth (size and weight) and development time (molt interval) for Schizocosa ocreata wolf

spiders in the laboratory and field. Frequency of autotomy in the field was high (12-19%) and body size, weight and condition were significantly lower in individuals with

missing/regenerating legs. Regeneration of a single leg increased development time in

laboratory reared spiders but decreased growth in field-caught individuals. Autotomy of

multiple legs decreased both growth and development time.

Laboratory experiments carried out in artificial conditions showed no effects of autotomy

or regeneration on prey capture efficiency or vibration sensory abilities. However, spiders tested in a semi-natural habitat (a leaf litter filled mesocosm) with a missing or regenerating leg had reduced prey capture rates. This suggests that the effects of autotomy and regeneration on foraging may only be apparent in more complex environments encountered in nature.

i

ii Acknowledgements:

In order to complete the following research I relied on the assistance and guidance of

several people. First, I would like to thank my advisor Dr. George Uetz for all of his advice and

assistance and for taking the time to answer my numerous questions even when he was very

busy. I would also like to thank my committee members Dr. Elke Buschbeck and Dr. Mike

Polak for their insights about my experimental design and statistics.

The following undergraduates provided much-needed help with spider collecting, care

and maintainence: S. Dougherty, C. Kluener, S. Pruiett, M. Salpietra, M. Skelton, R. Srivastava,

A. Wick, and E. White. I’d also like to offer a special thanks to Dr. Andy Roberts and the

graduate students in my lab (J. Gibson, C. Harris, J. Johns, and J. Milliser) for their ideas, advice and emotional support. Finally I would like to thank my family and friends for their constant support of my interest in arachnids.

This research was supported by the American Arachnological Society, Sigma Xi, the

National Science Foundation (to GWU) and a Wieman/Wendel/Benedict foundation research award through the University of Cincinnati Biology Department.

iii Table of Contents

Abstract i Acknowledgements iii List of Tables vi List of Figures vii

General Introduction 1 Study Organism 2 Research Objectives Chapter 1 3 Research Objectives Chapter 2 4 References 7

Chapter

I The impacts of leg loss/regeneration on fitness as measured by body 11 condition; growth and development in Schizocosa ocreata (Araneae: Lycosidae) Abstract 12 Introduction 13 General Methods 16 Objective 1 17 Methods 17 Data Analysis 18 Results 18 Discussion 21 Objective 2 24 Methods 24 Data Analysis 26 Results 26 Discussion 28 Conclusions 31 References 32 Tables and Figures 37

II The effects of autotomy and regeneration on prey capture rate, 47 efficiency, and sensory detection in a wolf spider (Araneae: Lycosidae). Abstract 48 Introduction 49 General Methods 52 Objective 1 53 Methods 53 Data Analysis 54 Results 54 Discussion 55

iv

Table of Contents

II Objective 2 57 Methods 57 Data Analysis 59 Results 59 Discussion 60 Objective 3 61 Methods 62 Data Analysis 63 Results 63 Discussion 63 Conclusions 65 References 67 Tables and Figures 71

General Conclusions 83 References 88

v List of Tables Table

1.1 Differences in frequency of leg loss in the field for S. ocreata.

1.2 Results of Three-Factor ANOVA (injury status, year, and season [year]) comparisons for autotomized and intact spiders of Spring and Fall 2003 and 2004

1.3 Results of Three-Factor ANOVA (injury status, season, and week [season]) comparisons for a generation of spiders (Fall 2003 and Spring 2004).

1.4 Comparisons of mean ±SE cephalothorax widths, weights, and body condition measures for autotomized (AUT) and intact (INT) spiders of the same generation by season and by week.

1.5 Two sample t-tests were used to compare molt intervals (days) between autotomized and intact lab-reared spiders in 2003.

2.1 Repeated measures MANOVA results for prey capture of spiders given crickets 33% or 50% of the spiders body weight at Trials 1-3.

2.2 Repeated measures MANOVA results for prey capture rate at 6 and 24 hours by manipulated or intact spiders over four trial periods.

2.3 Repeated measures MANOVA of latency to orient over five trial periods for spiders given crickets weighing 25%, 33%, or 50% of their weight.

2.4 Repeated measures MANOVA of latency to attack over five trial periods for spiders given crickets weighing 25% or 33% of their weight.

2.5 Repeated measures MANOVA of number of attacks over five trial periods for spiders given crickets weighing 25% or 33% of their weight.

2.6 Repeated measures MANOVA of subdue time over five trial periods for autotomized and intact spiders given crickets 25% or 33% of their body weight.

2.7 McNemar tests for paired comparisons between consecutive trial periods for autotomized and intact spiders given crickets weighing 25%, 33%, and 50% of their body weights

vi List of Figures

Figure

1.1 Percentage of legs lost or regenerated according to position for field caught S. ocreata

1.2 Mean (±SE) of A) Average weights, B) cephalothorax widths and C) Body conditions for autotomized and intact field caught S. ocreata captured during comparable times in spring and fall of 2003 and 2004.

1.3 Early and midseason comparisons of BCI (Residuals of weight x cephalothorax width) for A) Spring 2004 and B) Fall 2004.

1.4 Mean (±SE) for (A) cephalothorax widths and (B) weights between spiders of the same age which were caught in the field and then underwent autotomy in the lab, according to treatment (Fall 2004)

1.5A Mean (±SE) molt interval (days) for spiders that were intact or with one or both forelegs autotomized.

1.5B Mean (±SE) weight (mg) for spiders with one or both forelegs autotomized.

1.5C Mean (±SE) cephalothorax width (mm) for spiders that were intact or with one or both forelegs autotomized

2.1 Measurement of error angle of orientation to a prey stimulus.

2.2 Comparisons of prey capture by spiders (manipulated and intact) over 4 trial periods at 6 and 24 hours.

2.3 Responses of S. ocreata to vibrations from a prey stimulus in the absence of visual cues. A) Percentage of spiders which oriented. B) Percentage of spiders which oriented correctly (with error angle less than stimulus angle).

2.4 Mean (±SE) latency to orient to a vibratory prey stimulus by spiders with foreleg intact (int), autotomized (aut), regenerated partially (reg 1), or regenerated fully (reg 2).

2.5 Mean (±SE) error angles of orientation to a vibratory prey stimulus for spiders with a foreleg intact (int), autotomized (aut), regenerated partially (reg 1), and regenerated fully (reg 2).

vii General Introduction:

The ability to autotomize (self amputate) and later regenerate appendages is common across many animal taxa including vertebrates, echinoderms, and arthropods (Arnold, 1984;

Ramsay et al, 2001; Juanes and Smith, 1995; Roth and Roth, 1984). Tradeoffs exist between the

costs and benefits of each of these two processes, and not all animals that autotomize have the

ability to regenerate. Studies of the costs and benefits of autotomy and regeneration may

therefore lead to a greater understanding of the selection forces behind their evolution.

In arthropods, regeneration is linked to the process of molting (shedding of the

exoskeleton) during growth (Goss, 1969). Many arthropods molt frequently and can thus also regenerate lost limbs in a relatively short time, making them an ideal group for studying

autotomy and regeneration. Extensive studies have addressed both processes in crustaceans (see

Juanes and Smith 1995 for a review), and insects have received some attention in these areas as well (Bell and Adiyodi, 1981; Stoks, 1999; Bateman and Fleming, 2005). Comparatively little is

known about autotomy and regeneration in arachnids. These processes may be particularly

important in spiders, as they use their legs for multiple functions e.g. locomotion, prey capture,

and communication. Previous studies have addressed the effects of autotomy on competition

(Dodson and Beck, 1993; Johnson and Jakob, 1999; Dodson and Schwaab, 2001; Taylor and

Jackson, 2003), prey capture (Amaya et al, 2001, Brueseke et al, 2001), growth (Weissmann and

Vollrath, 1998) and mating (Brautigam and Persons, 2003; Taylor, Roberts and Uetz, .

unpublished) in a variety of spider species. Spiders of most species do not molt after becoming

mature, and thus also lose the ability to regenerate lost limbs at adulthood. Probably for this

reason, only a few studies have addressed regeneration in spiders (Roth and Roth, 1984;

1 Vollrath, 1990 and references therein) and its effects on prey capture (Vollrath, 1995) and mating

success (Uetz et al, 1996).

Study Organism:

In the following studies I used the wolf spider Schizocosa ocreata (Hentz) to examine

autotomy and regeneration. The behavior of this species has been well-studied, although only a

few of these studies have addressed regeneration and autotomy. Amaya et al. (2001) found no

effects of autotomy on prey capture (number of attacks) in adult females. Uetz et al. (1996)

found that males with a partially regenerated foreleg were preferred as mates less often. Taylo,

Roberts and Uetzl, (unpublished) discovered that autotomy of one or both forelegs affected

courtship patterns, and in a separate study found that females were less likely to mate with males

with both forelegs autotomized. Therefore much is left to be learned about autotomy and

regeneration as potentially important evolutionary forces in this species. These spiders are

commonly found in the leaf litter of Eastern deciduous forests (Cady, 1984), and exhibit a high

occurrence of autotomy and regeneration in the field (13-19%) (Uetz et. al.,1996). Additionally,

it is easy to induce autotomy in the lab for these spiders (personal observation). Finally, these

spiders are ideal study animals because they are abundant, easy to maintain in the lab, and have

short growth periods and frequent molts.

Objectives:

This thesis encompasses two main areas of study: 1) the effects of naturally-occurring

and experimental autotomy and regeneration on body measurements (as indicators of foraging success) as well as development time (intermolt interval); and, 2) the effects of autotomy and regeneration on prey capture ability. Thus, for the first chapter, the costs of autotomy and regeneration to the spider’s overall condition are considered from a developmental or energetic

2 perspective. The second chapter focuses on the specific mechanical and sensory functions which may be impaired by having a missing or regenerating leg. Taken together, the objectives of these

studies address some of the tradeoffs between the costs and benefits of autotomy and

regeneration in S. ocreata. The results of this study may provide insights to why these processes

have been selected for and maintained over evolutionary time in this species, and by extension, spiders in general.

Chapter I: The effects of leg loss/regeneration on body condition, growth and development time; characteristics with the potential to impact fitness

Objective 1: Naturally occurring leg loss frequency, size, weight and body condition

Autotomy occurs frequently in many groups of spiders (Roth and Roth, 1984; Vollrath,

1990) as 5-20% of all spiders collected in field populations are missing legs (Foelix, 1996). This

objective examined the frequency of autotomy and regeneration of juvenile S. ocreata in the

field and tested the hypothesis that leg autotomy and regeneration would negatively impact body characteristics linked to fitness. Size (cephalothorax width), weight, and body condition were measured in field caught spiders with either intact or autotomized/regenerating legs. Reduction in foraging at a juvenile stage could potentially affect size and weight as an adult (Beck and

Connor, 1992; Uetz et al, 1996, Weissmann and Vollrath, 1998). Larger spiders that are in better condition generally have higher fitness, as shown by increased fecundity in females and higher mating success in males (Wise, 1975; Beck and Connor, 1992; Simpson, 1995; Spence et al.,

1996). Therefore, decreased foraging due to autotomy and regeneration of appendages during development has the potential to affect fitness, although this has yet to be established.

3 Objective 2: Laboratory studies: growth and development time

This objective tested the hypothesis that regeneration would affect growth (increase in

weight and cephalothorax width) and development time (intermolt interval) in S. ocreata.

Regeneration has been shown to negatively impact these measures in other animals (Goss, 1969;

Juanes and Smith, 1995; Ramsey, et al., 2001). Crustaceans that are regenerating appendages often exhibit reduced growth when compared to intact counterparts (Juanes and Smith, 1995;

Brock and Smith, 1998). Since regeneration is dependent on molting, its occurrence can be affected by the time during the intermolt cycle at which the appendage was lost. As regeneration requires a certain amount of growth it will not occur at the next molt if the appendage was lost too late in the cycle (Goss, 1969; Foelix, 1996). Also, regeneration itself can either decrease or extend the length of the molt cycle, depending on the species and the number of appendages being regenerated (Goss, 1969; Juanes and Smith, 1995; Mykles, 2001). The effects of regeneration on growth and development time have been well studied in crustaceans (Juanes and

Smith, 1995) but to my knowledge have not been addressed in arachnids.

Chapter II: The effects of autotomy and regeneration on prey capture rate, efficiency, and sensory detection:

Objectives 1 and 2: Prey capture rate and efficiency:

These two objectives tested the hypotheses that autotomy and regeneration of a leg would

have negative impacts on prey capture rate and efficiency. Several studies have shown that

decreased foraging abilities are a major cost of autotomy (Brock and Smith, 1998; Vollrath,

1990; Ramsey et al., 2001). Wolf spiders do not build webs, but are instead sit-and-wait

predators (Edgar, 1969, Ford, 1977, Cady, 1984) which have legs specifically adapted for

capturing and subduing prey (Rovner, 1980). A few laboratory studies have focused on the

4 effects of leg autotomy on capture and physical restraint of prey in wolf spiders (Amaya et al,

2001; Brueseke et al 2001). Amaya et al, (2001) found no overall affect of autotomy on the

number of attacks needed to capture prey in the wolf spiders Varicosa terricola and Schizocosa

ocreata, although larger S. ocreata with an autotomized leg were less effective at prey capture.

Brueseke et al (2001) found that autotomy did not affect number of attacks or subdue time in the

wolf spider Pardosa milvina, although autotomized spiders tended to capture smaller crickets

To my knowledge, no studies have addressed the effects of autotomy on foraging by

hunting spiders under field conditions, and none have used juveniles, whose foraging patterns

may differ from adults (Beck and Connor, 1992; Persons, 1999). Additionally, as juvenile

spiders are still capable of regenerating (unlike most adults), this process may also affect

foraging. Spiders require multiple molts to fully regenerate a leg, with the number required

being dependent on species. Partially regenerated legs are nonfunctional in some species and

may actually impede prey capture (Vollrath, 1990). However, the effects of regeneration on

foraging have only been tested for one species of web spinning spider (Vollrath, 1995).

Objective 3: Sensory Prey Detection:

This objective tested the hypothesis that autotomy and regeneration of a leg would impair

a spider’s vibration sensory abilities. Before physical capture can occur, spiders must detect their prey through visual and/or vibratory cues. Two types of vibration sensing structures occur on the spider’s legs. Sensory hairs called trichobothria are used to detect airborne vibrations,

while substrate borne vibrations are detected through lyriform organs made up of a series of

flexible slits. Vollrath, (1995) found that lyriform organs on the regenerated legs of the web

spinning spider Araneus diadematus are sometimes deformed. Likewise, juvenile Ixodid ticks

show reduction in number of slits in their Haller’s sensory organ (similar to lyriform organs of

5 spiders) after regeneration (Belozerov, 2001). Changes in the shapes and structures of these organs have the potential to affect their function. Weissmann (1987) demonstrated that autotomy of a leg impaired sensory detection in two species of web spinning spiders (cited in Vollrath,

1995) but no studies have addressed the effects of autotomy and regeneration on sensory detection of prey in a hunting spider.

6 References:

Amaya, C., Klawinski, P., Formanowicz, D. 2001. The effects of leg autotomy on

running speed and foraging ability in two species of wolf spider (Lycosidae).

American Midland Naturalist. 145: 201-205.

Arnold, E. 1984. Evolutionary aspects of tail shedding in lizards and their relatives.

Journal of Natural History. 18 (1): 127-169.

Bateman, P., Fleming, P. 2005. Direct and indirect costs of limb autotomy in field

crickets, Gryllus bimaculatus. Animal Behaviour. 69:151-159.

