Determinants of Morphological Characteristics, Thermogenetic Heating and Scarabaeidae Pollination in Xanthosoma robustum Araceae Joshua A. Craig Department of Biology and Latin American Studies, U niversity of Pittsburgh, PA ___________________________________________________________________ ABSTRACT Xanthosoma robustum Araceae relies upon thermogenic heating just after sunset to attract its mutualist Scarabidae pollinator. All morphological characteristics measured number of leaves, leaf size, stalk height, spathe length, spadix length, and kettle diameter were correlated; demonstrating larger plants have more leaves, larger leaves, taller stalk heights, taller spathes, tal ler spadices and wider kettles. However, no measured characteristic was found to influence Scarabaeidae visitation. Maximum temperature of an inflorescence was found to be positively correlated with leaf size and negatively correlated with spadex height t test p values 0.0007 and 0.0150, respective ly . However, no recorded temperature measurements were found to influence Scarabidae visitation. RESUMEN Xant hosoma robustum contra con calentamiento tÃ©rmico despuÃ©s de la puesto del sol para atra er los escarabajos mutualistas de como polinizaros. Todos las caracterÃsticas morfolÃ³gicas medidas fueron correlacionadas, mostrando que las pla ntas mÃ¡s grandes tiene n mÃ¡s hojas, hojas mÃ¡s grandes, tallos mÃ¡ s altos, inflorescencias mÃ¡ s altos, espÃ¡dices mÃ¡s altos, y las bases mÃ¡ s anchas. Sin embargo, ninguna caracterÃstica m edida se encontrÃ³ que influyo en la visitaciÃ³ n de los escarabajos. La temperatura mÃ¡xima de una inflorescencia se correlacionado positivamente con el tamaÃ±o de la hoja y f ue correlacionado negat ivamente con altura de la inflorescencia . Sin embargo, ninguna medida de la temperatura se encontrÃ³ que influye en la visitaciÃ³n de los escarabajos. INTRODUCTION Aroids Araceae, exhibit an unusual characteristic of thermogenetic respiration in order to attract beetle pollinators, specifically Cyclocephala spp. Coleoptera: Sc arabaeidae. The plants release heat through their inflorescence when stored carbohydrates are metabolized to ATP and NADH is converted to heat Mauseth 1995. The central agent t hought to be responsible for heat production in at least one aroid, Sauromatum guttatum is salicylic acid, which may alter the respiration in the florets , although more research is needed to conclude this as fact Raven et al. 1999. Some aroids have the ability to produce extraordinary amounts of heat and generate some of the highest temperatures ever recorded in plants, over 45 Â°C Gottsberger and Gottsberger 1991. A
fl eshy spathe surrounds a spadix, housing several hundred vertically stratified white male florets atop a ring of more elongated, flat and white sterile florets. Below these, on the spadix are sticky, orange female pistillate flowers within a red "kettle'' t hat is attached to the spathe Figure 1. In Philodendron selloum, the sterile male florets produce most of the heat, a nd the rates of respiration are so high, they approach those of flying birds Seymour 1990. The heating of the spadix seems to volatize amines and indoles to attract pollinators by emitting an easily registered odor across a greater area than if not heated Raven et al. 1999. Usually the odor emitted by carrion beetle pollinated aroids is putrid , but Xanthosoma odor is actually quite sweet and pleasant to the human senses. Aroids showing thermogenesis are known to be pollinated by Scarabaeidae beetles, whose sense of smell is more highly developed than their visual perception Gottsberger and Got tsberger 1991. This suggests the volatilization of odors, and not just heat production alone, is a critical feature for ensuring Scarabaeidae visitation, pollination and fruit production. The sequence of mutualistic interactions between Xanthosoma spp. an d Cyclocephala spp. are quite evolved and interesting. On the first night of thermogenesis, based upon limited understanding through few measurements, the plant reaches a peak temperature of 40 42 Â°C between 18.00 and 18.30 hours Goldwasser 1987. The scar abs travel over great distances at dusk to an inflorescence and land on the spathe either in response to heat or the sweet smell associated with it. The scarabs then crawl into the chamber within the kettle and bring with them pollen from the last visite d flower. They then begin to eat the sterile parts of the inflorescence, and in the process transfer the pollen from the other plant to the female pistillate flowers of the newly visited flower. Apparently, the female flowers are only receptive to pollinat ion transfer at this time within 24 hours. As the temp erature of the inflorescence cool s over the course of the evening, the scarabs feed and mate within the chamber. The second evening, the inflorescence reaches a maximum temperature of only around 34 Â°C, at which time it releases its pollen onto the very sticky scarabs as they travel up and down the inflorescence. The beetles then leave individually on the third day, full of pollen, in search of another flower for food, shelter, and mates. Consequently , the inflorescence does not heat at all on the third day, thereby ensuring the scarabs leave in a timely fashion for more efficient pollen transfer at another flower Goldwasser 1987. At Monteverde, Costa Rica, tw o species of scarab beetle poll inators on Xanthosoma have been observed: Cyclocephala nigerrima, a Costa Rican species which is nearly black and C. sexpunctata, a species common from Mexico to Bolivia which is light brown with dark spots on both elytra Goldwasser 1987. In the early 1980's in Monteverde, the ratio between species was Cyclocephala nigerrima outnumbering C. Sexpunctata 18:1. The mean number of scarabs per inflorescence was observed to be 6.8 SD =1 . 