View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Aquatic Commons 22 NOTICIAS DE GALAPAGOS No.54

POLLINATOR AVAILABILITY: A POSSIBLE EXPLANATION OF IN INTER.ISLAND FLORAL VARIATION "IUSZCIA GAI-APAGAN A (A CANTHA CEAE)

By: Conley K. McMullen

INTRODUCTION ducted, resulting in 53 angiospenns being classified as self-compatible, \^/ith 50 of these capable of auto- Floristic studies in Galápagos have had an admi- gamy (Aide, 1986; Elisens, 1989; McMullen, 1990; rable history, and have culminated in works such as Rick, 1966). Only one has been classified as 'Flora of the Galápagos Islands' (V/iggins and Por- self-incompatible (Grant and Grant, 1981). ter, l97I) and 'An updated and annotated check list Based on the results of these breeding studies and of the vascular plants of the Galápagos Islands' the fact that there are relatively few potential pollina- (Lawesson et al., l98l ). However, an understanding tors in the Galápagos, it has been hypothesized that of the relationships that exist between the members the first angiosperms to colonize the islands were of this unique flora and their coinhabitants is those that possessed upon arrival, or developed soon still in its infancy. after, the ability to reproduce autogamously (Aide, Stewart (1911), after visiting the Galápagos Is- 1986; McMullen, 1981,1990; Rick, 1966). The ini- lands as part of the California Academy of Sciences tial scarcity of pollinating , including X. expedition in 1905-6, made one of the earliest refer- darwini, could also explain the small flower size and ences to this subject when he noted that most of the drab color of the maj ority of endemics. Without faith- endemic angiosperms had small flowers and suggest- ful pollinators there would be little or no apparent ed that this was due to the scarcity of insects in the selective advantage for large attractive corollas. The archipelago. The inference was that large showy co- small amount of pollen produced by most flowers rollas had little selective value since there were so further supports this scenario (Colinvaux and few insects present to attract as pollinators. Schofield, 1976;McMullen and Close, 1993). Studies on the pollination biology of various The objective of this study was to investigate the Galápagos angiosperms have followed (Aide, 1986; relationship between pollinator availability and plant Eiisens, 1989; Grant and Grant, 1981; Linsley et al., reproduction by comparing these variables on Pinta 1966; McMullen, 1987; Rick, 1963, 1966). Results Island and Santa Cruz Island. Xylocopa darwini is suggest that the endemic , Xylocopa common on Santa Cruz,buthas never been found on datwini Cockerell (:), is a ma- Pinta. These studies would help determine whether jor pollinator on the islands it inhabits. In total, 73 or not the carpenter bee, since its arrival, has influ- angiosperm taxa(70 species) have been recorded as enced the development of floral characters and visited by this bee. Non-endemic plants appear to be pollination strategies of selected plants on the islands favored as only 24 of tbese taxa are restricted to the it inhabits. Galápagos Islands. Linsley et al. (7966) suggested Hypothetically, the presence of X. darwini on Santa thatX. darwini was probably more important for the Cruz will have promoted more attractive fl oral displays establishment of the native and introduced elements than those found in the resident flora of Pinta. This of the Galápagos flora than for the older endemic attractiveness mightbe expressed in characters such as members. Reports of other insect visitors are limited inflorescence and flower sizes. Moreover, a character- (Grant and Grant, 1981; Linsley,7966; McMullen, istic such as larger flowers might mean that the plants 1986). Plant breeding studies have also been con- inhabiting Santa Cruz rely less on autogamy. November 1994 NOTICIAS DE GALAPAGOS 23

METHODS Other studies were performed while awaiting the results of the breeding experiments. First, a variety of Fieldwork for the first phase of this research was inflorescence and floral characters were measured performed on Pinta Island from 23 June to 26 Júy using vernier calipers. Second, 26 hours ofobserva-