Beck, M., Connor, E. 1992. Factors affecting the reproductive success of the crab spider

Misumenoides formosipes- the covariance between juvenile and adult traits. Oecologia.

92 (2): 287-295.

Belozerov, V. 2001. Regeneration of limbs and sensory organs in Ixodid ticks (Acari

Ixodoidea, Ixodidae, and Argasidae). Russian Journal of Developmental Biology. 32 (3):

129-142.

Brautigam, S., Persons, M. 2003. The effect of limb loss on the courtship and mating

behavior of the wolf spider Pardosa milvina (Araneae: Lycosidae). Journal of

Insect Behavior. 16 (4): 571-587.

Brock, R.E, Smith L.D. 1998. Recovery of claw size and function following autotomy in

Cancer productus (Decapoda: Brachyura). Biological Bulletin. 194 (1): 53-62.

Brueseke, M., Rypstra, A., Walker, S., Persons, M. 2001. Leg autotomy in the wolf

spider Pardosa milvina: a common phenomenon with few apparent costs. American

Midland Naturalist. 146:153-160.

7 Cady, A. 1984. Microhabitat selection and locomotor activity of Schizocosa ocreata

(Walkenaer) (Araneae: Lycosidae). Journal of Arachnology. 11: 297-307.

Dodson, G., Beck, M. 1993. Precopulatory guarding of penultimate females by male

crab spiders, Misumenoides formosipes . Animal Behaviour. 46: 951-959.

Dodson, G., Schwaab, A. 2001. Body size, leg autotomy, and prior experience as factors

in the fighting success of male crab spiders, Misumenoides formosipes. Journal

of Insect Behavior. 14 (6): 841-855.

Edgar, W.1969. Prey and predators of the wolf spider Lycosa lugubris. Journal of

Zoology London. 159: 405-411.

Foelix, R. 1996. Biology of Spiders. pp.219-233. Oxford University Press. New York, NY.

Ford, M.1977. Metabolic cost of the predation strategy of the spider Pardosa amentata (Clerck)

(Lycosidae). Oecologia. 28: 333-340.

Goss, R. 1969. Principles of Regeneration. pp. 91-112. New York, Academic Press.

Johnson, S., Jakob, E.1999. Leg autotomy in a spider has minimal costs in competitive

ability and development. Animal Behaviour. 57: 957-965.

Juanes, F., Smith, L.D. 1995. The ecological consequences of limb damage and loss in

decapod crustaceans: a review and prospectus. Journal of Experimental Marine Biology

and Ecology. 193(1-2): 197-223.

Kunkel, J.G. 1981. Regeneration. In : The American Cockroach. (Ed. Bell, Adiyodi, ). pp. 427-

443. Chapman and Hall. New York, NY.

Mykles, D. 2001. Interaction between limb regeneration and molting in decapod

crustaceans. American Zoologist. 41: 399-406.

8 Persons, M. 1999. Hunger effects on foraging responses to perceptual cues in immature

and adult wolf spiders. Animal Behaviour. 57:81-88.

Ramsey, K., Kaiser, M.J., Richardson, C.A. 2001. Invest in arms: behavioral and

energetic costs of multiple autotomy in starfish (Asterias rubens). Behavioral Ecology

and Sociobiology. 50(4):360-365.

Roth, V., Roth, B.1984. A review of appendotomy in spiders and other arachnids.

Bulletin of the British Arachnological Society. 6 (4): 137-146.

Rovner, J. 1980. Morphological and ethological adaptations for prey capture in wolf

spiders (Araneae, Lycosidae). Journal of Arachnology. 8:201-215.

Simpson, M.1993. Reproduction in two species of arctic arachnids; Pardosa glacialis

and Alopecosa hirtipes. Canadian Journal of Zoology. 71: 451-457.

Spence, J., Zimmerman, M., Wojcicki.1996. Effects of food limitation and sexual

cannibalism on reproductive output of the nursery web spider Dolomedes triton

(Araneae: Pisauridae). Oikos. 75:373-382.

Stoks, R. 1999. Autotomy shapes the trade-off between seeking cover and foraging in

larval damselflies. Behavioral Ecology and Sociobiology. 47:70-75.

Taylor P., Jackson, R. 2003. Interacting effects of size and prior injury in jumping spider

conflicts. Animal Behaviour. 65: 787-794.

Uetz, G., McClintock, W., Miller, D., Smith, E., Cook, K. 1996. Limb regeneration and

subsequent asymmetry in a male secondary sexual character influences sexual

selection in wolf spiders. Behavioral Ecology and Sociobiology. 38:321-326.

Vollrath, F. 1990. Leg regeneration in web spiders and its implications for orb weaver

phylogeny. Bulletin of the British Arachnological Society. 8(6):177-184.

9

Vollrath, F. 1995. Lyriform organs on regenerated spider legs. Bulletin of the British

Arachnological Society. 10 (3):115-118.

Weissmann, M., Vollrath, F. 1998. The effect of leg loss on orb-spider growth.

Bulletin of the British Aarachnological Society. 11 (3) 92-94.

Wise, D. 1975. Food limitation of the spider Linyphia marginata: experimental field studies.

Ecology. 56: 637-646.

10 Chapter I

The impacts of leg loss and regeneration on fitness in a wolf spider as measured by body

condition, growth and development time.

(For submission to American Midland Naturalist or Canadian Journal of Zoology)

Kerri M. Wrinn and George W. Uetz

11 Abstract:

Autotomy (self-amputation) of appendages and subsequent regeneration are common to many animal groups. Numerous studies have shown that these processes can affect foraging abilities, growth and development time in animals such as reptiles, echinoderms, and crustaceans. However, to our knowledge no one has studied these effects in arachnids. We addressed the effects of autotomy and regeneration on body condition, growth (size and weight) and development time (molt interval) for Schizocosa ocreata wolf spiders in the laboratory and field. Frequency of autotomy in the field was high for this species (12-19% of individuals).

Body size, weight and condition were significantly lower in individuals captured in the field with missing/regenerating legs. Spiders in the lab were induced to autotomize one or both forelegs in order to test the effects of regeneration on size, weight and molt interval. Spiders reared in the laboratory had longer molt intervals when regenerating a foreleg, but were similar in size to intact individuals. In contrast, spiders captured in the field as juveniles and induced to autotomize a foreleg in the lab had similar molt intervals to intact spiders, but were smaller and weighed less. Spiders regenerating two legs had reduced molt intervals, were smaller, and weighed less than individuals that were intact or regenerating one leg. The overall negative impacts of autotomy and regeneration on body measurements, growth and development time in juvenile S. ocreata are important because body size and condition have the potential to affect the spider’s fitness as an adult.

12 Introduction:

The ability to autotomize (self-amputate) appendages occurs in many animals, including

echinoderms (Ramsay et al; 2001), vertebrates (Arnold, 1984), crustaceans (Juanes and Smith,

1995), and arachnids (Roth and Roth, 1984). Autotomy provides direct fitness benefits,

including avoidance of predation (Formanowicz, 1990; Klawinski and Formanowitz, 1994;

Punzo, 1997) and poisoning by venomous prey (Eisner and Camazine, 1983). However,

autotomy may also result in decreased competitive abilities (Mariappan et al., 2000; Taylor and

Jackson, 2003; Dodson and Beck, 1993), reduced speed (Formanowicz, 1990, Bateman and

Fleming, 2005) and reduced foraging abilities (Vollrath, 1990; Brock and Smith, 1998; Stoks,

1999; Ramsey et al., 2001).

Many animals that autotomize appendages can also regenerate, which may be beneficial in negating some of the costs of missing a leg. However, regeneration also has its own set of potential fitness costs. These costs include increased developmental energy requirements, e.g. reduced growth and longer duration between growth periods (Goss, 1969; Vitt et al., 1977;

Juanes and Smith, 1995; Ramsey, et al., 2001), reduced function, which may affect locomotion and foraging (Brock and Smith, 1998); and behavioral impacts on competition (Brock and Smith,

1998) and mating (Uetz et al., 1996).

Arthropods molt (shed their exoskeleton) several times as they grow. As regeneration is linked to molting, arthropods that molt frequently can also frequently regenerate, making them an ideal group for studying the process. Extensive studies have addressed both autotomy and regeneration in crustaceans (see Juanes and Smith 1995 for a review), and insects such as cockroaches (Bell and Adiyodi, 1981), damselflies (Stoks, 1999), and crickets (Bateman and

Fleming, 2005). However, much less is known about these processes in arachnids. Autotomy

13 and regeneration are potentially important for spiders, as they use their legs for multiple

functions including locomotion, prey capture, sensory perception, prey detection, and

communication. Additionally, autotomy is a frequent occurrence in many spider taxa, as field

populations normally show 5-20% of spiders with missing legs (Foelix, 1996). Many (but not

all) spiders that autotomize can also regenerate lost legs (Vollrath, 1990, Johnson and Jakob,

1999). Regeneration appears to be an ancestral trait (Goss, 1969). Therefore, the loss of the

ability to regenerate by some spider groups over evolutionary time might indicate that costs of

this process could differ between groups. In order for regeneration to be maintained, its costs

must be outweighed by its benefits over the autotomized condition.

One of the major costs of autotomy is decreased foraging ability (Vollrath, 1990; Brock

and Smith, 1998; Stoks, 1999; Ramsey et al., 2001). A few laboratory studies have addressed

effects of autotomy on foraging in adult wolf spiders (Amaya et al, 2001; Brueseke et al 2001).

However, no studies have addressed the effects of autotomy on foraging in the field, or have

used juveniles whose foraging patterns may differ from adults (Beck and Connor, 1992; Persons,

1999). Additionally, juvenile spiders are still capable of regenerating (unlike most adults).

Regenerating a leg may affect foraging in spiders, but this has only been tested for one species of

web spinning spider (Vollrath, 1995).

Reduced foraging as a juvenile has the potential to affect size and weight as an adult

(Beck and Connor, 1992; Uetz et al, 1996). Additionally, size, weight, and condition have all

been correlated with fitness in spiders (Wise, 1975; Beck and Connor, 1992; Simpson, 1995;

Spence et al., 1996). Therefore, decreased foraging due to autotomy and regeneration of

appendages has the potential to affect the size, weight, condition and ultimately fitness of

spiders, although this has yet to be studied.

14 Regeneration can also lead to developmental costs (Goss, 1969; Juanes and Smith, 1995;

Ramsey, et al., 2001). In crustaceans, individuals regenerating appendages have been shown to

exhibit reduced growth, as they are often smaller after molting than intact counterparts (Juanes

and Smith, 1995; Brock and Smith, 1998). Additionally, regeneration of new appendages is

linked to molting and can be affected by the time during the intermolt cycle at which the

appendage was lost. If the appendage is lost too late in the cycle, there will not be enough time

for growth and it will not regenerate until a later molt (Goss, 1969; Foelix, 1996). Additionally,

regeneration itself can either decrease or extend the length of the molt cycle (Goss, 1969; Juanes

and Smith, 1995; Mykles, 2001). The number of appendages being regenerated can also affect

the molt cycle (Juanes and Smith, 1995; Goss, 1969). The effects of regeneration on growth and

development are well studied in crustaceans (Juanes and Smith, 1995) but have yet to be

addressed in any studies for spiders.

We used the well-studied wolf spider Schizocosa ocreata (Hentz) for this study. These spiders are commonly found in the leaf litter of Eastern deciduous forests (Cady, 1984). S. ocreata exhibit a high occurrence of autotomy in the field (Uetz et al, 1996). Additionally, it is easy to induce autotomy in these spiders in the lab (personal observation). Finally, these spiders are ideal study animals because they are abundant, easy to maintain, and have short growth periods and frequent molts.

The objectives of this study were 1) to observe the frequency of autotomy in S. ocreata and compare size (cephalothorax width), weight and body condition between autotomized/regenerating and intact spiders collected in the field, and 2) to compare the growth

(size and weight) and development time (molt interval) of spiders regenerating one or both

15 forelegs with intact controls under laboratory conditions where food was held constant between groups.

General methods:

Spiders were collected by hand from forest floor leaf litter at the Cincinnati Nature

Center, Rowe Woods, in Clermont County, OH. Spring collecting for both juveniles and adults occurred from April-June (2003, 2004) and fall collecting for juveniles only, occurred from

September-October (2003, 2004). All spiders collected from the field (juveniles and adults) were housed individually in the laboratory in circular opaque deli dishes (9 cm diameter x 6 cm

high) with clear lids. This visually isolated the spiders from each other but still allowed them to receive light from above. All spiders were maintained under controlled laboratory conditions including: temperature (21-24 º), stable humidity, and an 11:13 hour dark: light photoperiod.

The spiders were fed 10-day old crickets (Acheta domesticus) twice a week and provided water

ad libitum through a dental wick connected to a reservoir in a container below. In the lab, adult

females were checked daily for egg production and hatching. Upon hatching, spiderlings were

left with the mother for seven to ten days (spiderlings climb onto the mother’s abdomen and stay

attached for this time before becoming independent). After this time they were dispersed, placed

in separate 120 ml specimen cups with damp dental wicks for water, and were fed Collembola

(mixed species) or fruit flies (Drosophila melanogaster) twice a week. After reaching their

fourth instar, spiderlings were placed in deli dish containers (as above) and fed crickets (Acheta

domesticus) twice weekly.

16 Objective 1: Frequency of leg loss in the field and its impact on size, weight and body condition

If losing and regenerating a leg affects fitness due to increased developmental costs or decreased foraging ability, intact spiders would be expected to be larger and have higher body condition than spiders with missing or regenerated legs.

Methods:

Juvenile S. ocreata collected from the field were examined for leg loss and measured in the lab [2003 spring (n=693); fall (n=450); 2004 spring (n=938); fall (n=656)]. The occurrence, side (left or right) and position (legs I-IV) of any leg loss or regeneration was recorded for all spiders collected.

Size (cephalothorax width) and weight, both indicators of past feeding history, have been

related to fitness in spiders (Wise, 1975; Beck and Connor, 1992; Simpson, 1995; Spence et al.,

1996) and insects (Honek, 1993). However, body condition indices that account for both size

and weight together have begun to gain popularity as measures of fitness in spiders (Jakob et al,

1996; Danielson-Francois et al., 2002; Uetz et al., 2002) and other arthropods (Glazier, 2000;

Heg and Rasa, 2004). Therefore in addition to comparing weight and size between groups, I

used a body condition index as a measure of fitness in this study.

In order to determine body condition, all spiders with missing or regenerating legs and

equal numbers of randomly selected intact spiders from the same collection were weighed to

nearest milligram within 48 hours of collection (prior to being fed), and digitally photographed

using a Pixera 1.2 mega-pixel digital camera through a Wild M5 microscope. For each picture,

the spider was placed in a small Petri dish and allowed to settle so that the cephalothorax was flat against the substratum and the spider’s legs were stretched out in a resting position. From these

17 pictures, measurements were taken of cephalothorax width (distance across the widest point of the spider’s carapace) using Image-Tool version 2.00 (University of Texas Health Sciences

Center, San Antonio). To create a body condition index (BCI), log transformed weight (mg) was regressed against log transformed cephalothorax width (mm) and the residuals of these regressions were used for analyses (as in Jakob et al., 1996).

Data Analysis:

A 2x2 Chi square contingency table was used to test whether the number of spiders missing legs each year (2003 and 2004) was independent of the year of collection. Chi square goodness-of-fit tests were used to test for difference in leg loss by side and by position.