31; n =73, with a maximum of 38 Goldwasser 1987, Previous studies determined that plants of lower density and lower altitude, and with taller inflorescences were found to have more scarab visitors Sica 1999. Others have shown high variation betwee n different populations Goldwasser 1987. A similar study has ruled out Muridae and Aphidae as important pollinators, even though hundreds are found commonly on almost all inflorescences and some Xanthosoma populations lack scarabs altogether Cochoran 19 97. However, no research has investigated which morphological characteristics influence thermogenic heating in natural Xanthosoma populations, nor have these characteristics been linked to
reproductive success by the two species of Cycl ocephala pollinators found in Monteverde. I hypothesized that a larger plant would have larger floral characteristics spathe length, spadix length, kettle diameter, floret numbers, and greater mass of inflorescence. This would correlate significantly with the max imum temperature achieved by an inflorescence thermogenesis, ultimately correlating with the number of scarab visitors to an inflorescence. The purpose of this study is to determine the morphological plant characteristics that influence thermogenic heati ng and investigate what attributes of thermogenesis influence scarab visitation. MATERIALS AND METHODS The area of study was on the property of Skywalk/ Sky trek , located near the Santa Elena Reserve, Santa Elena, Costa Rica. The altitude is 1550 meters consisting of Lower Montane Wet forest. Xanthosoma robustum plants were observed between April 25 to May 5, 2001. Each day, I checked a population of plants for flowers that were going to be in their first night of thermogenic heating. First day flowers were distinguished by the following common temporal characteristics. Immature flowers have a greenish tint to the spathe and are completely sealed, preventing the entrance of visitors. However, fi rst day flowers have a whiter appearance to the outer spathe, which has slightly separated from the spadix, but still is somewhat closed. The occurrence of a few Muridae and Aphidae flies are observed if the spathe is physically pulled apart. The common "v irgin" appearance to the whit e spadix implies that thermogen t et ic heating had not yet occurred, and no scarab visitation had taken place. Second day flowers have completely separated spathes, exposing the male florets entirely, and will usually be covered with hundreds of Muridae, Aphidae and other Diptera. The spadix on a second day flower usually loses its pure white color due to the massive visitation of the hundreds of insects, especially when they transfer small amounts of orange, sticky female florets to the white male florets. A clear indicator for third day flowers is the release of pollen on the male florets. For each first night flower located, I used a tape measure to determine in cm plant stalk height from the ground to the base of the kettle, maximum leaf length, spa the length, and spadix length. I also counted the number of leaves and measured kettle diameter using a caliper. For three flowers every night, a dissecting pin was used to make a small hole in the spadix, into which an external digital Hobo XT thermometer probe was then inserted. Internal spadix temperature and ambient temperature was recorded automatically every minute. The Hobo thermometers were then digitally programmed or "launched" using a Macintosh laptop computer with the Box Car Pro PPC 3.5d2 program. Upon return to the same flower the following day, the scarab number was recorded and the temperature from the previous night downloaded into the laptop computer and analyzed. Flowers for wh ich temperatures were not recorded, the numbers of scarabs were counted, and the previously mentioned parameters were measured. In total, I collected morphological measurements on 179 individuals, including 28 individuals for which temperature data were ob served. To measure spadix weight and floret number, the inflorescences were cut from the parent plant on the second day, matted, pat into Ziploc bags, and placed into a freezer. Wet weight 0.01 g was taken for the sterile and male florets of each inflore scence using a digital scale. They were dried for two days using two 100 watt light bulbs in a dryin g box.
