Cooperative National Park Resources Studies Unit Department of Botany University of Hawaii at Manoa Honolulu, Hawaii 96822 (808) 948-8218

COOPERATIVE NATIONAL PARK RESOURCES STUDIES UNIT DEPARTMENT OF BOTANY UNIVERSITY OF HAWAII AT MANOA HONOLULU, HAWAII 96822 (808) 948-8218

PROCEEDINGS FIRST CONFERENCE IN NATURAL SCIENCES HAWAII VOLCANOES NATIONAL PARK

NATIONAL PARK SERVICE CONTRACT #CX8000 6 0031

Clifford W. Smith, Unit Director The National Park Service and the University of Hawaii signed the memorandum of agreement establishing this Cooperative National Park Resources Studies Unit on March 16, 1973. The Unit provides a multidisciplinary approach to studies on the biological resources in the National Parks in Hawaii, that is, Hawaii Volcanoes National Park, Haleakala National Park, City of Refuge National Historical Park, and Puukohola Heiau National Historic Site. Through the Unit Director, projects are undertaken in areas identified by park management. These studies provide information of resource management programs. The involvement of University faculty and students in the resource management of the National Parks in Hawaii lends to a greater awareness of the problems and needs of the Service. At the same time research not directly or immediately applicable to management is also encouraged through the Unit. PROCEEDINGS of the FIRST CONFERENCE IN NATURAL SCIENCES in Hawaii

held at Hawaii Field Research Center Hawaii Volcanoes National Park on August 19 - 20, 1976

edited by C. W. Smith, Director, CPSUJUH Department of Botany 3190 Maile Way University of Hawaii Honolulu, Hawaii 96822 CONTENTS

PREFACE DESCRIPTIVE SUMMARY OF A NORTH KONA BURIAL CAVE, ISLAND OF HAWAII by M.S. Allen and T.L. Hunt KOA AND LEHUA TIMBER HARVESTING AND PRODUCT UTILIZATION: RELIGIO-ECOLOGICAL RELATIONSHIPS IN HAWAII, A.D. 1778 by R.A. Apple THE OCCURRENCE AND ECOLOGICAL SIGNIFICANCE OF METALLIC SKINKS ON THE ISLANDS OF HAWAI'I AND KAUA'I by K. Baker STUDIES ON THE ENDEMIC HAWAIIAN GENUS: HIBISCADELPHUS (HAU-KUAHIWI) by K. Baker and S. Allen SAVING THE 'ALALA (CORVUS TROPICUS)--SOME PRELIMINARY MANAGEMENT RECOMMENDATIONS by P. Banko ROLE OF FOOD DEPLETION BY FOREIGN ORGANISMS IN HISTORICAL DECLINE OF HAWAIIAN FOREST BIRDS by W.E. Banko and P.C. Banko ECOLOGICAL PARK MANAGEMENT: A MANAGER'S PERSPECTIVE by R.D. Barbee ECOLOGY OF RARE DROSOPHILA SPECIES IN HAWAII VOLCANOES NATIONAL PARK by H.L. Carson THE PHYSIOGRAPHY AND MARINE FAUNA OF INSHORE AND INTERTIDAL AREAS IN THE PUU-KOHOLAHEIAU NATIONAL HISTORIC SITE by D.P. Cheney, D.E. Hemmes, and R.S. Nolan SUMMARY OF RESULTS OF PRELIMINARY INVESTIGATIONS AT THE HILINA PAL1 PETROGLYPH CAVE AND ASSOCIATED SITES by P.L. Cleghorn BIRD DISTRIBUTION AND ABUNDANCE ABOVE 3000 FEET IN HAWAII VOLCANOES NATIONAL PARK by S. Conant VARIATION OF HAWAIIAN METROSIDEROS ALONG THE SOUTH AND EAST FLANKS OF MAUNA LOA, HAWAII by C. Corn MITES ASSOCIATED WITH A FUMAROLE IN HAWAII VOLCANOES NATIONAL PARK by M.L. Goff HYDRATION RIND DATES FROM ARCHAEOLOGICAL SITES IN THE SOUTH POINT AREA: A CONTRIBUTION TO HAWAIIAN PREHISTORY by T.L. Hunt THE INFLUENCE OF FERAL PIGS ON A NATIVE ALPINE GRASSLAND IN HALEAKALA NATIONAL PARK by J. Jacobi TOPOCLIMATIC GRADIENTS IN HAWAII VOLCANOES NATIONAL FARK by J.O. Juvik and G.G. Clarke