Three-factor nested ANOVAs with the factors Injury status (autotomy or intact), Year

(2003, 2004), and Season [Year] (fall or spring), were used to test for differences in the response variables weight, cephalothorax width, and body condition. Further three-factor nested

ANOVAs with the factors Injury status (autotomized or intact), Season (fall or spring) and Week

(1 or 2) [Season] were used to compare the same three response variables between fall 2003 and spring 2004 as these represent two age classes of the same cohort. All data analysis was done using the statistics program JMP 4.0.2 (SAS Institute).

Results:

Frequency of leg loss:

Frequency of leg loss within the population was considerable for both spring and fall of

2003 and 2004; ranging from 12.96% to 19.3% (Table 1.1). The proportion of spiders collected with missing legs for 2004 did not differ between seasons (spring and fall) (X²1=0.113; p<0.75).

However, in 2003 a significantly larger proportion of spiders were missing legs in fall than in

2 spring field collections (X 1 = 9.56; p<0.005).

18 There was no difference between collecting periods for the side at which leg loss

2 occurred (R and L) so data were pooled for all periods (X 3=2.24, p<0.50), and a Chi square goodness of fit test of the pooled data showed no overall difference in leg loss between sides

(X2=0.86, p<0.5). Likewise, a 4 X 4 Chi square contingency table showed no significant difference in position of leg loss between any of the collecting periods (Spring and Fall 2003,

2 2004; X 9=4.681, p<0.9) so these data were also pooled for further analysis. A Chi square goodness of fit test showed a significant difference in leg loss by position for the pooled data, with legs being lost most often at the first and fourth positions (Figure 1.1).

Size, weight and body condition:

Overall, autotomized spiders were significantly smaller, weighed significantly less, and had significantly lower body condition than their intact counterparts, as shown by a three- factor nested ANOVA with injury status, year, and season [year] as factors (Table 1.2, Figures 1.2 A-

C). It should be noted that in laboratory tests the loss of the weight of an autotomized leg was not enough to measurably change a spider’s overall weight (personal observation). Therefore any weight differences between intact and autotomized spiders must be due to other factors.

In addition to being affected by autotomy, spider cephalothorax width, weight, and body condition differed significantly by season, and weight and body condition differed by year (Table

1.2, Figure 1.2 A-C). The interaction terms Season [year] x Injury status, and Year x Injury status were significant in several cases as well (Table 1.2).

One cohort of spiders which spanned two seasons (fall 2003 and spring 2004) was examined in order to tease apart some of the effects of season and year on cephalothorax width, weight and body condition in a biologically significant way. The life cycle of S. ocreata creates two different generations during each calendar year, with the fall population being the offspring

19 of the previous spring population. Spiders are born in the summer, overwinter as juveniles, and mature in the following spring. This created some overlap, as two separate samples of spiders from the same generation were measured in fall of 2003 and in spring 2004 respectively.

There was no difference in frequency of leg loss between spiders from the same cohort in fall

2 and the following spring (X 1=2.91, p<0.10). Overall, spiders captured with missing or regenerating legs were smaller, weighed less, and had lower body condition than intact spiders

(Table 1.3A,B). Spiders were larger and weighed more in the spring, as was expected from normal growth and development. There was also a significant injury status x season interaction for weight, with a larger weight difference between autotomized and intact spiders in spring than in fall (Table 1.3A,B).

There were some differences in body condition depending on the part of the season

(early vs. middle or late) that the spiders were captured in. Autotomized spiders measured in mid to late fall 2003 had significantly lower body condition than intact spiders (Table 1.3B). By early spring 2004, all spiders showed a marked decrease in condition from the previous fall and differences between groups were no longer significant. Both groups increased in condition by mid spring, but the increase was greater for intact spiders, leaving autotomized spiders in significantly lower condition (Table 1.3B, Figure 1.3A). Because of changes in the body condition of spiders from early to mid-spring 2004, early and middle period weeks were examined for fall 2004 as well. As in the previous spring, both autotomized and intact fall 2004 spiders were of approximately equal body condition early in the season, but became significantly different towards the middle of the season (Figure 1.3B).

20 Discussion:

Frequency of leg loss:

The frequency of leg loss in S. ocreata fell within the upper part of the range of values

reported in the literature for other species (5-20%; Foelix, 1996). In a previous study (Uetz et al,

1996), field-collected immature S. ocreata showed a comparable value (15%). As spiders that

have molted more than two times after autotomy are indistinguishable from intact spiders

(personal observation), these percentages may actually be an underestimate of the true amount of

autotomy occurring in the field. Regardless, this high percentage of leg loss indicates the

importance of autotomy for this species. There were significant differences in the percentage of

leg loss between spring and fall of 2004, but not 2003. This suggests that if predation is the

major cause of leg loss in the field, predation pressures may differ by season and/or year. The

differences between spring and fall of 2003 may have been due to lower predation occurring

over the winter, as both the spiders themselves and many of their potential predators were less

active during that time.

S. ocreata were found most frequently with legs missing at the first and fourth positions.

These results are similar to those found in the Uetz et al (1996) study. S. ocreata may lose legs

at the first or fourth position more often if most incidents of autotomy are due to predation or

cannibalism attempts. The first legs are often raised defensively if the spider is threatened

(Aspey, 1976) potentially making them more visible to an attacker. Alternatively the fourth pair

of legs in addition to being the longest pair may be easily accessible if the spider turns to flee

from an attacker. The second and third legs are in a more protected position and are slightly

shorter than the other legs, possibly making them less vulnerable to being grasped. This appears to be the case for some species of crabs where autotomy occurs most frequently in the two

21 longest most exposed limbs (Spivak and Politis, 1989). However, a study by Brueseke et al,

2001 demonstrated that Pardosa milvina, another wolf spider native to the area autotomized all

legs with equal frequency. This may indicate differences in morphology or predation pressure

between the two species.

Size, weight and body condition:

Size, weight and body condition were lower in field caught spiders that were missing or

regenerating legs. This was possibly due to reduced ability to forage, as foraging success has the potential to affect all three body measurements. Weight varies with food and water intake

(Anderson, 1974) making weight change an indicator of recent feeding history. Cephalothorax

width is fixed between molts, but can be an indicator of past foraging. Uetz et al. (1996) showed

that when S. ocreata were underfed as juveniles, they had smaller cephalothorax widths as adults. Likewise, Beck and Connor (1994) found that size differences in juveniles of the crab spider Misumenoides formosipes in the field were related to varying prey capture success. Body

condition indices (BCI: the residuals of a weight x cephalothorax width regression) measure both

past and recent foraging history. In a laboratory study of P. milvina, Jakob et al. (1996) showed

that feeding level (high or low) over a period of 8 weeks affected body condition in juvenile

spiders. Additionally, Uetz et al. (1996) showed that S. ocreata raised for their entire lives under

different feeding levels (high or low) differed in body condition as adults.

Laboratory studies have shown no significant effects of autotomy on foraging in three

species of wolf spiders (Amaya et al, 2001; Brueseke et al, 2001). However, no studies have

addressed the effects of autotomy on foraging of spiders in the field. Autotomy/regeneration has

been shown to affect the foraging abilities of other arthropods through mark and recapture

studies and under simulated field conditions (Juanes and Smith, 1995; Stoks, 1999). Therefore,

22 it is conceivable that it may have an effect on spiders as well. We were unable to address differences in size and body condition between spiders with missing and regenerating legs in the field as it was necessary to combine them to maintain a sufficient sample size for each collecting date. Future studies which are able to separate these two categories may be useful.

It is possible that rather than autotomy affecting foraging, spiders in the field that were smaller and in lower condition in the first place were more prone to predation and leg loss.

However, the pattern of changes in body condition of autotomized spiders across the season would indicate otherwise. Overall, body condition did not differ between intact and autotomized spiders early in the season (for spring and fall), but towards the middle of each season intact spiders had significantly higher condition than spiders missing/regenerating a leg. This indicates that the negative effects of autotomy/regeneration may be cumulative over time.

Regeneration of limbs may create energetic costs that affect growth and development time. Autotomized/regenerating spiders were on average somewhat smaller than intact spiders in the field. This could have been due to decreased foraging as suggested above, increased energetic costs, or a combination of both. An interaction between feeding and energetics was shown for the starfish Asterias rubens as level of diet affected regeneration time (Ramsay 2001).

Likewise, in S. ocreata a growth cost leading to reduction in size increase after a molt could hamper a spider’s foraging as well by constraining them to attack smaller prey. This makes it difficult to separate the effects of the processes at work on spiders in the field. Further experiments which cross leg autotomy with a high/low feeding treatment are needed to determine what type of interactions between growth and foraging may be occurring in the field.

In addition to autotomy/regeneration affecting size weight and body condition, there were also some yearly differences in these variables as well. All spiders in spring 2003 (autotomized

23 and intact) were on average one mm smaller than spiders in 2004, and in lower condition.

Conversely, the spiders in fall 2003 were almost one mm larger than fall 2004 spiders. This

indicates that there may be some yearly differences in field conditions that have the potential to

affect spider size, such as temperature and prey availability (Schaefer, 1987). Furthermore, it is

possible that prey availability changes over time within the season, as evidenced from body

condition increases over collecting weeks for both autotomized and intact spiders in spring and

for intact spiders in fall.

Objective 2: The growth/developmental costs of leg regeneration.

If regeneration affects growth/development, autotomized spiders would be expected to

have longer intervals between molts and/or be smaller in size (cephalothorax width) and/or gain less weight during a molt than intact spiders.

Methods:

Weight and molt interval data were compared for lab-reared spiders in fall 2003 (n=114).

These individuals were the offspring of females that had been collected as adults in the field the previous spring (Objective 1). To control for possible egg sac effects, spiderlings were taken from 11 egg sacs, each produced by a different individual. Owing to differences in initial egg

sac size and survivorship, approximately 5-6 spiderlings were used from each egg sac. Spiders

were checked twice a week for molts until they reached their fourth molt, then were checked

daily thereafter. When spiders reached their fourth molt, approximately half of them (n=60)

were chosen at random and induced to autotomize an arbitrarily selected foreleg, by restraining

the femur with a pair of forceps. The rest of the spiders (n=54) served as a control group and

were sham manipulated (touched with forceps but left intact). The number of days between the

24 4th and 5th molts and between the 5th and 6th molts were compared between control and

manipulated spiders. Additionally, spiders were weighed within 48 hours of each molt (prior to

being fed). Weight gains between molts were compared between manipulated and control

spiders.

In spiders, cephalothorax width is fixed between molts, unlike weight, which can change with feeding and water intake (Anderson, 1974). Therefore, an experiment to measure this additional factor was conducted in fall 2004 using spiders captured in the field as juveniles in the

3rd to 4th instar and then raised in the lab. Within two molts after autotomy, most individuals of

S. ocreata regenerate a normal sized leg (personal observation). Therefore, in order to compare

individuals of the same age (total n=180) under different conditions, 60 individuals were induced

to autotomize immediately upon reaching the lab, 60 were induced to autotomize after molting

once in the lab, and 60 were left intact. Comparisons between groups were made for the molt

interval (days) needed for the first group to regenerate a full sized leg from a partial one, the

second group to partially regenerate a leg, and the third group (intact) to undergo normal growth.

Weights and cephalothorax widths were compared for a subset of these spiders, n=30/group for a

total of n= 90 (see methods in objective 1 for details on weighing and measuring).

Several studies have shown that in crabs the effects of multiple leg loss on molt intervals

and size gain are often different from those caused by a single autotomy (Goss, 1969, Juanes and

Smith, 1995). Therefore using another subset of spiders (n=50) I compared molt intervals,

weight gains and cephalothorax width gains as above between intact spiders (n=17), and those

from which I had previously autotomized one foreleg (n=16) or both forelegs (n=17).

25 Data Analysis :

For Fall 2003, two sample t-tests were run to compare both molt duration and weight gain

between control and manipulated spiders (as there were only two groups; autotomy or intact).

For Fall 2004, separate one-way ANOVA’s were used to compare weight gain, molt interval,

and cephalothorax width gain between the three groups for the first experiment (single autotomy)

and for the second experiment (multiple autotomy). Data analysis for all experiments was done

using the statistics program JMP 4.0.2 (SAS Institute).

Results

Regeneration has the potential to affect spider size and weight through two mechanisms:

1) by decreasing foraging abilities and 2) by causing a reallocation of energy to the regenerating

part, thus decreasing general growth and development of the individual. Laboratory studies were

performed in an attempt to separate these two mechanisms. As spiders were randomly selected

from the population it was assumed that all groups of spiders began the study in equal body

condition, and they were fed equally to control foraging. In fall 2004, body condition (a measure

of foraging success) did not differ between groups at the end of the experiment, indicating that

all changes in size/weight were due to changes in growth and development (ANOVA:

F2,73=0.2540; p=0.7764) .

Autotomy of a single leg:

Spiders regenerating a single leg in the lab reared 2003 study had initial molt intervals that were significantly longer than control spiders (t112=2.266; p=0.026; Table 1.5). However,

there was no significant difference in molt interval between groups for the second molt after

autotomy (t104= -1.042; p=0.300; Table 1.5). Additionally there was no difference in weight

26 change between the groups for either molt interval (t108=0.283; p=0.7778) and (t76=0.583;

p=0.5179) respectively.

For lab-maintained spiders in the 2004 study there was a marginally significant difference

in cephalothorax width between groups (ANOVA: F2,77=3.0197, p=0.0548; Figure 1.4A). Tukey post hoc tests showed that spiders that were regenerating a second time had the greatest cephalothorax width (Figure 1.4A). There was also a significant difference in weight between

groups (ANOVA: F2,133=7.1403, p=0.0011; Figure 1.4B). Spiders which had just undergone a

second molt post-autotomy weighed the most, followed by unmanipulated intact spiders and spiders having undergone the first post-autotomy molt (Figure 1.4B). There was no difference between groups in the number of days between molts (ANOVA: F2,83=1.0499, p=0.3546).

Autotomy of two legs:

There was a significant difference in molt interval between groups that lost one or two

legs (ANOVA: F2,47=7.2824, p=0.0018; Figure 1.5A). Spiders which had undergone autotomy

of a single leg weighed significantly more than spiders which had undergone autotomy of two legs after the first regeneration (F1,29=5.5457, p=0.0255; Figure 1.5B). There was also a

significant difference between groups in cephalothorax width (ANOVA: F2,45=5.8329, p=0.0056;

Figure 1.5C) Post hoc tests showed that intact spiders were the largest but took the longest to

molt, while spiders that were regenerating two legs were the smallest but molted the fastest

(Figures 1.5A-C). There was also a significant difference in the molt interval between groups for

the second period of regeneration (ANOVA: F2,47= 5.6461, p=0.0063), with intact spiders once

again having the longest intervals (Figure 1.4). However, for this period neither weight nor cephalothorax width were significantly different between groups (ANOVA: (F1,30= 1.6282,

p=0.2118) and (F2,43= 0.3504, p=0.7064) respectively (Figures 1.5B,C).

27 Discussion:

Regeneration of a single leg:

Lab-reared spiders in 2003 that were regenerating a leg took an average of 3.7 days

(10.8%) longer to molt than intact spiders, although there was no difference in weight between the groups after molting. However, this difference was true only for the initial molt after autotomy. The regenerating spiders were able to compensate with decreased molt intervals in the following molt, so there was no significant difference between groups in the total number of days for both molt intervals. An increase in molt interval due to regeneration of an appendage has been shown for both crustaceans and insects. Most animals have a critical point during the molt cycle after which they can no longer regenerate (Goss, 1969). The lobster Homarus americanus will extend its molt interval if the autotomy occurs close to this critical point, but will shorten the molt if autotomy occurs early in the cycle (Waddy, et al, 1995). In contrast, regeneration of a leg in the cockroach Blattella germanica always results in a delayed molt cycle

(Bell and Adiyondi, 1981).