During this time, as the inflorescences were in the process of drying, all male and sterile parts of the individual inflorescences were counted. Once completely dried, a dry weight 0.01 g was taken for the sterile and male florets using a digital scale. RESULTS Correlation matrices were computed to determine a correlation between mea sures. Next, multiple regression tests were conducted and the best fit model was determined with all non correlated independent variables. Where two independent variables were correlated, only one was selected, random ly for the multiple regression analyses. Plant Attributes Xanthosoma robustum varied greatly in size throughout the study area. Stalk height averaged 135 cm SD = 56; n = 179, with the shortest flowering plant having a stalk height of 23 cm and the tallest 290 cm from the ground. The majority of plants had three leaves 51% with the occasional two leafed 26% or four leafed plant 20%. The remaining plants had just one leaf , but no pla nts had more than four leaves. Maximum leaf size showed a highly varied range from 34 cm to 165 cm, with a mean of 86 cm SD = 19. Ave rages of sixteen first night flowers were located each day. Floral attributes were less variable. Spadix length ranged from 11 cm to 23 cm mean = 16.7; SD = 2.1, while spathe length range d from 12 cm to 33 cm mean = 2 1 .5; SD = 3.1. Likewise, kettle diameter ranged from 3.01 cm to 5.64 cm mean = 4.40; SD = 0.47 . Correlation of Plant Attributes As might be expected, all measured plant m orphological characteristics of Xanthosoma robustum were significantly and positively correlated, including stalk height, number of leaves, maximum leaf size, spathe length, spadix length, and kettle diameter Table 1. Larger plants tended to have these larger morphological attributes, while smaller traits were seen in smaller plants. Temperature Measures Of the thirty first night spadices for which temperature was recorded, twenty eight heated as predicted. One spadix heated to a lower temperature 28.70 Â°C on the first night and a hig her maximum temperature 31.93Â°C on the second night; For this irregular pattern, this spadix was eliminated from all analyses. One other spadix was improperly measured because it was a second night inflorescence; and was also eliminated from all analyses. For the tw enty eight spadices with normal heating patterns Figure 2; mos t 61 % were in a peak range of 36Â°C to 39Â°C. The mean maximum temperature of an inflorescence was 36.22Â°C SE= 8.2; n = 28, with the lowest recording Of 30.19Â°C to t he highest temperature o f 40.59Â° C. The duration of time an inflorescence heated above 30Â°C had a mean of 89.9 minutes SE = 1.6; n = 28 and a widely varied range of 7 minutes to 206 minutes. The duration within 1Â°C of the maximum temperature had a mean of 27.1 minutes SE = 1.6 and a smaller range of 14 minutes to 43 minutes. The maximum t emperature and duration above 30 Â°C were correlated r = 0.889; p = <0.0001; n = 28 but this was not the
case for maximum temperature and duration within 1Â°C of maximum temperature r = 0.232; p = 0. 2535; n = 28. Mean ambient temperature recorded at the time of maximum inflorescence temperature was 17.11Â°C SE = 0. 18; n = 10 with a range of 16 . 38Â° C to 17. 90Â° C. Plant Size and Temperature The stalk height, number of leaves, maximum leaf size, spathe length, spadix length, and kettle diameter were tested for their affects on temperature, either maximum temperature or duration above 30Â°C. The best multiple regression model demonstrated that both maxim um leaf size and spadix length were important determinants of spadix temperature y = 0.112x Â 0.635x Â‚ + 37.150; maximum leaf size: x Â = O.112 ; p = 0.0007, spadix length: x 2 = 0.635; p = 0.0150, overall p = 0.0027, n = 28. Whereas plant size increases and spadix length decreases, spadix temperature increases. Therefor e, the warm est spadices were found on large individuals with unusually small spadices. The duration over30Â°C was also significantly correlated with maximum leaf size and spadix length. However, when paired separately with maximum temperature in a simple regression model, maximum leaf size was the only morphological plant characteristic that demo nstrated a significant correlation Figure 3. Oth er Floral Characteristics and Temperature Wet wei ght varied greatly from 17.45 g. to 52.17 g. mean = 29.06; SD = 9.18, n = 23. Dry weight was less variable, from 2.85g to 8.44 g