GROWTH OF THE ORB WEAVER, ARANEUS DIADEMA TUS, AND CORRELATION WITH WEB MEASUREMENTS*

It is a well-known fact that within any population of spiders of similar .age there is considerable variation in the size of individual spiders of the same species. In literature as early as 189o, McCook has observed this variation and repeated observations (Comstock, I94O; Savory, 1928) have verified this phenomenon. Although observations are frequent, explanations are few. Bristowe (I958) cites differences in feeding as a reason for differential size, but the reference is made merely in passing and to the authors’ knowledge is not elaborated upon elsewhere. This paucity of explanation lends itself to further analysis of the factors contributing to the phenomenon of differential size. Our purpose in this paper is to isolate some of the factors which contribute to differential size in zlraneus diadematus Clerck (for identification .of species, see Levi, 1971), .and to elaborate upon certain of these factor.s as we are able. Corresponding with differential size, in an orb-weaver such as ztraneus diadematus, differential growth is also manifested in changing dimensions of the web. That large differences in dimensions exist betwen the individual webs of spiders is also a well-known fact. An attempt to clarify so.me of the factors influencing web changes is also made.


INTRODUCTION
It is a well-known fact that within any population of spiders of similar .age there is considerable variation in the size of individual spiders of the same species. In literature as early as 189o, McCook has observed this variation and repeated observations (Comstock, I94O;Savory, 1928) have verified this phenomenon. Although observations are frequent, explanations are few. Bristowe (I958) cites differences in feeding as a reason for differential size, but the reference is made merely in passing and to the authors' knowledge is not elaborated upon elsewhere. This paucity of explanation lends itself to further analysis of the factors contributing to the phenomenon of differential size.
Our purpose in this paper is to isolate some of the factors which contribute to differential size in zlraneus diadematus Clerck (for identification .of species, see Levi, 1971), .and to elaborate upon cer-1973] Benforado KistlerAraneus diadematus 91 cocoon were placed in a separate rearing box. The spiderlings were kept in these boxes, living on a communal web with a consta.nt supply o1: loose gnats in the box, until they began to build individual webs .approximately three weeks a1:ter hatching. As each animal built her first web she was removed 1:rom the rearing box and placed in an individual glass tube, approximately X 7 cm, with the ends o1: the tube stoppered with cotton. From the time the animals were placed in the tubes until onset o1: the experiment they were 1:ed approximately  week, by placing the gnats in the tube with the spiderling. The animals were watered by wetting the cotton with water daily. DS:RIBV:ION: Seven weeks ater hatching the two sets o spiderlings were each .epara.ted into three equal groups by means o1: a random numbers chart. No attempt was made to distribute males and 1:emales evenly. Although the growth (body weight) ot: males and 1:emales differs, it has been shown that the early growth o1: both sexes is alike (Witt et al.,97:). Beta.use o1: the short duration ot: the experiment and the difficulty in identi1:ying male spiders be1:o.re the last molt, distribution o1: males and 1:emales was le1:t to chance.
At the time ot: initial grouping the two .sets of spiderlings numbered twenty and thirty respectively. It was decided to 1:eed each o1: the three groups o1: each set according to a different eeding schedule: one group every day, one group twice weekly, and one group every ten days. Thus there were six groups, one 1:or each set o offspring on each eeding schedule. A1:ter one week o1: this procedure, however, it' was decided because o1: the small size ot: the groups to reduce the number o1: schedules to two, and the middle schedule was dropped and its members distributed randomly between the' lighter and heavier-1:ed groups. Data for animals that died or escaped during the course of the experiment were removed, so figures represent only animals observed or the duration of the experiment.
WEIGHING" Each spider in the heavy-fed groups was weighed once a week, to o. rag, while animals in the light-fed groups were weighed on the day of feeding and the day .after feeding. wB nNYss: After eleven da.ys of controlled differential feeding in the tubes, the spiders were transferred to aluminum and glass laboratory cages, 5o X 5o X o cm. At this time the animals were eight weeks old. From this time on the spiders began to. build webs. Photographs of webs were taken daily and analyzed (Reed et al.,965). Daily records of web building were kept and the. webs were destroyed daily with the thread left in the cage or the spider to digest. FEEDING: While in the tubes, the spiders were fed by placing a previously weighed de-winged housefly in the tube daily o.r every ten days. Those spiders that wo.uld not eat a housefly had three, to seven unweighed gnats placed in their tube. By visual inspection the following day it was determined whether the fly had been eaten. The remains of the eaten flies were then weighed to obtain an approximation of the amount .eaten by each spider. The spiders were watered by wetting the cotton every o.ther day.
After being placed in the cages, if the spider had a web, feeding was by means of placing the housefly in the web; if there was no web, we attempted to induce the animal to eat by placing the fly in front of its mouth. The heavy-fed spiders were offered at least one fly per day and more, if they would accept it. The light-fed group was fed one fly once every ten days. If on the day of feeding of the light-fed group a .spider would not eat, a note was made and the attempt repeated until successful. All spiders were. watered on Mondays, Wednesdays and Fridays by spraying a small amount of water in each cage.
MOLTS: From the onset of the experiment molts were recorded by date of the molt to give an indication of the. maturation of the animal.

