xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xlink http:www.w3.org1999xlink xmlns:xsi http:www.w3.org2001XMLSchema-instance
leader 00000nas 2200000Ka 4500
controlfield tag 008 000000c19749999pautr p s 0 0eng d
datafield ind1 8 ind2 024
subfield code a M39-00193
Quinn, Audrey L.
Perdida de endospermo, germinacin de semillas y crecimiento de las plntulas en semillas grandes en los rboles tropicales
Endosperm loss, seed germination, and early seedling growth in large-seeded tropical trees
Digitized by MVI
Endosperm damage was studied in Monteverde, Costa Rica to determine its frequency, extent, and effect on seed germination and early seedling growth. Endosperm damage was found to occur in 72% of seeds from Chione sylvicola (Rubiaceae) and 94% of seeds from Quercus costaricensis (Fagaceae). The extent of endosperm damage in Q. costaricensis was not related to the root and shoot weight of germinated seeds. Experimental endosperm removal was carried out in Persea americana and Cinnamomum paratriplinerve (both Lauraceae) seeds. The three week time period of the study was inadequate to yield significant C. paratriplinerve germination, but seeds with intact endosperm were much less likely to suffer fungal infection. Endosperm removal did not inhibit germination in P. americana seeds, even when 80% of the original seed weight had been removed. After 23 days in soil, most seeds were just beginning to produce roots and shoots, the majority of which were still enclosed in endosperm. At this time, P. americana seeds with a higher percent of endosperm removed had significantly heavier roots and shoots, suggesting a reduced or even inhibitory role of endosperm in initial seed germination.
Se estudi el dao al endospermo en Monteverde, Costa Rica, para determinar su frecuencia, grado y efecto en la germinacin de las semillas y el crecimiento temprano de las plntulas. Se encontr que el dao al endospermo ocurri en 72% de las semillas de Chione sylvicola (Rubiaceae) y 94% de las semillas de Quercus costaricensis (Fagaceae). El grado de dao al endospermo en Q. costaricensis no estuvo correlacionado ni con el peso ya sea de las races ni el de los brotes de semillas germinados. Se llev a cabo un experimento de eliminacin del endospermo en las semillas de Persea americana y Cinnamomum paratriplinerve (Familia Lauraceae). Las tres semanas de crecimiento no fueron suficientes para la germinacin significativa de C. paratriplinerve, pero las semillas con endospermos intactos tuvieron una probabilidad menor de sufrir una infeccin de hongos. La eliminacin del endospermo no inhibi la germinacin de las semillas de P. americana, aun cuando el 80% del peso original de la semilla haba sido removido. La mayora de las semillas solamente haban empezado a producir races y brotes despus de 23 das en el suelo y muchas de ellas todava estaban encerradas en el endospermo. Para entonces, las semillas de P. americana con un porcentaje ms alto de endospermo eliminado tuvieron races y brotes significativamente ms pesados. Esto sugiere un papel ms reducido o inhibitorio del endospermo en la germinacin inicial de la semilla.
Text in English.