mean = 4.50; SD = 1.47; n = 23, but was highly correlated with wet weight y = 1.753 + 6.064; p = < 0.0001; R 2 = 0.949; n = 22 . The total number of florets r anged from 567 to 1020 mean = 824; SD = 129. Sterile florets had a range of 73 to 173 mean = 120; SD = 27 while male florets h ad significantly more, ra nging from 485 to 891 mean = 70 6; SD = 113. There was a significant correlation between wet weight, dry weight, total number of florets, ma le florets, and sterile florets Table 2. Therefore, male florets and dry weight were chosen for a multiple regression test with maximum temperature. The best model suggested no attribute of floral characteristics was important to predicting temperature y = 0.003x 1 + 0.160x Â‚ + 34.524; x 1 = male florets, p = 0.4921; x Â‚ = dry weight, p = 0.5968; overall p = 0. 4338, n = 22. Plant Size, Floral Characteristics, Temperature and Scarab Visitation A total of 194 scarab beetles were observed mean = 1.0; SD = 2. 0; n = 179. The mean scarab visitation per first night inflorescence for which temperature was taken was 1.0 SD = 2.0; n = 28, and ranged from zero to nine scarabs. Many flowers were not visited 54%, of those tha t were, a mean of three scarabs SD = 2 were found. Spadix length; maximum leaf size, dry weight, and maximum temperature were tested for their affect on scarab number. The best model suggested that none of these measured characteristics were important in predicting th e number of scarabs y = 0.022x Â 0.066x 2 0.005x 3 0.100x Â„ + 2.465; X Â = maximum temperature, p = 0.9125; x 2 = dry weight, p = 0.7814; X Âƒ = m aximum leaf size, p = 0.8484; x Â„ = spadix length, p = O. 5899; overall p = 0.7913, n = 22. Two different species of scarab beetles were found; 190 individuals 98% were Cyclocepha la sexpunctata and only 4 individuals 2% were C. n igerrima.
DISCUSSION My research did not support my hypothesis with respect to plant size and tem perature, however did support predicted morphological traits. The results of my study suggest that there are morphological plant characteristics of Xanthosoma robustum that help predict certain aspects of their pollination sequence. All measured morphological attributes were correlated. This finding may aid in predicting how old a plant may be and based on its age r what its relative: reproductive success may have bee n in the: past fruit setting. For instance, a larger plant will usually have larger floral characteristics which may help direct scarabs to an inflorescence after a scent has been recognized. The maximum leaf size and spadix length correlated in such a way that if a plant has larger leaves and smaller spadices, it will heat to a higher maximum temperature. However, maximum leaf size was the only morphological trait that significantly correlated with maximum temperature and may be one determining factor in morphological plant g rowth and energy expenderature. It is. possible, that if a larger plant has a smaller flower, it may need to compensate for the lack of visual attraction the scarabs need to land on the spathe by heating to a higher temperature as a means of helping guide scarabs to an inflorescence . Consequently, it may be that a larger plant produces a larger flow er f or visual attraction so as not to invest much more energy in excessive heat production. For the volatilization of the attractant indoles to occur, the inflorescence may only need to reach a given base temperature. At which point, the maximum temperature an d heating duration would not play such a key deterministic role. Previous findings in Philodendron selloum have suggested that it is the number of florets that is the unit of heat. Therefore, increased number of florets should equal an increased inflorescence temperature. This was not found to be the case in my study; this may be due to the variance between aroids or the variance between Xanthosoma populations. An exhaustive study is needed to further predict the determinants of temperature and its relation to scara b visitation. Although populations have been shown to have high variance with respect to scarab populations, my study site had a mean of 1.0 scarabs per inflorescence compared to 6.8 fou nd by Goldwasser in the mid 1980 's. Although Xanthosoma seems to thrive along forest edge and disturbed habitat, the numbers of Scarabaeidae pollinators observed in recent studies have dropped considerably. What appears to be more striking is the drastic shift in