1 990. Four angiosperm species thatinhabitboth Pinta tions were conducted over four days to determine and Santa Cruz had sufficient flowers for testing at what insects were flower visitors and might act as this time (McMullen,1993). However, only Justicia pollinators. Information such as how many visits g al ap a g an a Lindat (Acanthac e ae) is reported here were made and how long each visit lasted was ob- since it alone was known from previous studies to be tained. An LED stopwatch was used to make these regularly visited and pollinated by X. darwini onSanta measurements. Cruz (Linsley et aI.,1966; McMullen, 1985; Rick, Similar studies were undertaken on Santa Cruz re66). Island from 31 July to 10 August 1990. Two sites This plant is an endemic herb that may reach 1 m near the craters known as Los Gemelos (ca. 630 m in height. It produces axillary inflorescences that altitude) were used. No breeding experiments were typically have only a few flowers. The corollas are conducted as this information was available frompre- purple to lavender or white in color, and the former vious studies performed in 1983-84 (McMullen, often have white markings in the throat. No notice- 1 987). All other measutements and observations were able fragrance is present. The fruit is a capsule that performed as on Pinta. However, only five hours may produce up to four seeds. over two days were spent observing insect visitors on The study site was located at 580 m altitude on the this island. Once again, this was due to the fact that southeastern slope ofPinta. Breeding experiments were research had previously been conducted on Santa conducted to determine if the species was capable of Cruz. autogamy. Fruit and seed yields were compared be- tween inflorescences completely isolated from insects RESULTS (bagged) and others that were exposed to potential pol- linators before being covered (open-pollinated). The Breeding studies indicate thatJ. galapaganais at actual number of individuals studied was impossible to least facultatively autogamous (Table 1). This spe- determine without destroying the plants due to their cies showed ahigherpercentage of autogamous fruit close spacing and vegetative growth by runners. and seed set on Pinta than on Santa Crtz. Open-

Table 1. Breeding experiment results. Those for Santa Cruz are based on studies conducted in 1983-84 (McMullen, 1987).

# Flowers Vo Fruit 7o Seed Tested Set Set

Bagged Pinta 80 43.8 3t.3 Santa Cruz 68 32.4 18.0

Open-Pollinated Pinta 92 29.3 19.6 Santa Cruz t25 74.4 10.8 24 NOTICIAS DE GALAPAGOS No.54

pollination on both islands produced a lower fruit and for atotal of 95 seconds. Three insect species visited seed set than bagged flowers. Once again, however, this plant on Santa Cruz during the timed observa- ahigherpercentage was found onPintathan on Santa tions. These w ere Tbxome rus c rockeriCurran (Diptera: Cruz. S yrphidae), U r b anu s do r ant e s g alapagen sis Williams Table 2 shows the results of the measurements of (: Hesperiidae), and a short-horned grass- the inflorescences and flowers. Significant differences hopper nymph (Orthoptera: Acrididae). Tbxomerus in all characters except number of open flowers per crockeri was most frequently observed (44 visits, inflorescence and corolla lip width are found between 3,810 seconds). One untimed visit was made by Pinta and Santa Craz. The mean values of all char- Phoebis sennae L. (Lepidoptera: Pieridae). acters except corolla tube width are higher for Santa Xylocopadarwiniwas not observed atthe flowers Cruz. of J. g al ap a g ana durìng these studies. This was prob - A single insectwas observedvisiting.I. galapaga- ably because of the weather. The highlands were na onPinta (Table 3). This was a damsel bug nymph often wet due to rain and the seasonal mist known as (Hemiptera: Nabidae), which visited only one flower garua. It must be remembered, however, that previ-

--.= Table 2. Inflorescence and flower measurements (mm). Significant differences: P<.001, "= P<.05. (Independent samples t- test).

Mean SD N

Infl orescence Length*** Pinta 33.09 16.16 50 Santa Cruz 44.15 18.70 91

Flowers Open / Inflorescence Pinta 1.00 0.00 50 Santa Cruz 1.03 0.17 9l

Entire Corolla Length*** Pinta 9.52 0.53 50 Santa Cruz 10.07 0.68 100

Corolla Tube Length. Pinta 5.24 0.54 50 Santa Cruz 5.45 0.51 100

Corolla Tube'Width..* Pinta 2.36 0.21 50 Santa Cruz 2.r1 0.2r 100

Corolla Lip Width Pinta 7.87 r.24 50 Santa Cruz 8.27 r.6l 100 November 1994 NOTICIAS DE GALAPAGOS 25

Table 3. Insect visitors. Visitation times are in seconds. Total refers to the times of all visits combined. N refers to the number of visits.