Monteverde Biological Station (Costa Rica)
Costa Rica--Puntarenas--Monteverde Zone--Monteverde
Crecimiento de semillas
Estacin Biolgica de Monteverde (Costa Rica)
Costa Rica--Puntarenas--Zona de Monteverde--Monteverde
Tropical Ecology Fall 2005
Ecologia Tropical Otono 2005
t Monteverde Institute : Tropical Ecology
Endosperm loss, seed germination, and early seedling growth in large seeded tropical trees Aud rey L. Quinn Department of Biology, University of Washington ABSTRACT Endosperm damage was studied in Monteverde, Costa Rica to determine its frequency, extent, and effect on seed germination and early seedling growth. Endosperm damage was found to occu r in 72% of seeds from Chione sylvicola Rubiaceae and 94% of seeds from Quercus costarricenses Fagaceae. The extent of endosperm damage in Q. costarricenses was not related to the root and shoot weight of germinated seeds. Experimental endosperm removal was carried out in Persea americana and Cinnamomum paratriplinerve both Lauraceae seeds. The three week time period of the study was inadequate to yield significant C. paratriplinerve germination, but seeds with intact endosperm were much less likely to suffer fungal infection. Endosperm removal did not inhibit germination in P. americana seeds, even when 80% of the original seed weight had been removed. After 23 days in soil, most seeds were just beginning to produce roots and shoots, the majority of whi ch were still enclosed in endosperm. At this time, P. americana seeds with a higher percent of endosperm removed had significantly heavier roots and shoots, suggesting a reduced or even inhibitory role of endosperm in initial seed germination. RESUMEN Se estudiÃ³ el daÃ±o al endospermo en Monteverde, Costa Rica, para determinar su frecuencia , grado y efecto en la germinaciÃ³n de las semillas y el crecimiento temprano de la s plÃ¡ntulas. Se encontrÃ³ que el daÃ±o al endospermo ocurriÃ³ en 72% de las semillas de Chi one sylvicola Rubiaceae y 94% de las semillas de Quercus costarricenses Fagaceae. El grado de daÃ±o al endospermo en Q. costarricenses no estuvo correlacionado ni con el peso ya sea de las raÃces ni el de los brotes de semillas germinados. Se llevÃ³ a cabo un experimento de eliminaciÃ³n del endospermo en las semillas de Persea americana y Cinnamomum paratriplinerve Familia Lauraceae. Las tres semanas de crecimiento no fueron suficientes para la germinaciÃ³n significativa de C. paratriplinerve , pero las sem illas con endospermos intactos tuvieron una probabilidad menor de sufrir una infecciÃ³n de hongos. La eliminaciÃ³n del endospermo no inhibiÃ³ la germinaciÃ³n de las semillas de P. americana , aun cuando el 80% del peso original de la semilla habÃa sido removido . La mayorÃa de las semillas solamente habÃan empezado a producir raÃces y brotes despuÃ©s de 23 dÃas en el suelo y muchas de ellas todavÃa estaban encerradas en el endospermo. Para entonces, las semillas de P. americana con un porcentaje mÃ¡s alto de endos permo eliminado tuvieron raÃces y brotes significativamente mÃ¡s pesados. Esto sugiere un papel mÃ¡s reducido o inhibitorio del endospermo en la germinaciÃ³n inicial de la semilla. INTRODUCTION Seed endosperm provides nutrition to sust ain developing embryos during seed germination and initial growth Raven and Johnson 1986. Overall, tropical seeds are consistently larger in size, and therefore implicitly higher in nutrient endosperm content than temperate species. Despite being relat ively large on average, seeds in the tropics cover a size range of ten orders of magnitude, from tiny orchid dust seeds to giant coconuts. A significant relationship exists between growth form and seed size, with canopy trees and lianas having the largest seeds Leishman et al. 2000. Within a growth form, large seed size may reflect a greater competitive ability. Multiple studies have shown a positive relationship between initial seedling size and seed size, even within species Dolan, 1984; Moegenburg, 19 96; Wulff, 1986; Zhang and Maun,