RESULTS
FEEDING AND WEIGHT INCREASE: At the end o.f a period of five weeks the two feeding schedules resulted in two significantly dierent weight groups. This development is shown in Figure I, which illustrates the increasing dierence in weight between the two. groups. At the onset of differential feeding the mean weights of the two groups were alike, however, a T-test between the mean weights at the ,end of the experiment is significant at the o. 1% level.
An analysis of covariance was performed on the data. Because the original data was non-homogeneous, a transformation [log (x+ o)] was made (Winer, I962). The initial observation was used as a covariate in the .analysis of covariance. Because the analysis of covariance indicated no significant difference in the behavior (growth) of the two families, all figures are for both families combined. For the heavy-fed group the mean weight changed from 7.93 mg I.o4 o.n June 2 to 74.28 mg -+-IO.93 on July I7. The mean weight of the light-fed group changed from 6.4o mg -+o.98 Psyche [March-June on June 2 to I7.9I mg -+-.56 on July I3; there w,as a significant interaction between time and feeding schedule below the % level. FEEDING AND MATURATION: If the number of molts over tinae is taken as an indication of speed of maturation, then a relationship between feeding and rate of maturation can be .seen. During the period of differential feeding the number o.f molts of the heavy-fed and light-fed groups differed significantly at the 5% level. The heavy-fed group had a naean number of 3.0. molts while the light-fed group had a mean number of 2.3 molts.. These results are in agreement with the findings reported by Deevey (1949) with Latrodectus mactans (Fabricius) and indicate that in the laboratory with only food quantity as a variable, a relationship exists between the rate of weight increase and the rate of maturation.
INITIAL WEIGHT AND RATE OF GROWTH: From the beginning of the experiment we noted a wide variation of weights of the individual animals. At the onset of differential feeding individual weights ranged from I.I nag to 16.2 nag. In both the light-fed and heavyfed groups there existed a positive correlation betw.een initial weight .and final weight. F, or the light-fed group r o.7713 and fo, r the Measurements are divided into those which measure size (above the broken line) and those which measure regularity. Note the difference between the light-fed and heavy-fed animals in measures of web size at the late date.
While the heavy-fed gro.up increased in all size measures (see Fig. 2 Table   and are all reproduced to the same scale. Note that while both the heavy-fed, and light-fed animals began with webs of similar size, after two weeks of differential feeding the large, heavy-fed spiders' webs had increased in size while the webs ef the small, light changes that accompanied the growth of the animals. In measurements of web size both groups increased, with the heavy-fed group having a much larger increase as illustrated in Figure 2. In measurements of web regularity both heavy-ted and light-fed groups remained constant, as shown in figures ,o.f Table I.