Total Mean SD N

Pinta Damsel Bug nymph (Hemiptera: Nabidae) 95

Santa Cruz Toxomerus crockeri (Diptera: Syrphidae) 3,810 86.59 122.50 44 U rbanus do rante s g alap a g ensi s (Lepidoptera: Hesperiidae) 300 Short-Horned Grasshopper nymph (Orthoptera: Acrididae) 6 2.00 0.00 Phoebis sennae (Lepidoptera: Pieridae) Not Timed

ous studies have shown that this plant's flowers are not appear to come into contact with the anthers dur- frequently visited on Santa Cruzby X. darwini. In ing their brief visits. addition, Lepto te s parrhas ioide s Wallengren (Lepi- Toxomerus crockeri was the most common visitor doptera: ) and Wasmannia auropunctata to J. galapagana on Santa Cruz during this study. Roger (Hymenoptera: Formicidae) are visitors to its Individuals were often seen pushing their way into flowers (Linsley etal., 7966;McMullen, 1986, 1990; the corolla throats. This movement caused the in- Rick, 1966). sect's abdomen, and more often its dorsal surface, to rub against the flower's anthers and stigmas. This DISCUSSION appears to be an ideal movement for pollination. The

and grasshopper nymphs that visited "I. Justicia galapagana exhibited a higher level of galapagana during this study appear to be unimpor- autogamy and generally smaller floral characters on tant as pollinators. Leptotes pctrrhasioides and W. Pinta than on Santa Cruz. These results may be ex- auropunctata as well are unlikely pollinators of this plained by the scarcity of insect visitors to this plant species (McMullen, 1986). Although not observed on Pinta. Although the damsel bug nymph had pollen in 1990, X. darwini remains an important pollinator adhering to its body, it was seen only once. Areturn of this plant. visit to Pinta by myself in 1993 confirmed this ab- Open-pollinated fl owers produced fewer fruits and sence of faithful pollinators. Although no timed seeds than those that were bagged. With so few pol- observations were made, it was apparent that these linators, one might not expect open-pollinated flowers flowers were of little interest to the local insects. to produce a higher yield. This is especially true Leptotes parrhasioides was commonly seen flying since wind pollination is thought to be of no conse- about the plants. Only rarely did they land on a flow- quence to this plant (McMullen and Close, 1993). er, and then just for an instant. These butterflies did These results suggest that reproduction in this spe- 26 NOTICIAS DE GALAPAGOS No.54

cies is pollen-limited on both islands. In other words, Santa Cruz. An obvious next step would be to raise more fruit could be set if more pollen vectors were individuals from both islands in the uniform environ- available. However, an explanation for fewer open- ment of a greenhouse, and then make the same flower pollinated fruits and seeds is more difficult to come and inflorescence measurements. If the differences by. Perhaps some of the flowers were damaged by are still present, then this would eliminate environ- birds or insects before being bagged. mental factors as an explanation. Unfortunately, The results of this study support the hypothesis implementing such research is often much more dif- that autogamous angiosperms were favored in the ficult than formulating the idea. The logistics of initial colonization of the Galápagos Islands. The obtaining the necessary seeds, retuming with them, presence oT X. darwini on Santa Crtrz, along with and raising the plants to reproductive maturity have other insects such as T. crockeri, may have since se- yet to be overcome. lected for flowers and inflorescences better able to vie for the attention of potential pollinators than those ACKNOWLEDGMENTS of the same plant species on Pinta. This could also explain why ,I. galapaganahad a lower level of au- Appreciation is extended to Jeraldine W. Dailey, togamous fruit set on Santa Cruz. Bryan E. Dutton, Emma K. Dutton, Bret Kuss, André Although differences were observed between these Mauchamp, Sandra J. Naranjo, JimW. Ross, and Hugo islands, two explanations might be offered for why Valdebenito for their assistance with this project. I es- they weren't more obvious. First, although there is pecially wish to thank E. Gorton Linsley, His pioneering no evidence of this, it is possible that X. darwini work on insect pollination in the Galápagos has been a inhabited Pinta in the past. Several tree species that continuing source of inspiration. T.J. Henry, M. Lacey- this bee uses for nesting on Santa Croz are also found Theisen, D.A. Nickle, and F.C. Thompson (Systematic on Pinta. Second, perhaps X. datw ini has not been in Entomology Laboratory, United States Department of the archipelago long enough to have had more of an Agriculture, Beltsville, Maryland) graciously assisted effect. This may be true since it would have needed with the insect identifications, As always, the staffs of several food and nesting plants present on the islands the Charles Darwin Foundation, Charles Darwin Re- before it could survive and reproduce (Linsley et al., search Station, and Galápagos National Park were helpful 1966). In addition, the fact that the bees themselves in solving logistical problems. Finally, I thankTAME for do not differ between islands would seem to indicate providing reduced airfare for flights between the continent a recent arrival in the archipelago. This is another and Galápagos. This work was suppofted by National area in which work is needed. Geographic grant 4327 -90. One final question shouldbe explored. What oth- er explanations might account for the inter-island LITERATURE CITED floral variation encountered? Perhaps this is simply an example intraspecific variation among island pop- Aide, M. 1986. The influence oTXylocopa darwini ulations. In other words, the smaller flowers of Pinta on floral evolution in the Galápagos. Charles might be exhibiting a founder effect. The distance Darwin Research StationAnnual Report, 1 983: 19- between Pinta and Santa Cruz would serve as a strict 21. enforcer of reproductive isolation between these is- Colinvaux, P.4., and E.K. Schofield. 1916. Histor- land populations. Rick (1983) and Elisens (1989) ical ecology of the Galápagos Islands. A have demonstrated such inter-island variation in the Holocene pollen record from El Junco, Isla San morphological and allozymíc characters of Lycoper- Cristóbal. Journal of Ecology 64:989-1012. sicon cheesmanii Riley (Solanaceae) and Galvezia Elisens, W.J. 1989. Genetic variation and evolution leuc antha Wiggins (Scrophulariaceae). of the Galápagos shrub snapdragon. National Perhaps the observed morphological variation is Geographic Research 5:98-1 10. due to differing environmental regimes on Pinta and Grant, B.R., and P.R. Grant. 1981. Exploitation of November 1994 NOTICIAS DE GALAPAGOS