1991. When a steep light or moisture gradient exists within a few centimeters of the soil surface, even a few extra millimeters of shoot or root length can make a great difference competitively for a seedling Leishman et al. 2000. As larger seeds produce larger seedlings, larger seed seedlings are able to grow up out of greater depths of soil, where the seed can receive more moisture. Seedlings of larger seeds also emerge from leaf litter more successfully. Larger seeds t end to have a higher percentage of their mass stored as reserve that is gradually deployed to permanent structures. More reserve remains in the seed uncommitted, available for the seed to draw upon in case of undesirable growing conditions or seedling dist urbance. Slower germination and growth can put larger seeds at a competitive disadvantage, since less tissue is available for photosynthesis. However, studies have concluded that larger seeds have an advantage in times of drought, intense shade, and minera l nutrient deficiency Leishman et al. 2000. Dispersal modes are also associated with seed size Harper et al. 1970, Primack 1987. Larger seeded species are typical of the mammal dispersal syndrome, Foster and Janson 1985 but little evidence exists con clusively linking seed size with distance of dispersal Leishman et al. 2000. Loss of endosperm to predation or parasitism is widespread in tropical forests due to the length spent on the forest floor Calvo IrabiÃ©n and Islas Luna 1999; Dalling, Harms, a nd AizprÃºa 1997. Tropical canopy trees, seeds and seedlings commonly suffer severe damage from insects and fungi Clark and Clark 1991. The role of seed herbivores, or granivores, in selecting for seed size in the tropics has been reported in multiple st udies. Harms and Dalling 1997 examined 13 dicotyledon tree species from Barro Colorado Island BCI, with seeds ranging in weight from 0.2 to 107.6 grams. After each seed produced its first pair of fully expanded leaves, they clipped the above ground sho ot and found that only seeds heavier than five grams were capable of resprouting and producing fully functional seedlings. These results agree with a similar study for seed damage to Macuna andreana Fabaceae, a tropical liana with seeds weighing five to ten grams Janzen 1976. Further, Ichie et al. 2001 show that germination and seedling growth are accompanied by depleted starch and lipid reserves in Dryobalanops lanceolata Dipterocarpaceae. Once depleted, seedling growth ceases, suggesting that seed ling size is a direct result of seed reserves. Dalling, Harms, and AizprÃºa 1997 studied the germination, initial seedling production, and seed and seedling damage tolerance of Prioria copaifera Fabaceae, the largest seeded tree species on Barro Colorad o Island, PanamÃ¡. They found that experimental endosperm removal of up to 60% of seed mass did not affect the proportion of seeds germinated, but decreased shoot mass. This supports the portrayal of seed endosperm as a determining factor in seedling growth rather than germination. Janzen 1976 demonstrated that removal of only 1%, 5%, and 10% of M. andreana seed weight by drilling holes in the endosperm to mimic insect damage severely reduced seed fitness in the face of artificial granivory. Less seed rese rves remained to be utilized to compensate for the herbivore damage. In the present study I first aimed to confirm the commonness of endosperm damage in the Cloud Forest by examining the frequency of granivory in Chione sylvicola Rubiaceae, a small seed ed member of the coffee family whose seeds are commonly seen along Cloud Forest trails. I next examined the impact of natural and simulated granivory on germination and initial seedling growth for two other species of Cloud Forest trees, Quercus costarricenses Fabaceae and Cinnamomum paratriplinerve Lauraceae, and along with the commercial avocado, Persea americana Lauraceae. Specifically, I analyze the role of endosperm in seed germination ability and early seedling size. I note the degree of fungal i nfection in seeds with removed endosperm,