DISCUSgION
The observed differential gro,wth and devel,o.pment in Araneus diadematus seems to be a unction of sever,al factors. Although an exposure to a greater than normal .supply o o.od generally results in iaster than normal growth and development, even within a group exposed to the same oo.d supply there seems to be a great variation in growth rates. Evidence of these differences is expressed in the increasing standard errors in Figure , and seems to be dependent upon individual a.ctor.s in the animals rather than environmental variations. Large differences in the amount eaten by individual animals in the laborato, ry existed and presumably exist in nature.
These differences seem to correspond to differences in the rate o growth in agreement with the findings of Turnbull in other species o spiders (Turnbull,96o,965). However, whether these difi:erences in the amount o ood eaten are due to differences in proficiency in prey-catching or to differences in appetite or so,me other t:actor in the animal is not clarified by our findings.
Another important factor influencing differential growth is the initial weight of the animal. Variations in initial weights xvithin a family are generally retained during the course of development. Although several possible reasons for different initial weights within a family have been given by others, the authors are reluctant to. offer any explanations..
In an orb web weaving spider such as Araneus diadematus the amount of food available to the animal is roughly equivalent to, the number of prey which become entrapped in the web. The number of prey entrapped in the web is in turn determined by a number of variabl.es such as web-site, size and fine structure of the web, and frequency of web building. Thus., it can be seen that the interaction of the variables resulting in differential size and growth is complex and can be divided into thos.e factors which influence the amount of food available to the spider and those actors which influence the spider's use of the food available to it. Repeated attempts have been made to explain web characteristics in terms of characteristics of the individual .spider (Peters.,I936). More common, however, has been the notation of changes in the form of the orb web during the life of the spider (Tilquin, I942; Savory, I952) and the attempt to relate these changes to changes i'n the animal (Witt and Ba.um,96o;Witt,I963;Reed et al.,I969). Because influencing factors vary concurrently, it is frequently difficult to assess the causes of changes in the form of the web. In our experiment we attempted to. isolate the effect of one. variable (weight) while minimizing the effect of a variable which normally changes concurrently (time). All animals used hatched on the same date, however, one gr.oup (the heavy-fed) gained co.nsiderable weight over the period measured, The web changes accompanying these weight increases are summarized in Table I. Because all of the animals were hatched on the same date, we conclude that increases in web size are due to differences in size of he animals resulting from differential feeding rather than differences in age. If appetite were a factor influencing web size, it would appear that the hungrier, light-fed animals would build a larger web in an attempt to catch more t:ood; ho.wever, this is not the case.
The relationship between f.ood and the web of a spider is a delicate .one. xZithout food, the spider's web-building ability diminishes,. but without a web there is no food (Peakall, I968). Thus, like a businessman, the spider faces the law ot: diminishing returns. It appears that the hungry spider chooses to conserve its resources Psyche [March-June rather than gamble on a larger web trapping more food. Early food deprivation experiments (Witt,963) show that the spider continues to build the same size web when deprived o food, but with less thread until finally a decreasing in web size occurs. Because our hungry (light-fed) animals were kept on a diet closer to a maintenance level than a deprivatim level, we observed no decreases in w.eb dimensions..
Feeding conditions in a natural environment vary more than those imposed in a controlled laboratory. Yet the spider is able to survive in these naturally diverse conditions because o its adaptability. In situations where there is little food available, the spider is. able to survive by growing at a slow rate and maintaining the same size web. Where food is abundant, the spider takes advantage of the situation, growing at a fast rate and increasing the size of its web.. The spider has developed a method for coping with a wide range of feeding conditio.ns. By varying its body and web growth, the spider can survive under the diver.se conditions imposed by nature, thus minimizing the necessity o,f seeking new food supplies and relocating the vceb. Our findings provide new insight into the spider as an example of an animal that adapts itself successfully to, its enviro.nment. UMMARY Spiders from two cocoons of ztran.eus diadematus were exposed to /ive weeks of two different feeding schedules: one group was offered large amounts (one housefly per day) o] food, the other group scarce (one fly every ten days) amounts. Although both groups increased in weight, weight gains of the heavy-fed group were significantly greater than those o.f the light-fed group, regardless of cocoon origin. Within each group there was a wide variation in the growth of individual animals, indicating the presence of factors other than ]ood supply; i.e. animals with extreme weights within a group at the onset remained the extremes.
In conjunction with increases in 'weight, over the three week period of observation, webs of the heavy-fed spider's showed an increase in size but not in regularity and shape in comparison to, webs of the smaller, light-fed animals of the same age which did no,t change. Such data suggest an increased chance of survival o.f the species through variations in rate of growth and maturation dependent on environmental factors.

/kCKNOWLEDGE'MENTS
This work was carried out in the laboratories o.f the North Carolina Department o.f Mental Health and was supported by Grant Number GB 25274 from the National Science Foundation to Peter N. Witt. The authors gratefully acknowledge the assistance of Dr. Peter N. Witt during all stages .and the. assistance of Mrs. Mabel B.
Scarboro during the period of laboratory work.