In the 2004 experiment, field caught S. ocreata which underwent autotomy in the lab were significantly smaller (0.1 mm) in size and weighed significantly less (4.3 mg) after molting when compared to intact spiders. Similar effects have been demonstrated in crustaceans, as loss of a limb can reduce size increase at molting in the lobster H. americanus (Waddy et al, 1995).

There was no difference in molt intervals between regenerating and intact spiders. Additionally, the significant difference in size was only true for the initial molt after autotomy. After the second post- autotomy molt, regenerating spiders actually seemed to overcompensate by increasing size and weight so that they were slightly larger and heavier than intact spiders.

28 Regeneration of a single leg in S. ocreata led to either an increased molt interval (2003)

or decreased size and weight after molting (2004), but not both at once. These results differ from

those in a study of autotomy and regeneration of caudal lamellae in damselfly larvae (Stoks,

2001). In that study, damselfly larvae which had undergone autotomy and were subsequently

regenerating showed both reduced growth (size and mass) and development (molt interval).

However, other factors affecting growth and development time including feeding level and presence/absence of a predator were present as well (Stoks, 2001). In the current study, which was only addressing autotomy, a tradeoff was obviously occurring between size and development time; the spiders either took longer to molt in order to reach full size or developed at the normal rate but at a cost of size. Field caught spiders were subjected to autotomy and treated the same as lab reared spiders upon reaching the lab, yet were still affected differently by regeneration. This may indicate that differences in early rearing history of the groups contributed to the differential effects of regeneration between the years.

Nutrition is one factor that affects both growth and development in spiders (Foelix,

1996). Spiders in the field are often limited by quantity of food, (Anderson, 1979, Kreiter and

Wise, 2001), but have the potential to capture a wide variety of prey types. Conversely, lab-

reared spiders were not limited by food quantity, but were fed on a fairly monotypic diet (which

might have affected their growth) (Uetz et al, 1992). Light levels and temperature are another set

of factors that affect the molt cycle in several species of spiders (Nentwig, 1987). S. ocreata in

the field were subjected to naturally varying light levels and temperatures, while both of these

factors remained constant throughout the lives of the lab reared spiders. Finally, spiders in the

lab were maintained individually after their initial dispersal, while field caught spiders had

opportunities to interact with other conspecifics and potential predators. It is unknown however,

29 whether competition or predation rate affects the molt cycle. Although previous foraging

history, differences in light levels and temperature, and presence or absence of competition could

conceivably affect the balance between the developmental and growth costs of regeneration,

further studies in these areas are needed before any conclusions are made.

Regeneration of two legs:

Regeneration of two legs had a greater effect on development time in S. ocreata than regeneration of a single leg. In some cases molt interval may be lengthened due to multiple autotomy; e.g. in the cockroach B. germanica (Bell and Adiyodi, 1981). However, in S. ocreata autotomy of two legs significantly decreased the molt interval for both of the molts following autotomy. Decreases in regeneration time and/or molt interval due to multiple autotomy have been demonstrated for several other animals, including echinoderms, crustaceans, and chilopods.

However, the number of regenerating appendages necessary to decrease molt intervals differs between taxa. Zeleny (1905) showed that the brittlestar Ophioglypha lacertosa decreased

regeneration time when more arms were missing. Skinner and Graham (1972) determined that

the molt cycle of the land crab Gecarcinus lateralis was shortened by loss of 5 or more legs.

Zeleny (1905) showed a decreased molt interval in two species of marine crabs (Gelasimus

pugilator and Alpheus denatipes) which were regenerating two limbs. Finally, Cameron (1927)

demonstrated that the centipede Scutigera forceps decreased molt interval only when all of its

legs had been autotomized.

In addition to decreased molt interval, spiders regenerating two legs weighed

significantly less than those which were regenerating only one after the first but not the second molt post autotomy. Intact spiders were not weighed directly after molting so they were not compared to the autotomized spiders in this experiment. However, cephalothorax width does not

30 change between molts and could therefore be measured later in intact spiders as well. Spiders

which were regenerating two legs had cephalothorax widths that were significantly smaller than

intact spiders but not different from spiders which were only regenerating one leg, although this

difference was only significant for the first molt after autotomy. Similar results have been found for the crab Cancer productus which has significantly reduced body size only when regenerating

two or more limbs (Brock and Smith, 1998).

Conclusions:

Overall, autotomy occurred frequently for S. ocreata in the field, and spiders missing or

regenerating legs were smaller, weighed less and were in poorer condition than intact

individuals. Regeneration also had negative effects on growth and development time

which seemed limited to the first instar after autotomy. However, these effects could be

particularly detrimental to spiders which underwent autotomy in a penultimate stage and could therefore only molt one more time. Ultimately, the effects of autotomy and regeneration, despite their obvious survival benefits, can be costly to spiders. Decreases in condition, size, and weight as juveniles have the potential to affect reproductive success as adults (Wise, 1975; Beck and

Connor, 1992; Simpson, 1995; Spence et al., 1996, Uetz et al, 1996).

31 References:

Amaya, C., Klawinski, P, Formanowicz, D. 2001. The effects of leg autotomy on

running speed and foraging ability in two species of wolf spider (Lycosidae).

American Midland Naturalist. 145: 201-205.

Anderson, J. 1974. Responses to starvation in the spiders Lycosa lenta: Hentz and

Filistata hibernalis. Ecology. 55: 576-585.

Arnold, E. 1984. Evolutionary aspects of tail shedding in lizards and their relatives.

Journal of Natural History. 18 (1): 127-169.

Aspey, W. 1976. Response strategies of adult male Schizocosa crassipes (Araneae: Lycosidae)

during agonistic interactions. Psyche. 83 (1), 94-105.

Bateman, P., Fleming, P. 2005. Direct and indirect costs of limb autotomy in field

Crickets, Gryllus bimaculatus. Animal Behaviour. 69:151-159.

Beck, M., Connor, E. 1992. Factors affecting the reproductive success of the crab spider

Misumenoides-Formosipes- the covariance between juvenile and adult traits. Oecologia.

92 (2): 287-295.

Brock, R.E., Smith LD. 1998. Recovery of claw size and function following autotomy in

Cancer productus (Decapoda: Brachyura). Biological Bulletin. 194 (1): 53-62.

Brueseke, M., Rypstra, A., Walker, S., Persons, M. 2001. Leg autotomy in the wolf

spider Pardosa milvina: a common phenomenon with few apparent costs. American

Midland Naturalist. 146:153-160.

Cady, A.1984. Microhabitat selection and locomotor activity of Schizocosa ocreata

(Walkenaer) (Araneae: Lycosidae). Journal of Arachnology. 11: 297-307.

32 Cameron, J. 1927. Regeneration in Scutigera forceps. Journal of Experimental Zoology. 46:

169-179.

Danielson-Francois, A., Fetterer, C., Smallwood, P. 2002. Body condition and mate

choice in Tetragnatha elongata (Araneae, Tetragnathidae). Journal of Arachnology. 30:

20-30.

Dodson, G., Beck, M. 1993. Precopulatory guarding of penultimate females by male

crab spiders, Misumenoides formosipes . Animal Behaviour. 46: 951-959.

Eisner, T., Camazine, S. 1983. Spider leg autotomy induced by prey venom injections:

An adaptive response to pain. Proceedings of the National Academy of Sciences.

80: 3382-3385.

Foelix, R. 1996. Biology of Spiders. pp. 213-233. Oxford University Press. New York, NY.

Formanowicz, D.1990. The antipredator efficacy of spider leg autotomy. Animal

Behaviour. 40:400-401.

Glazier, D. 2000. Is fatter fitter? Body storage and reproduction in ten populations of the

Freshwater amphipod Gammarus minus. Oecologia. 122: 335-345.

Goss, R. 1969. Principles of Regeneration. pp. 91-112. New York, Academic Press.

Heg, D., Rasa, O. 2004. Effects of parental body condition and size on reproductive

success in a tenebrionid beetle with biparental care. Ecological Entomology. 29 (4): 410-

419.

Honek, A. 1993. Intraspecific variation in body size and fecundity in insects: a general

relationship. OIKOS. 66: 483-492.

Jakob, E., Marshall, S., Uetz, G. 1996. Estimating fitness: a comparison of body

condition indices. OIKOS . 77: 61-67.

33 Johnson, S., Jakob, E.1999. Leg autotomy in a spider has minimal costs in competitive

ability and development. Animal Behaviour. 57: 957-965.

Juanes, F., Smith, L.D. 1995. The ecological consequences of limb damage and loss in

decapod crustaceans: a review and prospectus. Journal of Experimental Marine Biology

and Ecology. 193(1-2): 197-223.

Klawinski, P., Formanowitz, D. 1995. Ontogenetic change in survival value of leg

autotomy in a wolf spider, Gladicosa pulchra (Keyserling) (Araneae: Lycosidae), during

scorpion attacks. Canadian Journal of Zoology. 72: 2133-2157.

Kreiter, N., Wise, D. 2001. Prey availability limits fecundity and influences the movement

pattern of female fishing spiders. Oecologia. 127: 417-424.

Kunkel, J.G. 1981. Regeneration. In : The American Cockroach. (Ed. Bell, Adiyodi, ). pp. 427-

443. Chapman and Hall. New York, NY.

Mariappan, P., Balsundaram, C., Schmitz, B. 2000. Decapod crustacean chelipeds: an

overview. Journal of Biosciences. 25 (3): 301-313.

Mykles, D. 2001. Interaction between limb regeneration and molting in decapod

Crustaceans. American Zoologist. 41: 399-406.

Persons, M. 1999. Hunger effects on foraging responses to perceptual cues in immature

and adult wolf spiders. Animal Behaviour. 57:81-88.

Punzo, F. 1997. Leg autotomy and avoidance behaviour in response to a predator in the

wolf spider, Schizocosa avida (Araneae: Lycosidae). Journal of Arachnology. 25: 202-

205.

34 Ramsey, K., Kaiser, M.J., Richardson, CA. 2001. Invest in arms: behavioral and

energetic costs of multiple autotomy in starfish (Asterias rubens). Behavioral Ecology

and Sociobiology. 50(4):360-365.

Roth, V., Roth, B.1984. A review of appendotomy in spiders and other arachnids.

Bulletin of the British Arachnological Society. 6 (4): 137-146.

Schaefer, M. 1987. Life cycles and diapause. In: Ecophysiology of Spiders. (Ed. Nentwig, W.)

pp. 331-347. Springer-Verlag. Berlin, Germany.

Simpson, M.1993. Reproduction in two species of arctic arachnids; Pardosa glacialis

and Alopecosa hirtipes. Canadian Journal of Zoology. 71: 451-457.

Skinner, D.; Graham, D. 1972. Loss of limbs as a stimulus to ecdysis in Brachyura (true

crabs). Biological Bulletin. 143: 222-233.

Spence, J., Zimmerman, M., Wojcicki.1996. Effects of food limitation and sexual

cannibalism on reproductive output of the nursery web spider Dolomedes triton

(Araneae: Pisauridae). Oikos. 75:373-382.

Spivak, E., Politis, M.1989. High incidence of limb autotomy in a crag population from a

coastal lagoon in the province of Buenos Aires, Argentina. Canadian Journal of

Zoology. 67: 1976-1985.

Stoks, R. 1999. Autotomy shapes the trade-off between seeking cover and foraging in

larval damselflies. Behavioral Ecology and Sociobiology. 47:70-75.

Taylor, P., Jackson, R. 2003. Interacting effects of size and prior injury in jumping spider

conflicts. Animal Behaviour. 65: 787-794.

Uetz, G, Bischoff, J., Raver, J. 1992. Survivorship of wolf spiders (lycosidae) reared on

different diets. Journal of Arachnology. 20: 207-211.

35 Uetz, G., McClintock, W., Miller, D., Smith, E., Cook, K. 1996. Limb regeneration and

subsequent asymmetry in a male secondary sexual character influences sexual

selection in wolf spiders. Behavioral Ecology and Sociobiology. 38:321-326.

Vitt, L., Congdon, J., Dickson, N. 1977. Adaptive strategies and energetics of tail

autotomy in lizards. Ecology. 58:326-337.

Vollrath, F. 1990. Leg regeneration in web spiders and its implications for orb weaver

phylogeny. Bulletin of the British Arachnological Society. 8(6):177-184.

Vollrath, F. 1995. Lyriform organs on regenerated spider legs. Bulletin of the British

Arachnological Society. 10 (3):115-118.

Waddy, S., Aiken, D., de Kleign, D. 1995. Control of growth and reproduction. In: Biology

of the lobster Homarus americanus. (ed. Factor, J.) pp. 217-266. San Diego, CA.

Academic Press.

Wise, D. 1975. Food limitation of the spider Linyphia marginata: experimental field studies.

Ecology. 56: 637-646.

Zeleny, C. 1905. Compensatory regulation. Journal of Experimental Zoology. 2: 1-102.

36

Table 1.1: Differences in frequency of leg loss in the field for S. ocreata

Season/Year N % of spiders autotomized/regenerating Spring 2003 693 12.96%

Fall 2003 451 19.3%

Spring 2004 749 15.49%

Fall 2004 656 16.16%

37

Table 1.2: Results of Three-Factor ANOVA (injury status, year, and season [year]) comparisons for autotomized and intact spiders of Spring and Fall 2003 and 2004. Significant p values are shown in bold text.

df SS F P Cephalothorax width Injury status 1 0.989 8.6841 0.0034 Year 1 0.192 1.6865 0.1948 Injury status x Year 1 0.718 6.2975 0.0125 Season [year] 2 90.476 397.0323 <0.0001 Injury status x Season [year] 2 0.565 2.4781 0.0852 Weight Injury status 1 0.00091 21.6454 <0.0001 Year 1 0.00072 17.3851 <0.0001 Injury status x Year 1 0.000089 2.1163 0.1465 Season [year] 2 0.0578 690.5843 <0.0001 Injury status x Season [year] 2 0.000346 4.1278 0.0168 Body condition Injury status 1 0.00029 20.5693 <0.0001 Year 1 0.00016 11.3113 0.0008 Injury status x Year 1 0.000054 3.7525 0.0535 Season [year] 2 0.00015 5.1200 0.0064 Injury status x Season [year] 2 0.00023 7.7702 0.0005

38

Table 1.3: Results of Three-Factor ANOVA (injury status, season, and week [season]) comparisons for a single generation of spiders (Fall 2003 and Spring 2004). Significant p values are in bold text.

df SS F P Cephalothorax width Condition 1 1.5130 10.5326 0.0121 Season 1 0.6319 4.3989 0.0084 Week [Season] 2 1.9302 6.7182 0.0336 Condition x Season 1 0.6549 4.5590 0.5118 Week x Condition [Season} 2 0.0663 0.2307 0.8631 Weight Condition 1 0.0007 23.6629 <0.0001 Season 1 0.0230 760.2510 <0.0001 Week [Season] 2 0.0014 23.4628 <0.0001 Condition x Season 1 0.0003 11.2403 0.0009 Week x Condition [Season} 2 0.00001 0.3173 0.7284 Body condition Condition 1 0.00005 5.8512 0.0162 Season 1 0.000009 1.0706 0.3017 Week [Season] 2 0.0003 20.9663 <0.0001 Condition x Season 1 0.0000003 0.0383 0.8451 Week x Condition [Season} 2 0.00003 1.6470 0.1945

39

Table 1.4: Comparisons of mean (±SE) cephalothorax widths and weights for autotomized (AUT) and intact (INT) spiders of the same generation by season and by week.