and quantify the amount of endosperm missing and development of root and shoot from the two Cloud Forest species. Further, I test for whether root and shoot masses are enhanced by soil nutrients or are accumulated from endosperm stores alone by growing half of the avocado seeds in water rather than soil. MATERIALS AND METHODS Field Observations In mid November 2005 C. sylvicola seeds along a forest trail near the Monteverde Biological Station were examined. The first 100 red C. sylvicola. seeds lying exposed on the trail were inspected for granivore damage. These seeds, which were typically around 20 grams in weight, were categorized as damaged if they exhibited exit holes or had portions of the seed removed. Whi le counting the 100 seeds a tally mark was recorded for each counted seed qualifying as damaged in order to determine the percentage of seeds preyed upon. During the same week 100 Q. costarricenses Fagaceae seeds mean weight = 2.26 g, sd = 0.68 g we re collected from the same trail. Q. costarricenses drops its seeds, or acorns, synchronously so all the seeds were of relatively the same age Burger 1975. Almost all seeds were in the early stages of germinating, with a small leafless shoot only a few ce ntimeters in diameter emerging from the center of the seed. To accurately measure seed size only seeds with both halves remaining were gathered. After the seeds were gathered any pieces of shell or debris attached to the seeds were removed and the root and shoots were clipped off. The number of insect exit holes was counted for each seed, and the percent of endosperm removed from the seed was estimated. Each pair of seed halves was weighed, along with their respective roots and shoots. Original seed weights were calculated using the measured seed weights and respective estimates of percent endosperm removed. Endosperm removal experiments 100 P. americana seeds were removed from their fruits. In order for this experiment to yield data, the seeds at least t he ones without endosperm removed needed to be capable of germinating within the short time span allotted for this study. Out of this concern the seed coat was removed from each seed, as well as a sliver of tissue from both ends, as this has been shown to hasten germination Eggers 1942. Each seed was weighed after this removal. To determine whether soil nutrients as well as endosperm impacted seedling growth, 50 of the P. americana seeds were grown in water and the remaining 50 were designated for soil t reatment. Seeds were equally divided between manipulations of 0%, 20%, 40%, 60% and 80% endosperm removal by weight Table 1. A knife was used to cut away tissue from the seed end distal to the embryo, leaving the actual embryo untouched. The water treatm ent seeds were each supported within the rim of a water filled plastic cup by three toothpicks, assuring that at least the embryo contain bottom two centimeters of the seed was submerged. The soil treatment seeds were planted in large germination trays wit h .5 1 cm of seed left exposed above the soil, which was a two to one mixture of forest soil and sand. Both treatments were placed indoors along a large westward facing window. The soil in the germination trays was kept moist and the water cups were kept a t a constant level of water.
TABLE 1. Mean weights of P. americana seeds in a the water treatment set and b the soil treatment set. a b % of endosperm remov ed Mean initial weight g Mean after cut weight g % of endosperm removed Mean initial weight g Mean after cut weight g 0 30.895 Â± 5.616 30.895 Â± 5.616 0 25.012 Â± 6.171 25.012 Â± 6.171 20 28.650 Â± 6.114 23.032 Â± 6.270 20 27.538 Â± 5.296 21.930 Â± 4.056 40 30.075 Â± 4.132 18.014 Â± 2.433 40 29.902 Â± 6.257 17.922 Â± 3.776 60 29.826 Â± 6.205 12.094 Â± 2.583 60 29.331 Â± 9.258 11.733 Â± 3.642 80 26.009 Â± 5.340 5.259 Â± 1.053 80 26.449 Â± 6.475 5.423 Â± 1.351 Fifty intact C. paratriplinerve seeds with their seed coats intact were weighed and endosperm was removed from them in the same manner as with the P. americana seeds. Ten seeds were left completely intact, and 20%, 40%, 60%, and 80% endosperm removal were each performed on ten seeds Table 2. All the C. paratriplinerve seeds were then planted in germination trays .5 cm deep in two:one forest soil and sand mixture, and placed in the same window as the P. americana seeds. TABLE 2 . Mean weights of C. paratriplinerve seeds planted in soil. % of endosperm removed Mean initial seed weight g Mean after cut seed weight g 0 0.493 Â± 0.077 0.494 Â± 0.077 20 0.493 Â± 0.060 0.395 Â± 0.049 40 0.503 Â± 0.057 0.300 Â± 0.032 60 0.478 Â± 0 .077 0.190 Â± 0.029 80 0.442 Â± 0.067 0.087 Â± 0.012 All P. americana seeds were removed from their treatments 23 days following planting. The seeds were cut open and the sprouting roots and shoots, none of which had grown longer than a few centimeters, we re then removed and