EVOLUTIONARY STUDIES OF TWO DROSOPHILA SPECIES FROM HAWAI'I: A MULTIDISCIPLINARY APPROACH by K.Y. Kaneshiro MANAGEMENT OF SCIENCE RESERVE AREAS by J.I. Kjargaard ENDANGERED PLANTS IN HAWAII VOLCANOES NATIONAL PARK by C.H. Lamoureux PHENOLOGICAL STUDIES IN HAWAII VOLCANOES NATIONAL PARK by C.H. Lamoureux CURRENT STUDIES OF KT-LAU-EA ANDMAUNA LOA VOLCANOES BY THE HAWAIIAN VOLCANO OBSERVATORY OF THE U.S. GEOLOGICAL SURVEY by P.W. Lipman, L.A. Anderson, R.Y. Koyanagi, and J.P. Lockwood PUZT-KOHOLXHEIAU NATIONAL HISTORIC SITE PLANT SURVEY by J.D. Macneil, Jr., L.K. Croft, and D.E. Hemes THE MAUNA KEA ADZ QUARRY COMPLEX, 'HAWAII: A FIRST ANALYSIS by P.C. McCoy

CATERPILLARS (EUPITHECIA SPP.) AS OBLIGATORY AMBUSH PREDATORS: A UNIQUE ADAPTIVE SHIFT IN THE HAWAIIAN ISLANDS by S.L. Montgomery THE MAJOR VEGETATION TYPES AND ECOLOGICAL ZONES IN HAWAII VOLCANOES NATIONAL PARK AND THEIR APPLICATION TO PARK MANAGEMENT AND RESEARCH by D. Mueller-Dombois LIVE COLOR PHOTOGRAPHY OF ENDEMIC HAWAIIAN INVERTEBRATES by W.P. Mull

HOST SPECIFICITY IN HAWAIIAN DROSOPHILA: NUTRITIONAL OR BEHAVIORAL? by A.T. Ohta THE EFFECTS OF FIRE UPON A HAWAIIAN MONTANE ECOSYSTEM by T.T. Parman HILINA PAL1 FIRE OF 1975 by T.T. Parman REESTABLISHMENT OF NATIVE FLORA AND FAUNA IN HAWAII VOLCANOES NATIONAL PARK by D. Reeser MMANEFOREST DECLINE ON MAUNA KEA: A REALITY OR MYTH by P.G. Scowcroft INTERPRETIVE TRAINING by J. Shimoda FIFTEEN YEARS OF VEGETATION INVASION AND RECOVERY AFTER A VOLCANIC ERUPTION IN HAWAII by G.A. Smathers THE CPSU UH by C.W. Smith HALEAKALA RBI by C.W. Smith DISTRIBUTION AND VEGETATIVE ANATOMY OF HAWAIIAN SANDALWOOD by R.L. Stemermann THE INFLUENCE OF FOOD SUPPLEMENTATION UPON THE REPRODUCTIVE STRATEGY AND MOVEMENT PATTERNS IN THE HAWAI'I 'AMAKIHI (LOXOPS VIRENS) by C. van Riper I11 PATTERNS IN THE DISTRIBUTION AND ABUNDANCE OF THE PALILA ON MAUNA KEA by C. van Riper I11 and J. M. Scott ROLE OF RESEARCH IN THE MANAGEMENT OF NATURAL RESOURCES IN AREAS ADMINISTERED BY NATIONAL PARK SERVICE IN HAWAII by O.L. Wallis THE KALA-PANA EXTENSION: ITS VARIETY, VEGETATION, AND VALUE by F.R. Warshauer LIST OF PARTICIPANTS SUBJECT INDEX PREFACE

The First Conference in Natural Sciences in Hawai'i was held at the Hawaii Field Research Center, Hawaii Volcanoes National Park, on August 19 and 20, 1976. The conference was sponsored by the Hawaii Natural History Association, Hawaii Volcanoes National Park, and the Cooperative National Parks Resources Studies Unit at the University of Hawaii. Forty papers were read on pure and applied subjects ranging from archaeology to zoology. There were 93 registered participants, and a sizeable number of others attended at various times. The objective of the conference was to bring together the various people doing research in the National Parks or for the managers of the Parks. However, the conference was not limited to National Park research. By discussing their research and problems it was hoped to increase the communication among researchers and managers. In this latter respect the conference was eminently successful. The success of the meeting prompted suggestions that such conferences be held annually. Though this idea was enthusiastically supported, later discussions resulted in a provisional agreement that future conferences should be held biennially. In the alternating years, seminars on specific subjects might be held. The problems caused by feral pigs, the subject of an informal discussion on the evening of August 19, is being considered as the subject for the first seminar. The Proceedings are published to disseminate the information presented as widely as possible. Papers are arranged in alphabetical order, by author's name. I have done as little editing as possible. Unfortunately, the number of illustrations has been severely curtailed because of the cost in preparing the plates. For consistency, Hawaiian words have been treated as follows. They have not been underlined as foreign words normally are, because to most of us some Hawaiian words are part of our normal scientific vocabulary, e.g., 'Chi'a, pahoehoe, etc. Furthermore, the spelling follows that in Pukui, Elbert, and Mookini's Place Names of Hawaii, or Pukui and Elbert's Hawaiian ~ictionr(ezept in some cases in proper names where +t ey are spelled in the customary way, e.g. Haleakala National Park, cf. Hale-a-ka-la). I realize that this is not conventional; however, I think that the reader unfamiliar with Hawaiian words will have a better chance at pronouncing the words correctly when presented in this manner. I would like to thank the following for their assistance during the conference: Suzy Allen, Cindy Furukawa, Marsha Isoshima, Larry Katahira, and Deborah Weiner. My special thanks to the CPSU/UH secretary, Deborah Weiner; for her assistance in the editing and publication of these Proceedings, and Maura Garibay and Dora Tsuha for secretarial assistance.