Cephalothorax width (mm) Weight (mg)

SEASON WEEK AUT INT AUT INT

3.25 3.34 6.3 7.7 Fall 2003 1 ±0.06 ±0.06 ±0.9 ±0.9

3.44 3.45 7.9 8.6 Fall 2003 2 ±0.05 ±0.06 ±0.8 ±0.8

Spring 3.25 3.50 20.2 25.0 1 2004 ±0.07 ±0.06 ±1.0 ±0.9

Spring 3.45 3.68 26.4 32.6 2 2004 ±0.08 ±0.08 ±1.1 ±1.1

40

Table 1.5: Two sample t-tests comparisons of molt intervals (days) between autotomized and intact lab-reared spiders in 2003, with significant results boldfaced.

Autotomized Intact

Spiders Spiders Mean Mean N (days) N (days) t df p

Molts 4/5 60 34.22± 54 30.56± 2.290 112 0.024 1.10 1.16

Molts 5/6 56 33.26± 50 34.89± -1.042 104 0.300 1.72 1.81

Total Molts 56 66.60± 50 64.44± 0.625 104 0.5331 2.4 2.47

41

35%

30% s 25%

20%

15%

10% % of missing% leg of 5%

0% I II III IV Leg position

Figure 1.1: Percentage of legs lost or regenerated according to 2 position for field caught S. ocreata (X 3=19.5, p<0.001). Data was pooled for the spring and fall of 2003 and 2004 (N=517).

42

A) B)

45 Intact 4 * Intact 40 * * Autotomy 3.5 Autotom y 35 3 * 30 2.5 25 2 20 1.5 15 * weight (mg) weight 1 10 * 0.5 5 0 0 Spring Spring Fall Fall Cephalothorax width (mm) Spring Spring Fall Fall 2003 2004 2003 2004 2003 2004 2003 2004

C)

0.006

0.005 Intact 0.004 Autotomy 0.003 * 0.002 0.001 * 0 -0.001 Spring Spring Fall Fall -0.002 2003 2004 2003 2004

Body condition indexBody -0.003 -0.004 -0.005

Figure 1.2: Mean ± SE of A) Average weights, B) cephalothorax widths and C) Body

conditions for autotomized and intact field caught S. ocreata captured during

comparable times in spring and fall of 2003 and 2004. * indicate a significant

difference between intact and autotomized spiders at P<0.05.

43

A) B)

Intact Intact Spring 2004 Fall 2004 Autotomy * Autotom y 0.004 0.001 0.003 0.002 0.0005 0.001

0 BCI 0

BCI -0.001 -0.002 -0.0005 -0.003 A)-0.004 -0.001

Collection Early 1 Collection Middle 2 Collection Early 1 CollectionMiddle 2

Figure 1.3: Comparisons of BCI (Residuals of weight x cephalothorax width) in early vs. middle season for A) Spring 2004 and B) Fall 2004. * indicate significant differences between autotomized and intact spiders at p<0.05.

44

(A) (B)

3.45 35 3.4 3.35 30 3.3 25 3.25 20 3.2 3.15 15

3.1 Weight (mg) 10 3.05 ab a b a ba 5 3 Cephalothorax width (mm) width Cephalothorax 2.95 0 Intact Regenerate 1 Regenerate 2 Intact Regenerate 1 Regenerate 2

Figure 1.4: Mean ±SE for (A) cephalothorax widths and (B) weights between spiders of the same age which were caught in the field and then underwent autotomy in the lab, according to treatment (Fall 2004). Intact spiders were unmanipulated, Regenerate 1 spiders molted once after autotomy, and Regenerate 2 spiders molted twice before measurements were made. Tukey posthoc tests were used to compare the means (p<0.05). Identical letters indicate groups that are not significantly different.

45 A. Int 60 Figure 1.5: a a b Aut 1 leg Differences were measured 50 Aut 2 legs between spiders that were intact or with one or both 40 forelegs regenerating. Identical letters indicate 30 a ab b a a b

Days groups that are not significantly different 20 according to Tukey posthoc

10 tests (p<0.05).

0 A) Mean ± SE molt interval (days) were measured. Interval 1: Days until partial B. regeneration; Interval 2: Days Aut 1 leg between partial and full 80 aa Aut 2 legs regeneration Total: Days for 70 both intervals.

60 B). Mean ± SE weight (mg) 50 a b were measured. Spiders were 40 weighed initially before manipulation and within 24 a a 30 hours after the following two Weight (mg) 20 molts. Weights of Intact 10 spiders were not included as they were not obtained within 0 24 hours of molting and would thus have been inaccurate.

C. C). Mean +SE cephalothorax 6 Int widths (mm) were measured. aaa Spiders were measured Aut 1 leg 5 a ab b initially before manipulation Aut 2 legs and after the following two 4 a a a molts.

3

2

1

Cephalothorax width (mm) width Cephalothorax 0

46 Chapter II:

The effects of autotomy and regeneration on prey capture rate, capture efficiency and sensory

detection in Schizocosa ocreata (Araneae, Lycosidae).

(For submission to Journal of Insect Behavior)

Kerri M. Wrinn and George W. Uetz

47 Abstract:

Previous laboratory experiments have shown no effects of leg autotomy on prey capture

in adult wolf spiders. However, these effects may not be the same for juveniles, which have

different foraging patterns and are able to regenerate lost appendages. This study tested the effects of autotomy and regeneration on prey capture in juvenile Schizocosa ocreata wolf spiders in both artificial and semi-natural settings. Spiders were tested for capture efficiency (i.e., measures of latency to orient to, capture and subdue prey) in a 15 cm diameter circular arena with cricket prey. Sensory detection through vibration (i.e., measures of accuracy of orientation)

was also tested by placing spiders in the same type of arena but visually isolating them from their

prey. Subsequent analyses showed no effects of autotomy or regeneration on any measures of

prey capture efficiency. Similarly, spiders’ vibratory sensory abilities were not significantly

affected by autotomy or regeneration. However, when spiders were tested in a semi-natural

habitat (a mesocosm filled with leaf litter), individuals with a missing or regenerating leg had

reduced prey capture rates. This suggests that while negative effects of autotomy and

regeneration do not appear to be directly attributable to mechanical or sensory impacts on

foraging, they may only be apparent in more complex environments such as the spider would encounter in nature.

48 Introduction

Autotomy (self amputation) of appendages and subsequent regeneration of the lost parts

are common across many taxonomic groups including vertebrates (Dial and Fitzpatrick, 1983),

echinoderms (Ramsay et al; 2001), crustaceans (Juanes and Smith, 1995), and arachnids

(Formanowitz, 1990). Both of these processes are presumed ancestral across the animal

kingdom (Goss, 1969), however, there must be a balance between the costs and benefits of

autotomy and regeneration in order for them to be maintained within a taxon over evolutionary

time.

The ability to autotomize an appendage that is being grasped by a predator may allow the

animal to escape with its life, providing a direct survival benefit (Formanowitz, 1990; Klawinski

and Formanowitz, 1994; Punzo, 1997). However, autotomy can incur fitness costs later. Loss of

an appendage has been known to impair foraging abilities (Brock and Smith, 1998; Vollrath,

1990; Ramsey et al., 2001), locomotion (Amaya et al, 2001), competition (Mariappan et al.,

2000; Taylor and Jackson, 2003; Dodson and Beck, 1993), and mating (Bateman and Fleming,

2005).

Regeneration of lost appendages allows animals to negate some of the costs of autotomy.

However, regeneration can be energetically costly, affecting growth and development (Goss,

1969; Vitt et al., 1977; Juanes and Smith, 1995; Ramsey, et al., 2001). Additionally, the

regenerated appendage may have reduced function, affecting locomotion and foraging (Brock

and Smith, 1998), competition (Brock and Smith, 1998) and mating (Uetz et al., 1996, Taylor et

al., in press).

Autotomy and regeneration have been widely studied in arthropods. However, these

studies have focused mainly on crustaceans (Juanes and Smith, 1995) and insects (Bell and

49 Adiyodi, 1981; Stoks, 1999; Bateman and Fleming, 2005). There remains a paucity of

knowledge on autotomy and regeneration for arachnids, a group of arthropods for which these

processes are very important.

Spiders and other arachnids use their legs extensively for all aspects of prey capture,

from web building to sensory detection and physical restraint of prey. For this reason, autotomy

and regeneration have the potential to greatly affect the foraging process and thereby fitness.

The effects of autotomy on the physical capture and restraint aspects of foraging in wolf spiders

have been tested in a few laboratory studies (Amaya et al, 2001; Brueseke et al 2001). As wolf spiders do not build webs, they have leg adaptations for capturing and subduing prey i.e. spines, adhesive hairs, and specialized flexor muscles (Rovner, 1980). Amaya et al. (2001) measured the number of attacks needed to capture prey in autotomized and intact individuals of the wolf spiders Varicosa terricola and Schizocosa ocreata. Brueseke et al (2001) measured number of attacks and time to subdue prey in autotomized and intact individuals of the wolf spider Pardosa milvina. Neither study showed any reduction in overall prey capture in autotomized spiders.

However, Amaya et al (2001) found that larger S. ocreata with autotomized limbs were less effective at capturing prey, while Brueseke et al (2001) found that P. milvina with autotomized limbs tended to capture smaller crickets. Additionally, both of the previous studies were conducted using adult spiders in a laboratory setting, and the ability to regenerate is lost in most adult spiders as they cease molting. In spiders, legs take multiple moults to regenerate and thus newly regenerated legs are not fully sized. In many species these partial in-sized limbs are held rigidly away from the body and are thus nonfunctional (Vollrath, 1990). These nonfunctional legs may potentially be an impediment during prey capture. So far no studies have addressed the effects of autotomy and/or regeneration on foraging under more natural conditions, or have used

50 juveniles whose foraging patterns may differ from adults (Beck and Connor, 1992; Persons,

1999).

In addition to physical capture and restraint of prey, spiders also use their legs for sensory detection. Sensory hairs located on the legs known as trichobothria are used to detect airborne vibrations, while lyriform organs, each made up of a series of flexible slits, detect substrate borne vibrations. Lyriform organs on regenerated legs are sometimes malformed (Vollrath, 1995), which could potentially affect their function. Although losing a leg has been shown to impair sensory detection and prey capture in two species of web spinning spiders (Vollrath, 1995), no studies have addressed the affects of autotomy and regeneration on sensory detection of prey in a hunting spider.

For the following experiments I used the well-studied wolf spider species, Schizocosa ocreata, which readily autotomizes and regenerates legs. These spiders are abundant in the leaf litter of Eastern deciduous forests (Cady, 1984). In addition to being readily available, this species has short growth periods (10-30 days/period in the lab) and molts 6-8 times before reaching adulthood, making it a good group for which to study autotomy and regeneration.

This study addresses how autotomy and regeneration affect both the functional and sensory aspects of prey capture in S. ocreata. The objectives were: 1) To determine the effects of autotomy and regeneration on prey capture rate (number of crickets consumed/time period) under simulated natural conditions; 2) To measure the effects of autotomy and regeneration on prey capture efficiency (ie. latencies to orient to, attack and subdue prey) under controlled laboratory conditions, and; 3) To examine how autotomy and regeneration affect sensory perception (i.e., measures of accuracy of orientation).

51 General Methods:

For the following three experiments, spiders were captured from the forest floor leaf litter at the Cincinnati Nature Center, Rowe Woods, Clermont County, OH in Spring 2003.

Upon capture, each spider was brought into the lab, placed in an opaque 10 cm diameter deli dish

with a clear lid and given water ad libitum through a dental wick connected to a water dish below. Late juvenile/adult spiders were fed 10 day old crickets (Acheta domesticus) twice a week. For the duration of the experiments spiders remained under controlled laboratory conditions including stable humidity, room temperature (21-24 º) and a light: dark cycle of 13:11

hours. Data analysis for all experiments was done using JMP version 4.0.2 software.

Adult females were captured during June of 2003 and their offspring were raised in the

lab to produce individuals for experiments 1 and 2. Spiders were examined daily for egg sac

production and hatching. Spiderlings for the following two experiments were randomly taken

from a total of 11 egg sacs (each by a different female) in order to control for possible egg sac

effects. Upon hatching, spiderlings were left with the mother for 7-10 days. Spiderlings

normally cling to the mother’s abdomen for this length of time before becoming independent.

Upon dispersal, spiderlings were placed in separate 120 ml specimen cups and given water ad

libitum through a damp dental wick placed in the container. Spiderlings were fed 4-5

Collembola twice a week until they reached their third molt. At that point they were fed fruit

flies (Drosophila melanogaster) or pinhead crickets (Acheta domesticus) twice a week. As spiders reached their fourth molt they were randomly assigned to two experiments: 1) capture

rate, or 2) capture efficiency. At that time they were transferred over to the larger containers

used for adult spiders (see general methods above).

52 Objective 1: Effects of autotomy/regeneration on prey capture rate

If losing and regenerating legs affects prey capture rate, then spiders would be expected

to capture more crickets/time while intact than when missing or regenerating legs.

Methods:

Each spider (n=105) was placed along with five crickets in a clear plastic shoebox (32 cm x 17 cm x 9 cm) with 2-3cm of leaf litter covering the bottom. This served as a mesocosm simulation of their natural habitat and approximate home range size (Cady, 1984; Uetz, unpublished). Since cricket size may affect prey capture in autotomized individuals (Brueseke et al., 2001), spiders were randomly assigned crickets equal to either 33% or 50% of their body weight. Preliminary experiments showed that prey capture rate was highest during the first day.

Therefore, boxes were checked after 6 and 24 hours had passed and the number of crickets alive at each time period was recorded in order to get early and late comparisons of prey capture rates.

Each experimental spider was tested intact, then induced to autotomize one randomly chosen foreleg by restraining the femur with forceps. The forelegs were chosen for autotomy, as these are involved in prey capture (Rovner, 1980) and are more frequently seen missing in the field (Chapter I this thesis). Spiders that autotomized legs were rested 24 hrs prior to a second trial (Brueseke, et al, 2001). They were then tested a 3rd and 4th time within a week of each consecutive molt (first and second regeneration), as these spiders required at least two molts for the regenerated leg to become full-sized. A few spiders molted during a trial and were removed from further analysis as this process can affect foraging behavior (Morse, 2000). A separate set of intact spiders was tested in each time period to account for possible age or learning effects.

These intact spiders, like the experimental spiders, were divided into groups given crickets either

33% or 50% of their body weight.

53 Data Analysis:

A repeated measures MANOVA with cricket weight and manipulation as factors was

used to analyze cricket consumption / time period for the four groups (Intact spiders: 33% and

50% cricket weight, and Manipulated spiders: 33% and 50% cricket weight). The MANOVA

form of repeated measures test was used as this does not require the strict assumptions of

sphericity (e.g. equal variances between all repeated measurements) which other repeated

measures designs require (O’Brien and Kaiser, 1985). Violation of sphericity can make results

of a repeated measures ANOVA approach highly prone to Type 1 error. Because it is highly

unlikely that variances remained equal across all measurements for the current study, the

repeated measures MANOVA approach seemed the most appropriate (O’Brien and Kaiser,

1985). If significant factors were found in the initial repeated measures MANOVAs, two way

ANOVAs were then run with Bonferroni adjusted α-levels for multiple tests to determine which time periods were significant for the other factors (Schiener and Gurevitch, 2001).