weighed separately. Bivariate regression analyses were performed to test for correlations between the percent of endosperm removed and the resulting root and shoot weight. The C. paratriplinerve seeds were removed after 22 days in soil and all seeds were examined for shoots or roots. If the seed had germinated the shoot and root were weighed together. Each seed was also inspected for fungal infection. RESULTS Field Observations Endosperm damage is common. Out of the 100 C. sylvicola seeds observed, 72 seeds exhibited herbivore damage. This damage was characterized by exit holes or partially missing endosperm. In damaged seeds, the soft tissue on the outside of the seed either had chunks taken out of it, showed small dark round bruise s around tiny insect exit holes, was split laterally, or showed any combination of these signs of granivory. Seed damage ranged in severity from only one or two insect exit holes to the removal of over half of the endosperm. Quercus costarricenses seeds suf fer an even higher frequency of damage than C. sylvicola . Furthermore, damage in this species is cause by insects both boring holes into the seed endosperm and consuming its contents. Some of the seeds still contained insect larvae within
them when examine d in the lab. Only six of the 100 Q. costarricenses seeds gathered show no signs of endosperm damage. Seeds had an average of 11 Â± 13% of their endosperm estimated to be missing Figure 1a. Ten of the seeds had so little intact tissue remaining that it was impossible to count the number of insect exit holes in the seed. These seeds were excluded from exit hole analyses. With these seeds excluded the number of exit holes in a seed averaged 28 Â± 17 Figure 1b. FIGURE 1. Q. cost arricenses seeds with a estimated percent of seed endosperm removed n = 100 and b number of exit holes n = 90. Insects do not prey preferentially upon larger Q. costarricenses seeds. The number of exit holes in a Q. costarricenses seed did not relate to either its measured weight r 2 = 0.006, P = 0.486, n = 90; Figure 2a nor its calculated original weight r 2 = 0.026, P = 0.130, n = 90; Figure 2b. Since insect damage did not vary significantly by seed size, seeds of smaller weight on average suffered a greater proportion of endosperm loss. FIGURE 2. Significant regressions relating the percentage of endosperm removed from Q. costarricenses seeds to a the measure d seed weight r 2 = 0.21, P < 0.0001, n = 100 and to b the calculated estimated original seed weight r 2 = 0.086, P = .0032, n = 100. Insects damaging Q. costarricenses endosperm do not regularly damage the seed embryo or seedling. The shoot and root g rowth of the germinated seeds was not found to be significantly 0 5 10 15 20 25 30 35 0 5 15 25 35 45 55 65 75 Number of seeds Estimated percent of seed endosperm removed a 0 2 4 6 8 10 12 14 16 0 6 to 10 16 to 20 26 to 30 36 to 40 46 to 50 56 to 60 66 to 70 Number of seeds Number of exit holes b 0 10 20 30 40 50 60 70 80 0.0 2.0 4.0 Percent of seed endosperm removed Measured seed weight g a 0 10 20 30 40 50 60 70 80 0.0 2.0 4.0 Percent of seed endosperm removed Estimated original seed weight g b
hindered by any amount of endosperm damage. The number of exit holes in a seed neither correlated to its root and shoot weight r 2 = 0.0004, P = 0.845, n = 90; Figure 3a, nor to its root and shoot weight as a percentage of its calculated original weight r 2 = 0.0006, P = 0.813, n = 90; Figure 3b. FIGURE 3. There was no significant correlation between either the number of ins ect holes removed from a Q. costarricenses seed and a the weight of its respective root and shoot nor b the weight of the root and shoot relative to the original seed weight. Endosperm removal experiments The large avocado seeds did not have inhibited germination even when 80% of their endosperm had been removed. Twenty three days after planting, 97% of the seeds had germinated. This germination was so subtle, however, that only about a quarter of the shoots had grown taller the top of the seed mean r oot and shoot weight of water treatment seeds = 0.170 g, sd = 0.092 g; mean root and shoot weight of soil treatment seeds = 0.142 g, sd = 0.111 g. Most shoots and roots combined were only a centimeter longer than the embryo, and none had begun to visibly deplete the endosperm