McMullen, C.K. 1993. Angiosperm breeding sys- Opuntia cactus by birds on the Galápagos. tems and pollination ecology in the Galápagos

Oecologia (Berlin) 49:119-181 . Islands. National Geographic Research and Ex- Lawesson, J.E., H. Adsersen, and P. Bentley. 1987. ploration 9:380-382. An annotated checklist of the vascular plants of McMullen, C.K., andD.D. Close. 7993. Windpol- the Galápagos Islands. Reports Botanical Insti- lination in the Galápagos Islands. Noticias de tute 16: l-14. University of Aarhus, Denmark. Galápagos 52:12-17. Linsley, E.G . 1966. Pollinating insects of the Galápa- Rick, C.M. 1963. Biosystematic studies on Galápa- gos Islands . Pp.225-232 inR.I. Bowman( e d , ) gos tomatoes. Occasional Papers of the The Galápagos. University California Press, Ber- California Academy of Sciences 44:59-77 . keley. Rick, C.M. 1966. Some plant- relations on Linsley, E.G., C.M. Rick, and S.G. Stephens. 1966. the Galápagos Islands. Pp.215-224 in R.I. Bow- Observations on the floral relationships of the man (ed.) The Galápagos. University California Galápagos carpenter bee. Pan-Pacific Entomolo- Press, Berkeley. gist 42:1-18. Rick, C.M. 1 983. Genetic variation and evolution of McMullen, C.K. 1985. Observations on insect vis- Galápagos tomatoes. Pp. 97-106 in R.L Bow- itors to floweringplants oflslaSantaCruz. I. T h e man, M. Berson, andA.E. Levitan (eds.) Patterns endemic carpenterbee. Noticias de Galápago s 42:24. of evolution in Galápagos organisms. American McMullen, C.K. 1986. Observations on insect vis- Association for the Advancement of Science, San itors to flowering plants of Isla Santa Cruz. II. Francisco. Butterflies, moths, ants, hoverflies, and stiltbugs. Stewart, A. 1911. Abotanical survey of the Galápa- Noticias de Galápagos 43:21- 23. gos Islands. Proceedings of the California McMullen, C.K. 1987. Breeding systems of selected Academy of Science, Ser. 4, l:l -288. Galápagos Islands angiosperms. American Wiggins, I.L. and D.M. Porter. 1971. Flora of the Journal of Botany 14:1694-1105. Galápagos Islands. Stanford Univ. Press, Stan- McMullen, C.K. 1990. Reproductive biology of ford, California. Galápagos Islands angiosperms. Monographs Conley K. McMullen, Department of Biology and in Systematic Botany from the Missouri Botani- Chemistry, West Liberty State College, West Lib- cal Garden 32:35-45. erty, West Virginia, 26074, USA.

Justicia galapagana. Photograph by C. McMullen.