Results

A subset of spiders (n=33 of n=105 total) molted to adulthood prior to the fourth trial (full

regeneration), and were thus only run in the first three trials. A repeated measures MANOVA performed on the first three trials at 24 hours including these individuals showed that both juvenile state (penultimate or antepenultimate at trial 3) and the juvenile state x manipulation interaction term were significant (Table 2.1). Because of these differences, those spiders maturing before the fourth trial were considered a separate population from the younger spiders and were removed from further analysis. The differences in foraging behavior between penultimate and antepenultimate spiders may have been due to the fact that the penultimate spiders were, on average, significantly larger. Two sample t-tests showed significant differences

54 in (natural log transformed) weights by juvenile state, with spiders penultimate at the third trial

weighing more at both the second (137 ± 6.1 vs 113 ± 4.4 mg; t1,81=-3.349, p=0.0012) and third trials (361 ± 15.8 vs 230 ± 11.3 mg; t1,81=-7.035, p<0.0001). Further experiments are needed to explore the relationship between juvenile stage and foraging in greater detail.

After the subset of early maturing spiders was removed, further repeated measures

MANOVAs were run on the remaining spiders. Analysis was run for two time periods (6 and 24 hours) in order to determine possible hourly differences. After 6 hours there was no overall affect of manipulation on prey capture rate, however cricket weight and the cricket weight x manipulation interactions were significant (Table 2.2). In the 50% weight group, manipulated spiders consumed significantly fewer crickets than intact spiders at the second (autotomy) and third (partial regeneration) trials at 6 hours (Figure 2.2C). By 24 hours there was no longer a significant effect of manipulation on prey capture rate for this group at trial two, although the trend for higher capture in intact spiders remained (Figures 2.2D; Table 2.2). The negative effect of manipulation on capture rate was not present until 24 hours in the 33% group and only affected spiders at the second trial (autotomy) (Figures 2.2 A,B Table 2.2).

Discussion:

It is not surprising that the effects of autotomy appeared sooner when larger prey items were involved. A laboratory study by Brueseke et al, (2001) also showed an interaction between cricket weight and autotomy in the wolf spider Pardosa milvina, where individuals with a foreleg autotomized were less likely to capture larger crickets. A possible reason why regenerating spiders in the 50% prey weight group had reduced capture whereas regenerating spiders in the 33% prey weight group did not is due to increased cricket size/weight at the third trial. Before the third trial spiders had molted once in order to allow the autotomized group to

55 regenerate. Thus they were given heavier crickets proportionate to their new size. It is possible that the actual size of the crickets increased faster than their weight, making them more difficult to catch for the spiders regenerating a leg.

Autotomy of a leg had a negative impact on rate of prey capture, but only at the second and third trials. The differences between manipulated and intact spiders at each of these time periods were due to a slight increase in prey capture rate from the initial trial by the intact spiders and a slight reduction in prey capture by injured spiders. This may indicate learning on the part of the intact spiders, which allowed them to capture more crickets with experience. A study by

Heiling and Hergerstein (1999) showed that the web spinning spiders Argiope keyserlingi and

Larinioides sclopetarius changed the structure of their webs in response to past prey capture experiences, allowing a higher capture rate. Additionally, Morse (2000) discovered that the juvenile crab spider Misumena vatia increased efficiency of prey capture with experience.

It is possible that due to their injury, spiders that had autotomized a leg were unable to increase their prey capture rate between trials as the intact spiders did, and in fact had a slight decrease in capture rate. The average prey capture rate at 24 hours for the second trial was lower for autotomized spiders in both cricket weight categories. A few studies have similarly found negative affects of autotomy on prey capture rate in other animals. Feeding rate was significantly lower in the crab Callinectes sapidus only when it was missing both chelipeds

(Juanes and Smith, 1995). Additionally, Cooper (2003) found in a field setting that lizards

(Holbrookia propinqua) which had undergone tail autotomy made fewer attempts to capture prey per time than intact lizards. If autotomized spiders did give up sooner when attempting to capture crickets, this could have contributed to the lower rate of capture. Regeneration of a leg also negatively impacted prey capture rate for the 50% cricket weight group. It was not

56 unexpected that the spiders would have decreased prey capture rate with a partially regenerated

leg, as newly regenerated legs are often nonfunctional (Vollrath, 1990). Reduced feeding efficiency, and switching to suboptimal prey occur in several species of crabs when they are regenerating a front claw (Juanes and Smith, 1995). Similarly, a study by Weissman, 1987 (cited in Vollrath, 1990) demonstrated that prey capture rate was lower in the web spinning spiders

Zygiella and Nephila when regenerating legs.

This study had a few logistical problems with the potential to affect results. Several of the spiders molted to maturity before being run in the last trial, reducing the sample size. There were also occasional occurrences of crickets being found dead and unconsumed by the spiders.

It is possible that these crickets died from other causes, although “wasteful killing” has been previously shown for this species, where spiders killed more prey than they consumed (Persons,

1999). Regardless of the cause, this cricket death without consumption usually only occurred several days after the beginning of each trial, so it did not interfere with the reported measures at

6 and 24 hours.

Objective 2: Effects of autotomy/regeneration on prey capture efficiency

If losing and regenerating a leg affects prey capture efficiency, then intact spiders would be expected to have decreased capture latencies compared to spiders with missing or regenerated

legs.

Methods:

Each spider (n=120) was tested five times (intact, 24 hr post-autotomy, one week post-

autotomy, and after the two molts needed for full regeneration). As cricket weight may affect

prey capture in autotomized spiders (Brueseke et al, 2001), spiders were randomly assigned to

57 three groups (n=15-20) and fed crickets 25%, 33%, or 50% of the spider’s body weight. A second set of spiders also divided into groups by cricket weight (25%, 33%, or 50%), was run intact each time period to account for age or learning effects.

During experimentation, the feeding schedules were reduced so that each spider received only two crickets per week. As demonstrated in preliminary experiments, this was enough to maintain the spiders, but kept their hunger levels high enough so they were more likely to feed during trials. Spiders were placed in the experimental container 24 hours prior to being run, as preliminary experiments showed many of the spiders did not capture prey immediately when placed in the container without the acclimation period.

For each trial, the spider was restrained under a vial in the center of a circular plastic testing arena (15 cm diameter x 6.5 cm high) with filter paper covering the bottom. The vial was opaque on the sides (so the spider was not distracted visually), but clear on the top (to reveal the direction the spider was facing). The filter paper was changed between trials and the arena wiped with cotton and 70% ethanol to remove any traces of chemical signals from the previous spider. For each trial, a cricket was introduced using a modified syringe via a hole in the side of the container behind the spider and approximately 4 cm above the bottom. When the cricket entered the arena, the spider was released and the trial began.

The spider/cricket interactions were recorded from above using a Watec America

Corporation WAT-902C video camera until the spider began to feed (or 10 min). Trials were scored upon playback for the following parameters: 1) latency (s) to orient to prey; 2) latency (s) to attack; 3) number of attacks necessary for capture; 4) Whether capture was successful (y/n); and 5) time (s) to subdue prey. Latency to orientation was scored as the time between when the spider was released and when it either turned or moved towards the cricket. If the spider’s

58 orientation and attack could be separated, attack latency was measured as the time between when

the spider oriented to the cricket and when it initially lunged at and attempted to grab the prey. If

the spider’s orientation and attack could not be separated (e.g. if the cricket approached the spider from the front) then attack latency was scored as 0.5 seconds. This value was used as it was greater than 0, but less than the 1 second which was the smallest measurable difference between orientation and attack times. Sometimes spiders would require multiple attempts to

capture the prey, so the number of attacks was recorded as well. Subdue time was measured as

the time between attack and when the spider successfully restrained the cricket.

Data Analysis:

The nominal variable of prey capture success (+/-) was analyzed by pairing trials (e.g.

spiders (cricket weight 33%): trial 1(intact) vs. trial 2 (autotomized)) and using McNemar’s tests

for paired sample data (Zar, 1999). Data for orient latency, attack latency, and subdue time were

natural log transformed for normality. For the reasons stated previously, separate repeated

measures MANOVAs were used to analyze the four continuous parameters being measured

(orient latency, attack latency, # attacks, subdue time) for each of the six groups: (Manipulated

spiders: cricket weight 25%, 33%, 50% and Intact spiders: cricket weight 25%, 33%, 50%).

Results

For all of the following response variables except for orient latency the 50% group was

dropped from overall analysis as spiders in this group were often unresponsive, drastically

reducing sample size (especially in the last two trials). According to repeated measures

MANOVAs there was no effect of manipulation (autotomy vs. intact) on orient latency, attack latency, number of attacks, or latency to subdue a cricket (Tables 2.3-2.6). Likewise, McNemar tests for paired sample nominal-scale data showed no significant differences in likelihood of prey

59 capture between trials for either intact or autotomized spiders at any of the cricket weights (Table

2.7). Trial week was a significant factor for two of the dependent variables (attack latency and

subdue time; Tables 2.4, 2.6). Natural log transformed attack latency decreased for all spiders

over the first three trials, then increased again for the 33% prey weight group only in the last two

trials. Latency to subdue also increased in the last two trials for all spiders in the 33% prey

weight group, but not the 25% prey weight group, as indicated by the significant trial week x

cricket weight interaction (Table 2.6).

Discussion:

This experiment showed no effects of autotomy/regeneration on prey capture efficiency,

which matches results of other laboratory studies performed on adult wolf spiders. Amaya et al,

2001 found no difference in number of attacks between autotomized and intact spiders of the

species S. ocreata or V. terricola. Likewise Brueseke et al, (2001) found no difference in the

number of attacks or time to subdue prey in autotomized vs. intact individuals of P. milvina wolf

spiders. The Amaya et al (2001) study did not specify what size crickets were used, and the

Brueseke et al. (2001) study used only crickets 25% of the spider’s body weight. As Brueseke et al (2001) demonstrated a trend for autotomized spiders to be less likely to capture larger prey

(within the range of size for the prey presented) it was expected that S. ocreata would show

similar trends in the current study. There was however, no significant autotomy x cricket weight

interaction at any of the trial periods (Tables 2.3-2.7).

The changes in attack latency and subdue time by trial regardless of leg manipulation

may be attributable to two factors: experience by the spider and change in spider/cricket size

over time. Morse (2000) showed a decrease in both orient and attack latencies accompanying

experience in the juvenile crab spider Misumenata vatia. Likewise, the decrease in attack

60 latency during the first three trials for S. ocreata, may indicate increased efficiency due to

experience as well. However, the increases in attack latency and subdue times during the last

two trials for spiders in the 33% prey weight group may have been due to a change in cricket

size. As spiders molted prior to the last two trials, the weights of the crickets they were given

were adjusted accordingly. Although the crickets still weighed the same proportionate to the

spiders, they were often as large as the spiders in the 33% prey weight group, and even larger in

the 50% prey weight group (personal observation). A study by Nentwig and Wissel (1986)

demonstrated that the wolf spiders Alopecosa cuneata and Pardosa lugubris would not accept

crickets greater than 150% of their body length and that acceptance rate was highest at 50-80%

of the spider’s body length. If this is also true for S. ocreata it would explain the increase in

attack latency and lengthened subdue times by the 33% prey weight group in the fourth and fifth

trials. Likewise, this change in prey size may also explain the greatly reduced prey acceptance

rates by the spiders in the 50% prey weight group which led them to be dropped from the

analysis. Reduced capture rate of larger crickets also occurred for spiders in the 50% prey weight group in the mesocosm experiment (Objective 1). However, the fact that the spiders in that experiment were allowed a longer capture time before being measured (6 hours vs.10 minutes) ensured that those spiders did not refuse the prey entirely.

Objective 3: Effects of autotomy/regeneration on vibration sensory ability

If losing and/or regenerating a leg affects vibration sensory ability, then spiders would be

expected to orient more often, more quickly and show lower error in orientation when intact.

61 Methods:

Juvenile spiders (n=160) were collected in the field from the Cincinnati Nature Center in fall 2004 and maintained in the lab (see general methods). All spiders were randomly assigned to one of four groups (n=40/group) and were allowed to molt twice within the lab before being run in trials. In order to ensure that spiders in all groups were approximately the same age when they were run in experiments, spiders were induced to autotomize a foreleg at different time periods based on treatment. Thus, one group of spiders had a fully regenerated leg when they were run in trials, a second group had a partially regenerated leg, a third group was missing a leg, and the fourth group was intact, having been unmanipulated. Spiders were fed two crickets

(Acheta domesticus) one week before trials in order to standardize hunger levels.

For each trial the spider was placed inside a 3.5 cm wide opaque circle of plastic in the center of a 15 cm diameter plastic arena with filter paper on the bottom (as in experiment 2). A cricket, also restrained by a 2 cm wide circle of plastic, was then placed on the filter paper at an approximately 90° angle to the spider. Thus the spider and cricket could sense each other through vibrations in the substratum but were isolated visually. For spiders which had an autotomized or regenerated leg, the cricket was placed on the side where autotomy had occurred, while side placement was chosen randomly for intact spiders. Each spider was weighed prior to the trial, and assigned a cricket weighing approximately 50% of its weight. Preliminary trials showed that crickets of this weight, while still falling under the spider’s normal prey size range, were more likely to be detected by the spider than smaller crickets. Each trial (duration: 5 minutes) was taped from above using a Canon XL-1 digital video camera and later analyzed frame by frame for angle measurements using the program Image J (Wright Cell Imaging

62 facility, Toronto Western Research Institute). Both the spider’s time of initial orientation to the

prey and the angle of orientation to the prey were measured (Figure 1).

Data Analysis:

Orientation to prey (+/-) was compared between groups using a 4 X 2 Chi square

contingency table. Further analyses were conducted on the spiders that did orient. A Kruskal-

Wallis test was used to compare error angle between treatments (Zar, 1999). Also a 4 X 2 Chi

square contingency table was used to determine any differences between the numbers of spiders

in each group which oriented in the correct direction. Finally a one-way ANOVA was used to

compare time to initial orientation.

Results:

There were no statistically significant effects of autotomy and regeneration on sensory

detection of prey. There were no significant differences between groups in likelihood to orient at

2 all (X 3 = 1.755, p=0.6248; Figure 2.3A), or if the spiders did orient, no significant difference in

2 likelihood of orienting correctly (X 3= 3.136, p=0.3711; Figure 2.3B). For the purposes of this

experiment what was considered “correct” orientation was broadly defined as spiders with an

error angle (γ) smaller than the stimulus angle (β), as error angles tended to be high for all

groups. Latency to orient measured in seconds was natural log transformed for normality. Once again, latency to orient did not differ significantly between groups (ANOVA: F3,198=0.9236,

p=0.4304; Figure 2.4). A Kruskal-Wallis test showed no significant difference in mean error

2 angles of orientation between the groups (X 3=1.5658, p=0.6672, Figure 2.5).

Discussion:

Hergenroder and Barth (1983) showed that vibratory information from the forelegs was most important in determining turning angle when the hunting spider Cupiennius salei was

63 orienting to prey. If this is also true for S. ocreata then spiders with an autotomized or regenerated foreleg would be expected to orient more slowly and less accurately than intact spiders. However, in most cases, regardless of leg condition, S. ocreata tended to have inaccurate orientation to isolated vibrations, as the average error angle of even the intact spiders was high (72°). This may have masked any differences between the groups somewhat. As with

C. salei (Hergenroder and Barth, 1983), S. ocreata tended to under-orient, turning at less of an angle then they needed to face the prey. Additionally, in this experiment spiders would sometimes reorient if the first angle was incorrect. This is unlike other spiders such as the web spinning species Zygiella x-notata and Nephila clavipes, which have low error angles of 4-7°

(Klarner and Barth, 1982). Hergenroder and Barth (1983) suggested that as the prey of C. salei is moving and not stuck in a web then these spiders only orient far enough to reach the prey before it can escape. It is possible that this is also true for S. ocreata, which hunt prey capable of quickly fleeing as well.