Cyclocepha la species abundance. Goldwasser found in the mid 1980's, Cyclocephala nigerrima outnumbering C. sexpunctata 18:1. Sica 1999 found the opposite abundance with C. sexpunctata comprising 8 7% of the population. In my study, I found a drastic difference, agreeing with Sica's findings, except 98% were C. sexpunctata. There have been many changes t o the ecosystem of Monteverde in the last 15 years, including the encroachment of humans into and development of forest habitats coupled with climatic changes. This may be a contributing factor for the lack of scarabs and change in species relative abundance found in my study. Since C. sexpunctata is found in such a wide range, they may be better adapt ed for these recent disruptive changes. More research must be done on the relative abundance of these scarab pollinators, especially with respect to C. nigerrim a, if its population size is indeed declining at such drastic rates, More studies must
be don e o n different Xanthosoma populations in more diverse areas to find a correlation with interspecies Scarabae idae pollinators and thermogenic heating. My research suggests a good example of how humans are impacting their natural world around them. Although the extinction of one beet le w ould not be a world news flash, it certainly would have many direct effects on many forest organisms. Without trying to find human correlated negat ive effects on our environment , I could not escape the possibilities in my researc h. On a worldwide scale, we all must recognize the effects we have on our environment, and do all w e can to try and conserve all species and habitats. ACKNOWLEDGEMENTS I would like to give an extra special thanks to Alan and Karen Masters, Mauricio Garcia, Tim Kuhman and Andrew Rodstrom for their patience and help with my research; everyone at Skywalk/Skytrek for allowing me to study on their property, giving me transpor tation each day and preserving the forest through ecotourism and awe inspiring canopy tours; the Valera Mendez family for being the most incredibly helpful and loving family in Costa Rica; la EstaciÃ³n BiolÃ³gica de Monteverde; all of my friends and fellow C IEE students, especially the unforgettable Amy Hanna for being my best friend and support through it all; and lastly all of those working each day to preserve our glorious Mother Earth all around the world. LITERATURE CITED Cochoran, Carolyn. 1997. Elephan t Ear Xanthosoma jaquinii: Araceae Pollinators and the Effect of Heat as an Attractant. Council on International Education and Exchange. Goldwasser, Lloyd Paul 1987. Mutualism and its Ecological and Evolutionary Consequences. Doctoral Dissertation: University of California, Berkeley. Gottsberger, Gerhard and I ise Silberbauer Gottsberger. 1991. Olfactroy and Visual Attraction of Erioscelis emarginata Cyclocephalini, Dynastinae to the Inflorescences of Philodendron selloum Araceae, Biotropica 231; 23 28. Mauseth, James. 1995. Botany: An Introduction to Plant Biology. Saunders College Publishing. Raven, Peter, Evert, Ray and Eichhorn, Susan. 1999. Biology of Plants. Sixth Edition. Worth Publishers. Sica, Audrey. 1999. The Effectiveness and Abundance of Scarabid Pollinators of Xanthosoma sp. Araceae. Council on International Education and Exchange. Seymour, Roger. 1991. Analysis of Heat Production in a Thermogenic Arum Lily, Philodendron selloum, b y Three Calorimetric Methods. Thermochimica Acta. 193; 91 97. Zar, Jerrhold H. 1984. Biostatistical Analysis Second Edition. Prentice Hall Inc., Englewood Cliffs, New Jersey.
Table 1 . Correlation matrix demonstrating all measured morphological plant characteristic values as being significantly correlated. All bold values show statistical correlation absolute value > 0.145. Stalk Height Number Of leaves Maximum Leaf size Spathe Length Spadex Length Kettle Diameter Stalk Height 1.00 0.161 0.239 0.306 0.293 0.320 Number of Leaves 1.00 0.421 0.380 0.357 0.295 Maximum Leaf Size 1.00 0.572 0.535 0.535 Spathe Length 1.00 0.806 0.502 Spadex Length 1.00 0.512 Kettle Diameter 1.00 Table 2 . Correlation matrix demonstrating all floret numbers and measured weight values as being significantly correlated. All bold values show statistical correlation absolute value > 0.404. Wet Weight Dry Weight Total Florets Male Florets Sterile Florets Wet Weight 1.00 0.974 0.584 0.559 0.437 Dry Weight 1.00 0.651 0.630 0.470 Total Florets 1.00 0.982 0.659 Male Florets 1.00 0.509 Sterile Florets 1.00