reserves. Upon the extraction of seedlings from soil treatment seeds it was discovered that the 80% of the endosperm cut from one seed had been inadvertently removed from the wrong end, so that the seed no longer contained an embryo. This seed was excluded from analysis. By examining the contents of each seed when removing seedlings it was confirmed that this error did not occur in any of the other 99 P. americana seeds used in the study. Seeds grown in soil differed from those grown i n water in that removal of their endosperm increased their root and shoot weights. The water treatment seeds showed no significant correlation between either the percent of endosperm removed and the respective root, shoot, root and shoot combined weight, o r root and shoot combined weight as a percentage of the original weight of each seed. The P. Americana seeds grown in soil, however, showed a strong positive linear relation was found for each of these relationships except the percent of endosperm removed versus root weight Fig. 4. 0 20 40 60 80 0.0 0.2 0.4 0.6 0.8 1.0 Number of insect exit holes Shoot and root weight g a 0 20 40 60 80 0% 10% 20% 30% 40% Number of insect exit holes Shoot and root weight as % of original seed weight b
FIGURE 4. The effect of the percent of endosperm removed on a shoot weight r 2 = 0.131, P = 0.011, n = 49, b combined root and shoot weight r 2 = 0.104, P = 0.024, n = 49, and c combined root and shoot weight as a percentage of the initial seed weight r 2 = 0.084, P = 0.044, n = 49 of P. americana seeds grown in soil. Only one C. paratriplinerve seed had germinated by the end of the growi ng experiment. This was insufficient data to draw conclusions on the effect of endosperm removal on germination growth. However, a large number of the seeds had developed a fungal infection. The fungus infested endosperm swelled up above the edges of the s eed coat where the seed had been cut, and the endosperm was white and liquefied. All infected seeds were dissected to inspect for germination, and none of their embryos showed any sign of growth; some were so degraded they were barely distinguishable from the rest of the dissolving tissue. Intact seeds showed no infection, but all other categories, even those with as little as 20% removed, had complete or nearly complete infestation Figure 5. A Chi squared sample test showed that that the number of seeds infected with the fungus varied significantly between the percent endosperm removed groups X 2 = 9.85, df = 4, p < 0.05;. FIGURE 5. Cinnamomum paratriplinerve seeds with removed endosperm were more susceptible to fungal infecti on, while those still intact were untouched by the fungus. DISCUSSION Large seeds may be adaptive for many reasons. They are better equipped to res pond to abiotic conditions important for tropical trees: shade, seasonal rain, depleted soils Leishman et al. 2000, as well as the biotic conditions that are stronger in Tropical forests: competition, 0.00 0.05 0.10 0.15 0.20 0.25 0% 20% 40% 60% 80% shoot weight g Percent of seed endosperm removed a 0.0 0.1 0.2 0.3 0.4 0.5 0% 20% 40% 60% 80% Root and shoot weight g Percent of seed endosperm removed b 0.0% 0.5% 1.0% 1.5% 2.0% 0% 20% 40% 60% 80% Root and shoot weight as a % of initial seed weight Percent of seed endosperm removed c 0 2 4 6 8 10 12 0 20 40 60 80 Number of seeds infected Percent of seed endosperm removed
predation, and parasitism. Further, these conditions are not necessarily mutually exclusive. The current study explores the importance of granivory to germination and subsequent immediate seedling growth. Seed endosperm damage is common in lowland tropical forests Calvo IrabiÃ©n and Islas Luna 1999, Clark a nd Clark 1991, Dalling et al. 1997. Cloud Forest seeds seem to suffer the same high rates of endosperm loss as seeds from lower sites. Of the two seeds studied in the field, both showed high rates of endosperm damage. Predation of C. sylvicola seeds was n ot directly witnessed so its specific granivore is unknown, but insect larvae still remaining in the Q. costarricenses allowed the identification of their damager. Larger oak seeds suffered less damage proportionately than smaller oak seeds, as insects seem ed to attack both large and small seeds with equal frequency. Larger seeds may also escape fungal infections longer, though