Although the effects of autotomy and regeneration on sensory ability were not statistically significant in this study, they may be biologically significant in the field. It is possible that it would be more difficult for spiders missing or regenerating a leg to compensate for reduced vibration sensory abilities in a more complex environment. This may be especially true for spiders like C. salei, which hunt at night relying entirely on vibrations to sense their prey

(Barth, 2002). S. ocreata, however, are active both day and night and rely heavily on vision as well as vibratory cues for prey capture (Persons and Uetz, 1996). Persons (1999) found that generally S. ocreata tended to spend more time in patches where they were given both visual and vibratory prey cues as opposed to either cue separately. However, all spiders regardless of age or sex would respond to visual cues alone, but only immature females would reside longer in

64 patches with vibratory cues alone (Persons, 1999). In the closely related species, Schizocosa rovneri, juvenile spiders (but not adults) responded to vibratory cues alone when determining patch residence time. However, visual cues only and visual cues and vibratory cues together still led to a greater response (Persons and Uetz, 1999). The spiders in the current study were juveniles and therefore vibratory cues alone may have still been important in foraging. However, in nature vibratory signals would usually only be present without the corresponding visual cue at night. Further experiments and/or field observations would need to be run to determine the extent of prey capture by these spiders in the dark. With the addition of the visual sensory mode, spiders should have even less impairment when orienting. Therefore for daytime hunting, losing or regenerating a leg would probably have minimal effect on the sensory aspects of prey capture.

Even though the current study was done under laboratory conditions, some factors varied between trials. The crickets were not all equally active, which may have affected the spider’s ability to detect prey vibrations. Additionally, sometimes a cricket would be moving as the spider was orienting so the spider would have to adjust while turning to have the correct response angle. However, it was assumed that these differences in cricket behavior were the same across all groups so the overall averages for each group would not be affected.

Conclusions:

The results of laboratory experiments indicate that autotomy and regeneration do not appear to have significant impacts on foraging as measured by sensory detection (orientation latency or accuracy) or physical capture of prey (number of attacks, subdue time, etc.).

However, these experiments by virtue of being in the lab, were lacking in the environmental complexity that occurs in the field. The prey capture rate experiment carried out in the leaf litter mesocosms provided a more natural habitat, and it was in this experiment that autotomy and in

65 some cases regeneration was detrimental to rate of prey capture. No other studies to my

knowledge have examined the effects of autotomy and regeneration on hunting spiders in a field

setting. In order to truly understand the effects of autotomy and regeneration on spiders in

natural populations, further studies should be run to address the many factors with the potential to interact with leg loss and affect prey capture in the field, e.g., habitat complexity, competition from conspecifics and chances of predation.

66 References

Amaya, C., Klawinski, P, Rormanowicz, D. 2001. The effects of leg autotomy on

running speed and foraging ability in two species of wolf spider (Lycosidae).

American Midland Naturalist. 145: 201-205.

Barth, F. 2002. A spider’s world: senses and behavior. Springer-Verleg. Berlin.

Bateman, P., Fleming, P. 2005. Direct and indirect costs of limb autotomy in field

Crickets, Gryllus bimaculatus. Animal Behaviour. 69:151-159.

Beck, M., Connor, E. 1992. “Factors affecting the reproductive success of the crab spider

Misumenoides formosipes- the covariance between juvenile and adult traits.” Oecologia.

92 (2): 287-295.

Brock, R.E, Smith LD. 1998. “Recovery of claw size and function following autotomy in

Cancer productus (Decapoda: Brachyura).” Biological Bulletin. 194 (1): 53-62.

Brueseke, M., Rypstra, A., Walker, S., Persons, M. 2001. “Leg autotomy in the wolf

spider Pardosa milvina: a common phenomenon with few apparent costs.” American

Midland Naturalist. 146:153-160.

Cady, A.1984. “Microhabitat selection and locomotor activity of Schizocosa ocrata

(Walkenaer) (Araneae: Lycosidae).” Journal of Arachnology. 11: 297-307.

Cooper, W. 2003. “ Shifted balance of risk and cost after autotomy affects use of cover, escape,

activity, and foraging in the keeled earless lizard (Holbrookia propinqua). Behavioral

Ecology and Sociobiology. 54: 179-187.

Dial, B.E, Fitzpatrick, LC. 1984. Predator escape success in tailed versus tailless

Scincella lateralis (Sauria, Scincidae). Animal Behaviour. 32: 301-302.

67 Dodson, G., Beck, M. 1993. Precopulatory guarding of penultimate females by male

crab spiders, Misumenoides formosipes . Animal Behaviour. 46: 951-959.

Formanowicz, D.1990. The antipredator efficacy of spider leg autotomy. Animal

Behaviour. 40:400-401,

Goss, R. 1969. Principles of Regeneration. Pp. 91-112. New York, Academic Press.

Heiling, A., Herberstein, M. 1999. The role of experience in web-building spiders

(Araneidae). Animal Cognition. 2: 171-177.

Hergenroder, R., Barth, F. 1983. Vibratory signals and spider behavior: how do the

sensory inputs from the eight legs interact in orientation? Journal of

Comparative Physiology. 152:361-371.

Juanes, F., Smith, L.D. 1995. The ecological consequences of limb damage and loss in

decapod crustaceans: a review and prospectus. Journal of Experimental Marine Biology

and Ecology. 193(1-2): 197-223.

Klarner, D; Barth, F. 1982. Vibratory signals and prey capture in web spiders (Zygiella x-

notata, Nephila clavipes). Journal of Comparative Physiology. 148: 445-455.

Klawinski, P.; Formanowitz, D. 1995. Ontogenetic change in survival value of leg

autotomy in a wolf spider, Gladicosa pulchra (Keyserling) (Araneae: Lycosidae), during

scorpion attacks. Canadian Journal of Zoology. 72: 2133-2157.

Kunkel, J.G. 1981. Regeneration. In : The American Cockroach. (Ed. Bell, Adiyodi, ). pp. 427-

443. Chapman and Hall. New York, NY.

Mariappan, P., Balsundaram, C., Schmitz, B. 2000. Decapod crustacean chelipeds: an

overview. Journal of Biosciences. 25 (3): 301-313.

68 Morse, D. 2000. The effect of experience on the hunting success of newly emerged

Spiderlings. Animal Behaviour. 60: 827-835.

Nentwig, W.; Wissel, C. 1986. Comparison of prey lengths among spiders. Oecologia. 68:

595-600.

O’Brien, R., Kaiser, M. 1985. MANOVA method for analyzing repeated measures

designs: an extensive primer. Psychological Bulletin. 97 (2): 316-333.

Persons, M. 1999. Hunger affects on foraging responses to perceptual cues in immature

and adult wolf spiders. Animal Behaviour. 57:81-88.

Persons, M., Uetz, G. 1996. The influence of sensory information on patch residence time in

wolf spiders (Araneae: Lycosidae). Animal Behaviour. 51:1285-1293.

Persons, M,; Uetz, G. 1999. Age and sex-based differences in the use of prey sensory cues in

the wolf spiders (Araneae: Lycosidae). Journal of Insect Behavior. 12(6): 723-736.

Punzo, F. 1997. Leg autotomyand avoidance behaviour in response to a predator in the

wolf spider, Schizocosa avida (Araneae: Lycosidae). Journal of Arachnology. 25: 202-

205.

Ramsey, K., Kaiser, MJ., Richardson, CA. 2001. Invest in arms: behavioral and

energetic costs of multiple autotomy in starfish (Asterias rubens). Behavioral Ecology

and Sociobiology. 50(4):360-365.

Rovner, J. 1980. Morphological and ethological adaptations for prey capture in wolf

spiders (Araneae, Lycosidae). Journal of Arachnology. 8:201-215.

Stoks, R. 1999. Autotomy shapes the trade-off between seeking cover and foraging in

larval damselflies. Behavioral Ecology and Sociobiology. 47:70-75.

69 Taylor P., Jackson, R. 2003. Interacting effects of size and prior injury in jumping spider

conflicts. Animal Behaviour. 65: 787-794.

Uetz, G., McClintock, W., Miller, D., Smith, E., Cook, K. 1996. Limb regeneration and

subsequent asymmetry in a male secondary sexual character influences sexual

selection in wolf spiders. Behavioral Ecology and Sociobiology. 38:321-326.

Vitt, L., Congdon, J., Dickson, N. 1977. Adaptive strategies and energetics of tail

autotomy in lizards. Ecology. 58:326-337.

Vollrath, F. 1990. Leg regeneration in web spiders and its implications for orb weaver

phylogeny. Bulletin of the British Arachnological Society. 8(6):177-184.

Vollrath, F. 1995. Lyriform organs on regenerated spider legs. Bulletin of the British

Arachnological Society. 10 (3):115-118.

Von Ende, C. 2001. Repeated measures analysis. In: Designs and analysis of ecological

experiments. (ed. Scheiner, S., Gurevitch, J.) pp.134-157. New York, NY. Oxford

University Press.

Zar, J. 1999. Biostatistical Analysis. Prentice Hall. New Jersey.

70

Table 2.1: Repeated measures MANOVA results for prey capture of spiders given crickets 33% or 50% of the spiders body weight at Trials 1-3. Tests were run at 24 hours. All spiders were included (penultimate and antepenultimate at Trial 3). CW=cricket weight, Juv St= juvenile state, Man=manipulation. Significant p values are in bold text. Effect F df p Between Subjects All between 5.816 7,83 <0.0001 Man 0.0508 1,83 0.8223 CW 22.631 1,83 <0.0001 CW x Man 0.3102 1,83 0.5791 Juv St 5.855 1,83 0.0177 Juv St x CW 0.2534 1,83 0.6160 Juv St x Man 8.5076 1,83 0.0045 Juv St x CW x Man 1.1596 1,83 0.2847 Within Subjects All within 1.513 14,184 0.1108 Trial wk 5.64 2,82 0.0051 Trial wk x Man 0.7797 2,82 0.4619 Trial wk x CW 4.8255 2,82 0.0104 Trial wk x CW x Man 0.9610 2,82 0.3868 Trial wk x Juv St 1.667 2,82 0.1951 Trial wk x Juv St x CW 0.166 2,82 0.8470 Trial wk x Juv St x Man 0.287 2,82 0.7516 Trial wk x Juv St x Man x CW 0.670 2,82 0.5145

71

Table 2.2: Repeated measures MANOVA results for prey capture rate at 6 and 24 hours by manipulated or intact spiders over four trial periods. Spiders were given crickets weighing either 33% or 50% of their body weight. Significant p values are in bold text. 6 hours 24 hours Effect F Df p F Df p Between subjects All between 5.249 3,52 0.0031 8.3369 3,53 0.0001 Man 1.579 1,52 0.2144 4.8019 1,53 0.0419 CW 6.618 1,52 0.0130 15.5674 1,53 0.0002 CW x Man 5.189 1,52 0.0269 2.481 1,53 0.1499 Within subjects All within 1.8477 9,156 0.0638 2.1579 9,159 0.0277 Trial wk 0.6251 3,50 0.6022 1.5106 3,51 0.2229 Trial wk x Man 1.1213 3,50 0.3494 2.1728 3,51 0.1026 Trial wk x CW 4.1668 3,50 0.0104 2.9137 3,51 0.0431 Trial wk x CW x Man 0.8572 3,50 0.4695 1.5988 3,51 0.2011

72

Table 2.3: Repeated measures MANOVA of latency to orient over five trial periods for spiders given crickets weighing 25%, 33%, or 50% of their weight.

Effects Value F DF P Between subjects All between 0.00888 0.0941 5,53 0.9928 Manipulation 0.000212 0.0112 1,53 0.9160 Cricket weight 0.00222 0.0589 2,53 0.9429 Manipulation x cricket weight 0.00567 0.150 2,53 0.861 Within subjects All within* 0.372 1.0874 20,212 0.3644 Trial wk 0.0999 1.249 4,50 0.3025 Trial wk x manipulation 0.0443 0.554 4,50 0.697 Trial wk x cricket weight 0.135 0.924 8,102 0.5004 Trial wk x cricket weight x manipulation 0.202 1.433 8,102 0.1917 *Pillai’s trace reported measurements for within subject interactions, as this test is considered the most robust (Scheiner and Gurevitch, 2001)

73

Table 2.4: Repeated measures MANOVA of latency to attack over five trial periods for spiders given crickets weighing 25% or 33% of their weight. Significant p values are in bold text.

Effects Value F DF P Between subjects All between 0.0529 0.669 3,38 0.5759 Manipulation 0.00145 0.0553 1,38 0.8154 Cricket weight 0.0382 1.452 1,38 0.2356 Manipulation x cricket weight 0.00531 0.202 1,38 0.6559 Within subjects All within* 0.3728 1.361 12,111 0.2214 Trial wk 0.889 7.787 4,35 0.0001 Trial wk x manipulation 0.222 1.945 4,35 0.1248 Trial wk x cricket weight 0.147 1.290 4,35 0.2926 Trial wk x cricket weight x manipulation 0.0646 0.566 4,35 0.6892 *Pillai’s trace reported measurements for within subject interactions

74

Table 2.5: Repeated measures MANOVA of number of attacks over five trial periods for spiders given crickets weighing 25% or 33% of their weight.

Effects Value F NumDF P Between subjects All between 0.00688 0.0826 3,36 0.9691 Manipulation 0.00225 0.0811 1,36 0.7775 Cricket weight 0.00427 0.1518 1,36 0.6991 Manipulation x cricket weight 0.0000062 0.0002 1,36 0.9881 Within subjects All within* 0.424 1.441 12,33 0.1591 Trial wk 0.106 0.878 4,33 0.4875 Trial wk x manipulation 0.218 1.798 4,33 0.1529 Trial wk x cricket weight 0.0695 0.574 4,33 0.6837 Trial wk x cricket weight x manipulation 0.208 1.718 4,33 0.1695 *Pillai’s trace reported measurements for within subject interactions

75

Table 2.6: Repeated measures MANOVA of subdue time over five trial periods for autotomized and intact spiders given crickets 25% or 33% of their body weight. Significant p values are in bold text.

Effects Value F DF P Between subjects All between 0.1181 0.9846 3,25 0.4159 Manipulation 0.0000089 0.00002 1,25 0.9882 Cricket weight 0.07994 1.9985 1,25 0.1698 Manipulation x cricket weight 0.0568 1.4202 1,25 0.2446 Within subjects All within* 0.7724 2.0806 12,72 0.0289 Trial wk 0.221 6.0837 4,22 0.0019 Trial wk x manipulation 0.4098 2.2538 4,22 0.0960 Trial wk x cricket weight 0.5634 3.0989 4,22 0.0364 Trial wk x cricket weight x manipulation 0.2917 1.6046 4,22 0.2085 *Pillai’s trace reported measurements for within subject interactions

76

Table 2.7: McNemar tests for paired comparisons between consecutive trial periods for autotomized and intact spiders given crickets weighing 25%, 33%, and 50% of their body weights.