C. W. Smith Director, CSPU/UH DESCRIPTIVE SUNMARY OF A NORTH KONA BURIAL CAVE ISLAND OF HAWAII

M. S. Allen and T. L. Huntjr Mauna Loa Field Station Hawaii Volcanoes National Park Hawaii 96718

The science of archaeology has had a continuing interest in the mortuary customs of human cultures. Burials and their associated material items lend insight into certain aspects of human behavior. Mortuary patterning of a particular culture may provide data on social structure, social ranking, religious ideology, artistic expression and utilitarian craft. Both prehistoric and post-contact Hawaiian burial customs have a limited representation in archaeological recording, analysis and interpretation. It was with this in mind that the authors conducted an archaeological survey of the North Kona burial cave in June 1975 to record 16 primary -and secondary burials with the accompanying well-preserved material items. It appears that both prehistoric and historic mortuary patterning are represented in this cave. Our immediate objective was a preliminary recording of the site prior to any vandalism or other burial disturbance. Field recording included a sketch map, complete burial descriptions, and photographs in both color slides and black and white print film. No samples of any kind were collected. Bowen's (1961) thesis, the only summary of Hawaiian disposal customs, provides several ethnohistorical data pertinent to our discussion. Bowen lists ten prehistoric methods of Hawaiian disposal: 1. exposure to the elements 2. cremation 3. sea or fresh water disposal 4. pit burial 5. monument interment 6. earth burial 7. cist interment 8. heiau burial 9. house floor burial 10. cave interment