here fungal infection resulted just from any endosperm removal. Seed size had no effect on the extent of its insect infestation nor did insect damage reduce root and shoot size. Insect infestation in acorns occurs most frequently on the basal end of the seed, opposite from the embryo, possibly explaining why germination frequencies are not affected by endosperm removal Steele et al. 1 993. Germination experiments showed that early seedling development is not impacted by endosperm removal, even up to 80% removal. This initially seems to conflict with the P. copaifera study by Dalling, Harms, and AizprÃºa 1997. However, that study was conducted over a 38 week growing time, after which all the germinated seedlings had a fully developed pair of leaves. Seedlings from avocadoes were smaller than those observed in the oak seeds, and much smaller than those from the P. copaifera study, so it is likely due to their small size that only 20% of endosperm remaining was sufficient for their early growth. This successful germination after endosperm damage coincides with the initial results of a study on Acacia albida Mimosaceae seeds by Hauser 1 994. Seeds showing bruchid beetle exit holes initially had a high germination rates. A few months later, though, they all had died. Other studies Dalling et al. 1997, Ichie et al. 2001 suggest that later seedling development would expose the negative im pact of endosperm removal on seedling size, however. It is predicted that though the experimentally endosperm deficient P. Americana seeds had greater initial germination growth, had the experiment been allowed to run longer their roots and shoots would ha ve eventually been surpassed in weight by those of seeds containing more endosperm. This could put plants with more endosperm removed at a competitive disadvantage, as taller seedlings and saplings are more likely to win a gap Brokaw and Busing 2000. Finally, soil seedlings actually grew larger with more endosperm removed. It is likely that this was due to their earlier germination. Apparently the greater amount of endosperm remaining on the more intact seeds inhibited their germination. The function o f endosperm is to fuel growth of a seedling until it is capable of gathering energy from photosynthesis Ichie et al. 2001; the mechanism of germination may be more significantly fueled by an interaction between the seed embryo and the soil rather than th e use of energy reserves in the endosperm. Having less endosperm between the embryo and the soil would have expedited the embryo s contact with soil compounds and thus may have hastened germination in soil treatment seeds with more endosperm removed. Perha ps avocado seeds grown in water did not show a similar t rend in seedling size due to their lack of soil substrate whose proximity could have given seeds with less endosperm around their embryo an advantage. Earlier germination among endosperm deprived soil treatment seeds may also explain why gap plants have smaller seeds, if fast germination time is more important for them than ultimate seedling size. Large seeds may
germinate slower, but their seedlings are larger which is more important for growth f orms such as canopy trees and lianas that have a lot of time on their hands at the seedling level, but once their endosperm is exhausted harvest less energy to draw upon for growth because they receive little light. ACKNOWLEDGMENTS Thank you to Alan Mas ters for his extensive guidance during this study. Thank you to Lorenzo Vargas for supplying the Cinnamomum paratriplinerve seeds, and to Bill Haber and Willow Zuchowski for their help in seed identification. Thank you to the Arguedas family for their kind ness while having me in their home. Lastly, thank you to all the other CIEE Fall 2005 students for their wonderful comradery, I do not know how I would have been able to complete this study without them. LITERATURE CITED Brokaw, N. and Busing, R. T. 2000 . Niche versus chance and tree diversity in forest gaps. Tree. 15:183 188. Burger, W. C. Quercus costaricensis Encino, Roble, Oak. In: Janzen, D. H., ed. Costa Rican Natural History. Chicago: The University of Chicago Press. 7:318 319. Calvo Irab iÃ©n, L. M., and Islas Luna, 1999. A. Predispersal predation of an understory rainforest herb Aphelandra aurantiaca Acathaceae in gaps and mature forest. American Journal of Botany. 