Condition/Cricket Condition/Cricket weight Trials X2 p weight Trials X2 p Autotomy 25% 1 vs 2 3.2 p<0.10 Autotomy 25% 3 vs 4 0.5 p<0.5 Intact 25% 1 vs 2 0 p<0.99 Intact 25% 3 vs 4 0 p<0.99 Autotomy 33% 1 vs 2 0.5 p<0.5 Autotomy 33% 3 vs 4 0 p<0.99 Intact 33% 1 vs 2 1.5 p<0.25 Intact 33% 3 vs 4 2.29 p<0.25 Autotomy 50% 1 vs 2 3.125 p<0.10 Autotomy 50% 3 vs 4 0.57 p<0.5 Intact 50% 1 vs 2 1.5 p<0.25 Intact 50% 3 vs 4 0.25 p<0.75 Autotomy 25% 2 vs 3 0 p<0.99 Autotomy 25% 4 vs 5 0 p<0.99 Intact 25% 2 vs 3 1.33 p<0.25 Intact 25% 4 vs 5 2.7 p<0.10 Autotomy 33% 2 vs 3 0.167 p<0.75 Autotomy 33% 4 vs 5 0.125 p<0.75 Intact 33% 2 vs 3 0.1 p<0.75 Intact 33% 4 vs 5 0.5 p<0.5 Autotomy 50% 2 vs 3 0 p<0.99 Autotomy 50% 4 vs 5 2.25 p<0.25 Intact 50% 2 vs 3 0.167 p<0.75 Intact 50% 4 vs 5 0.5 p<0.5

77

Figure 2.1: Measurement of error angle of orientation to a prey stimulus. The difference between the prey vibration stimulus angle (β) and the spiders responding orientation angle (α) was used to calculate the angle of error (γ) (Hergenroder and Barth, 1983). In this illustration, the spider has oriented farther past the stimulus then it needed to.

78

A) B)

4 Man 4 * Man Int Int 3 3

2 2

1 1 # of crickets consumed

# of crickets consumed 0 0 Trial 1 Trial 2 Trial 3 Trial 4 Trial 1 Trial 2 Trial 3 Trial 4 (Int) (Aut) (Reg 1) (Reg 2) (Int) (Aut) (Reg 1) (Reg 2)

C) D)

Prey capture at 6 hours Prey capture at 24 hours Man Man 4 4 Int * Int 3 ** 3

2 2

1 1

0 0 # of# crickets consumed Trial 1 Trial 2 Trial 3 Trial 4 # of crickets consumed Trial 1 Trial 2 Trial 3 Trial 4 (Int) (Aut) (Reg 1) (Reg 2) (Int) (Aut) (Reg 1) (Reg 2)

Figure 2.2: Comparisons of prey capture over 4 trial periods by spiders (manipulated and intact). For manipulated spiders: (Int=intact, Man=manipulated, Reg 1=partial regeneration, Reg 2= full regeneration. *denotes trials where there was a significant difference in prey capture rate between groups according to One-Way ANOVAs. A) 33% weight group at 6 hrs. B) 33% weight group at 24 hrs (trial 2:F1,1 =5.7993, p=0.0241). C) 50% weight group at 6 hrs (trial 2:F1,1:5.6006, p=0.0251; trial 3: F1,1=8.4933, p=0.0069). D) 50% weight group at 24 hrs (trial 3: F1,1=8.934, p=0.0057).

79

A) B)

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% % of orientation of % 10% 10%

0% orientation correct of % 0% Int Aut Reg 1 Reg 2 Int Aut Reg 1 Reg 2

Figure 2.3: Responses of S. ocreata to vibrations from a prey stimulus in the absence of visual cues. A) Percentage of spiders which oriented. B) Percentage of spiders which oriented correctly (with error angle less than stimulus angle). Int= intact, Aut= autotomy, Reg 1= regeneration 1, Reg 2= Regeneration 2.

80

160 140

120

100 80

60 40

20

(seconds) latency Orient 0 Int Aut Reg 1 Reg 2

Figure 2.4: Mean (±SE) latency to orient to a vibratory prey stimulus by spiders with foreleg intact (int), autotomized (aut), regenerated partially (reg 1), or regenerated fully (reg 2). One way ANOVA for natural log transformed data: F3,198=0.9236, p=0.4304. Spiders which did not orient were included as the maximum time of 300 seconds.

81

100 90 80 70 60 50 40 30 20 10

Mean error angle (degrees) 0 Int Aut Reg 1 Reg 2

Figure 2.5: Mean (±SE) error angles of orientation to a vibratory prey stimulus for spiders with a foreleg intact (int), autotomized (aut), regenerated partially (reg 1), and regenerated fully (reg 2). Results of a Kruskal-Wallis test for ranked mean angles showed no significant difference between 2 groups (X 3=1.5658, p=0.6672).

82 General Conclusions:

Impact of leg loss/regeneration on fitness as measured by body condition and

growth/development

S. ocreata have a high frequency of autotomy in the field (13-19%) which is at the upper

range of percentages reported for other species (Foelix, 1996). In this study autotomy was found

most frequently at the first and fourth legs. The position of legs autotomized most often may

vary by species (Uetz et al, 1996; Brueseke et al 2001, Weissmann and Vollrath, 1998). These

differences could be indicative of differences in the spider’s behavior in response to attack or

differences in types of predators the spider is exposed to. However legs could be lost more often

at one position simply due to differences in leg length (Spivak and Politis, 1989).

Autotomy/regeneration of limbs was correlated with decreased size, weight and body condition in field caught S. ocreata. These reductions were probably due to decreased foraging success (Beck and Connor, 1994; Jakobs et al, 1996; Uetz et al, 1996). In the web spinning spider Zygiella x-notata autotomized individuals showed reduced size after molting when living

in all areas but those with very high prey density (Weissmann and Vollrath, 1998). Differences in prey density by area were not measured in the current study. However there may have been

significant differences in prey availability between years as the spiders demonstrated significant

yearly differences in size and body condition regardless of autotomy. Further studies addressing prey availability in the field and its correlation with size, weight, and body condition in intact and autotomized/ regenerating spiders might elucidate these effects further.

Growth and development time were affected by regeneration differently for spiders in the lab depending on their rearing background. Completely lab reared spiders tended to take longer to molt when regenerating, but after molting did not differ in weight from intact spiders. In

83 contrast, field caught spiders brought into the lab were more likely to have no difference in molt

interval but were smaller after molting than intact spiders. This tradeoff between molt interval

and size bears further study. Smaller spiders generally are less fit because females are less

fecund, and males are less attractive to females or are out-competed by larger males (Wise, 1975;

Beck and Connor, 1992; Simpson, 1995; Spence et al., 1996, Uetz et al, 1996, Dodson and

Schwaab, 2001). However, having a prolonged molt interval prior to maturity may be

detrimental to male spiders as they lose potential chances to mate during that time. Further

studies would need to address what factors, e.g., nutrition, competition, etc. affect the tradeoff

between size and molt interval in regenerating S. ocreata.

In the current study the negative effects of regeneration on growth and development time in the lab seemed limited to the first instar after autotomy. The spiders were able to increase growth or decrease molt interval during the second instar after autotomy to compensate for previously slowed growth or development. It is unclear however whether the effect would be the same for spiders in the field which were under variable feeding conditions. Furthermore, these effects on growth or development time could be particularly detrimental to spiders which underwent autotomy in a penultimate stage and could therefore only molt one more time.

Regeneration of multiple legs had greater effects on growth and development than regeneration of a single leg. It is therefore not surprising that most of the spiders captured in the field were missing only a single leg. It could be that loss of multiple legs does not occur often, or if it does results in survival costs that few spiders are able to overcome. Other studies have shown that effects of autotomy on foraging occur only after multiple leg loss has occurred in some harvestmen (arachnids) (Guffey, 1999) and crabs (Juanes and Smith, 1995).

84 The effects of autotomy and regeneration on prey capture rate and efficiency and sensory

detection.

The results of the prey capture rate study showed that in a semi natural lab mesocosm

autotomy and in some cases the initial stage of regeneration had a negative impact on prey

capture. A previous study by Amaya et al (2001) showed that larger autotomized adult S.

ocreata had more trouble capturing prey. However, in the present study all of the spiders which were penultimate at the third trial and thus tended to be larger had a lower prey capture rate regardless of manipulation. Therefore, in addition to weight being a factor it is also possible that autotomy may affect spiders differently depending on their physiological stage of development.

All other autotomy studies thus far have focused on autotomy in adult spiders (but see

Weissmann and Vollrath, 1998). Follow up studies are necessary to determine the importance of the interactions between autotomy and spider weight and/or stage of development in juveniles.

Cricket weight, as expected, also had an impact on prey capture rate, with the larger group of crickets being captured at a slower rate. Additionally, the effects of autotomy (and regeneration) on prey capture rate appeared at 6 hours for spiders given larger weight crickets, but not until 24 hours for the spiders in the smaller cricket weight group. This autotomy x cricket weight affect was also somewhat expected as a previous laboratory study showed that larger crickets were harder to capture for autotomized individuals (Brueseke et al, 2001).

The laboratory study showed no significant effects of autotomy/regeneration on the physical aspects (efficiency) of prey capture. These results match those of other studies conducted on adult wolf spiders (Amaya et al, 2001; Brueseke et al, 2001). There was some effect of trial period, with spiders becoming more efficient at capture (decreased attack latency) as trials progressed. This is not entirely unexpected as other studies have shown that spiders can

85 increase prey capture efficiency with experience (Heiling and Herberstein, 1999; Morse, 2000).

However, these effects were only present in the first three trials (before the spider molted). In

the fourth and fifth trials (spiders molted before each trial) efficiency of prey capture decreased

again as the spiders were given larger crickets in proportion to their new weights.

We did not observe significant effects of autotomy or regeneration on the ability of

spiders to orient to a vibratory prey stimulus. S. ocreata generally relies heavily on visual as

well as vibratory cues for hunting (Persons and Uetz, 1996). If spiders are able to use both types of cues to detect their prey, the impact of any impairment due to missing or regenerating legs is probably slight. Future studies should focus on groups of spiders which hunt with more reliance on vibratory cues and are therefore more likely to be significantly impacted by losing or regenerating a leg.

The lack of any negative effects of autotomy and regeneration on prey capture efficiency and sensory abilities would seem to indicate that S. ocreata is able to compensate for missing or regenerating a leg while foraging. It is possible that the spiders are able to utilize their second leg for some of the functions that the missing/regenerating first leg would normally carry out.

However, autotomy and in some cases regeneration decreased the rate of prey capture in the semi-natural habitat (shoeboxes of leaf litter). Also, field caught spiders that were missing or regenerating a leg generally had lower size, weight, and body condition than intact spiders. This seems to indicate that some foraging cost of autotomy and/or regeneration is present. No other studies to my knowledge have examined the effects of autotomy and regeneration on hunting spiders in a field setting. It is possible that spiders are able to compensate for autotomy or regeneration of a leg under controlled laboratory conditions but that other factors in the field may interact with leg loss to make it more detrimental there. Further studies should be run to address

86 the many factors (such as competition from conspecifics and chances of predation) that have the

potential to interact with leg loss/regeneration and affect prey capture in natural populations.

Laboratory studies showed that regeneration of a leg in S. ocreata had some growth or

developmental costs in addition to potential foraging costs. In some web spinning spiders

autotomy occurs but the ability to regenerate has been secondarily lost (Vollrath, 1990). This

would be expected to have occurred in some hunting spiders as well if the benefits of

regeneration did not outweigh its costs. Because regeneration is energetically costly in the

hunting spider S. ocreata, autotomy should therefore be costly as well despite its obvious

survival benefits. Decreases in foraging due to autotomy could lead to decreases in condition, size, and weight as juveniles which have the potential to affect reproductive fitness as adults

(Wise, 1975; Beck and Connor, 1992; Simpson, 1995; Spence et al., 1996, Uetz et al, 1996).

Taken together, the results of these studies suggest that autotomy of a single leg may be more costly in hunting spiders than previously thought. Insights into the foraging, growth and developmental costs of regeneration were provided as well. Overall a better understanding was gained for the costs of autotomy and regeneration, providing evidence for why regeneration has been maintained over evolutionary time.

87 References:

Amaya, C., Klawinski, P, Formanowicz, D. 2001. The effects of leg autotomy on

running speed and foraging ability in two species of wolf spider (Lycosidae).

American Midland Naturalist. 145: 201-205.

Beck, M., Connor, E. 1992. Factors affecting the reproductive success of the crab spider

Misumenoides-Formosipes- the covariance between juvenile and adult traits. Oecologia.

92 (2): 287-295.

Brueseke, M., Rypstra, A., Walker, S., Persons, M. 2001. Leg autotomy in the wolf

spider Pardosa milvina: a common phenomenon with few apparent costs. American

Midland Naturalist. 146:153-160.

Dodson, G., Schwaab, A. 2001. Body size, leg autotomy, and prior experience as factors

in the fighting success of male crab spiders, Misumenoides formosipes. Journal

of Insect Behavior. 14 (6): 841-855.

Foelix, R. 1996. Biology of Spiders. pp. 213-233. Oxford University Press. New York, NY.

Guffey, C. 1999. Costs associated with leg autotomy in the harvestmen Leiobunum

nigripes and Leiobunum vittatum (Arachnida: Opiliones). Canadian Journal of Zoology.

77:824-830.

Heiling, A.; Herberstein, M. 1999. The role of experience in web-building spiders

(Araneidae). Animal Cognition. 2: 171-177.

Persons, M., Uetz, G. 1996. The influence of sensory information on patch residence

time in wolf spiders (Araneae: Lycosidae). Animal Behavior. 51:1285-1293.

Jakob, E., Marshall, S., Uetz, G. 1996. Estimating fitness: a comparison of body

condition indices. Oikos. 77: 61-67.

88 Juanes, F., Smith, L.D. 1995. The ecological consequences of limb damage and loss in

decapod crustaceans: a review and prospectus. Journal of Experimental Marine Biology

and Ecology. 193(1-2): 197-223.

Morse, D. 2000. The effect of experience on the hunting success of newly emerged

Spiderlings. Animal Behaviour. 60: 827-835.

Persons, M., Uetz, G. 1996. The influence of sensory information on patch residence

time in wolf spiders (Araneae: Lycosidae). Animal Behaviour. 51:1285-1293.

Simpson, M.1995. Covariation of spider egg and clutch size: the influence of foraging

and parental care. Ecology. 76(3): 795-800.

Spence, J., Zimmerman, M., Wojcicki.1996. Effects of food limitation and sexual

cannibalism on reproductive output of the nursery web spider Dolomedes triton

(Araneae: Pisauridae). Oikos. 75:373-382.

Spivak, E., Politis, M.1989. High incidence of limb autotomy in a crag population from

a coastal lagoon in the province of Buenos Aires, Argentina. Canadian Journal of

Zoology. 67: 1976-1985.

Uetz, G., McClintock, W., Miller, D., Smith, E., Cook, K. 1996. Limb regeneration and

subsequent asymmetry in a male secondary sexual character influences sexual

selection in wolf spiders. Behavioral Ecology and Sociobiology. 38:321-326.

Vollrath, F. 1990. Leg regeneration in web spiders and its implications for orb weaver

phylogeny. Bulletin of the British Arachnological Society. 8(6):177-184.

Weissmann, M., Vollrath, F. 1998. The effect of leg loss on orb-spider growth.

Bulletin of the British Arachnological Society. 11 (3) 92-94.

89 Wise, D. 1975. Food limitation of the spider Linyphia marginata: experimental field

studies. Ecology. 56: 637-646.

90