86:1108 1113. Clark and Clark. 1991. The impact of physical damage on ca nopy tree regeneration in tropical rain forest. Journal of Ecology. 79:447 457. In: Dalling, J. W., Harms, K. E., and AizprÃºa A. 1997. Seed damage tolerance and seedling resprouting ability of Prioria copaifera in PanamÃ¡. Journal of Tropical Ecology. 13:48 1 490. Dalling, J. W., Harms, K. E., and AizprÃºa A. 1997. Seed damage tolerance and seedling resprouting ability of Prioria copaifera in PanamÃ¡. Journal of Tropical Ecology. 13:481 490. Dolan, R.w 1984. The effect of seed size and maternal source on ind ividual size in a population of Ludwigia leptocarpa Onagraceae. American Journal of Botany. 71:1302 1307. In: Leishman, M. R., Wright, I. J., Moles, A. T., and Westoby, M. The evolutionary ecology of seed size. In: Seeds: The Ecology of Regeneration in P lant Communities, 2 nd edition. CAB International. 2:31 57. Eggers, E.R. 1942. Effect of the removal of the seed coats on avocado seed germination. California Avocado Society 1942 Yearbook. 27:41 43. Foster, S. A. and Janson, C. H. 1985. The relationship between seed size and establishment conditions in tropical woody plants. Ecology. 66:773 780. In: Leishman, M. R., Wright, I. J., Moles, A. T., and Westoby, M. The evolutionary ecology of seed size. In: Seeds: The Ecology of Regeneration in Plant Communiti es, 2 nd edition. CAB International. 2:31 57. Harms, K. E and Dalling, J. W. 1997. Damage and herbivory tolerance through resprouting as an advantage of large seed size in tropical trees and lianas. Journal of Tropical Ecology. 13:617 621. Harper, J. L., Lovell, P. H., and Moore, K. G. 1970 The shapes and sizes of seeds. Annual Review of Ecology and Systematics. 1:327 356. In: Leishman, M. R., Wright, I. J., Moles, A. T., and Westoby, M. The evolutionary ecology of seed size. In: Seeds: The Ecology of Rege neration in Plant Communities, 2 nd edition. CAB International. 2:31 57.
Hauser, T. P. 1994. Germination, predation and dispersal of Acacia albida seeds. Oikos. 70:421 426. Ichie, T., Ninomiya, I., and Ogino, K. 2001. Utilization of seed reserves during g ermination and early seedling growth by Dryobalanops lanceolata Dipterocarpaceae. Journal of Tropical Ecology. 17:371 378. Janzen, D. H. 1976. Reduction of Mucuna andreana Leguminosae seedling fitness by artificial seed damage. Ecology. 57: 826 828. Leishman, M. R., Wright, I. J., Moles, A. T., and Westoby, M. The evolutionary ecology of seed size. In: Seeds: The Ecology of Regeneration in Plant Communities, 2 nd edition. CAB International. 2:31 57. Moegenburg, S. M. 1996. Sabal palmetto seed size causes of variation, choices of predators, and consequences for seedlings. Oecologia. 109:539 543. In: Leishman, M. R., Wright, I. J., Moles, A. T., and Westoby, M. The evolutionary ecology of seed size. In: Seeds: The Ecology of Regeneration in Plant Com munities, 2 nd edition. CAB International. 2:31 57. Primack, R. B. 1987. Relationships among flowers, fruit and seeds. Annual Review of Ecology and Systematics. 18:409 430. Raven, P. H. and Johnson, G. B. 1986. Diversity of Plants. In: Biology. St. Louis: Times Mirror/Mosley College Publishing. 29:583 606. Steele, M. A., Knowles, T, Bridle, K, and Simms, E. L. 1993. Tannins and partial consumption of acorns implications for dispersal of oaks by seed predators. American Midland Naturalist. 130:229 238. Wenny, D. G. 2000. Seed dispersal, seed predation, and seedling recruitment of a Neotropical montane tree. Ecological Monographs. 70:331 351. Wulff, R. D. 1986. Seed size variation in Desmodium paniculatum II. Effects on seedling growth and physiological performance. Journal of Ecology. 74:99 114. In: Leishman, M. R., Wright, I. J., Moles, A. T., and Westoby, M. The evolutionary ecology of seed size. In: Seeds: The Ecology of Regeneration in Plant Communities, 2 nd edition. CAB International. 2:31 57. Zhan g, J. and Maun, M. A. 1991. Establishment and growth of Panicum virgatum L. seedlings on a Lake Erie sand dune. Bulletin of the Torrey Botanical Club. 118:141 153. In: Leishman, M. R., Wright, I. J., Moles, A. T., and Westoby, M. The evolutionary ecology o f seed size. In: Seeds: The Ecology of Regeneration in Plant Communities, 2 nd edition. CAB International. 2:31 57.