*T. L. Hunt, Dept. of Anthropology, University of Hawaii at Hilo, Hilo, Hawaii 96720. Our paper deals specifically with cave interment. In brief, Bowen concludes the following regarding cave interment: Caves were convenient places of disposal for the general population, with protection against the elements and molestation. Ethnohistorical sources suggest that the common people, maka'ainana class, remained attached to their land despite warfare and changes in ruling ali'i. Therefore, family caves were probably used continuously over long periods of time. In most cases, disposal caves were set apart exclusively for the dead and used for no other purpose. Several ethnohistorical sources outlined by Bowen provide data of considerable variation, and some contradiction, as to corpse positioning (e.g., extended vs. flexed) used in cave interment. Nearness of caves to dwelling areas does not seem to have prevented their use in disposal. Bowen also mentions that caves were used for all types of human disposal well into the Christian period. Bowen continues that historically, commoner caves seem to differ from ali'i caves in the following ways: though commoner caves were carefully hidden, the location was known to the entire family and not just to guardians. They were ancestral and used for generations, whereas ali'i caves were concealed and their location lost with the death of the kahu, or guardian. Ali'i remains consisted of a light, compact bundle of bones (secondary burial). The entire body of commoners went into the cave with some quantity of associated objects. Ali'i, in contrast, had a few associated materials. Commoner caves were kept in repair, altered and improved by those families to whom they belonged. Ali'i caves were abandoned and never spoke of once the remains were deposited. Archaeological survey and recording of sites has provided further data concerning Hawaiian disposal caves. Bowen summarizes the archaeologically known geographical distribution of disposal caves as follows: disposal caves have been reported from Necker, Nihoa, Kaua'i, O'ahu, Moloka'i, Maui, and Hawai'i islands. Only legendary references have come from Ni'ihau, LE-na'iand Ka-ho'olawe. It is evident that disposal caves are widely distributed throughout the major islands of the Hawaiian chain from Necker in the northwest to Hawai'i in the southeast. Lava tubes, niches, and crevices are found throughout the islands and have been utilized extensively for disposal. Both the immediate and general area surrounding the North Kona burial cave are rich in archaeological remains, indicative of the once extensive human occupation. In the immediate vicinity are historic petroglyphs dated from the 18701s,additional burial sites, numerous living caves containing quantities of midden, prehistoric fish ponds, an ancient coastal trail, a historic horse trai1,and several brackish water ponds. The adjacent ahupua'a contain several documented coastal settlements. Inhabitants of this area probably relied heavily on maritime resources. Numerous tide pools, pebble beaches and a lagoon characterize the coastline. Several brackish water ponds in the area would have provided a fairly adequate water supply. We can not be sure what plant resources were available to the inhabitants. Today the area is heavily infested with feral goats, and kiawe (Prosopis pallida) predominates. Typical coastal plants such as lau hala (Pandanus odoratissirnus), 'ilima (Sida fallazi and native grasses were probably once present. The cave contains sixteen primary and secondary burials. Upon entering the cave, in a short spur off to the right are the scattered incomplete remains of an individual which we designated as burial #I. No artifacts accompany the skeletal material. Burial #2 is the scattered remains of an infant, placed on a small ledge protruding from the cave ceiling. Lau hala mat fragments are also present. Burial 113 is a wooden casket lying on the cave floor in a short spur off to the right of the main chamber. The casket is made of wood planks, the sides being of a single piece, the top and bottom from three planks each. Both square and round-headed nails were used in construction. Large fragments oflauhala mat are under the casket and perhaps once surrounded the entire box. Inside the casket the majority of the remains are unidentifiable. The body appears to have been wrapped in a blue and white plaid quilt and a patterned cloth, possibly linen, was tied around the head. Accompanying the body is a broken shell comb, a large shell button, two wood weaving shuttles and two metal cans. The head lies on a cloth pillow. Burials #4 through $112 are situated on a large ledge, 1.3 meters above the floor, that extends out into the main chamber. All nine are extended burials. Burial $14 is on the outermost edge of this ledge. The body had been placed on a litter of eight poles; some have crudely rounded ends and others are pieces of planks with squared edges. The ends of the poles are tied with a narrow twisted cordage and a plaited coconut mat rests directly on top of the poles. Above the coconut mat is a lau hala mat, possibly double weave, that appears to have once enshrouded the entire body. A fine white tapa with watermark patterns was placed over the lower part of the corpse. Also associate with the burial was a thicker tapa of red and blue designs on white. The patterns included red and blue stripes, blue six- pointed stars and connected rows of chevrons. Tapa of a similar design was tied around the pelvis region, presumably a malo. European cloth was also present, wrapped around a long bone. Lying on the floor directly below the ledge was a lau hala pillow, probably associated with burial $14. The pillow was not woven but consisted of strips of lau hala, bound together with a piece of blue cloth and stuffed with rolls of lau hala. Next to burial 84 is Feature 1, an empty curved plank of about three meters length. Perforations mark both sides of the lower end of the plank and a fragment of knotted cloth remains inside one. Lying directly under the plank is a red bandana. This feature may have been a lid or top for the adjacent burial, #5. Near to burial 84 and feature 1 is a lone skull resting in a gourd fragment. It is not apparent which burial it is associated with. Burial #5 lies in a canoe half, possibly of Koa (Acacia koa). The open end had been closed off with a snugly fitting plank and the underside had been burned. The canoe rests on top of a lau hala mat. There is evidence of a canoe patchwork but also a number of unmended holes. The skull rests on a pillow and white cotton cloth is wrapped around the pelvis, ribs and near the skull. Brown tapa fragments lie near the feet. Burial $/6 lies in a canoe section. Both ends are open and a piece of cordage is tied around the head end. Lying near the feet are remnants of a straw hat, a pair of leather shoes, and a thin copper ring. The body is placed on top of a lau hala mat and cloth covers the lower portions. Under the canoe lie several thicknesses of lau hala mat. Burial #7 is on a slightly curved, crudely roughed out board of soft, unidentified white wood. Burial 117 represents the only example of this type of wood. Lau hala mat accompanies the disturbed skeletal remains. Burial //8 is also a canoe section with the head end open and a plank across the foot end. The canoe had been mended with a piece of wood lashed across the hole. A fragment of blue'h1oha"print cloth lies near the skull, possibly of recent introduction. Burial //9 is situated almost directly under 88. The skeleton lies on an upcurved board that is much wider than any of the others examined. A board closes off the foot end and a piece of cordage runs across the top of the head end, through a hole on each side of the plank. The skull lies on a pillow and both tapa and European cloth fragments are present. Under the canoe is a lau hala mat, possibly of double weave. The next two corpses, while designated $/I1and 1/12, appear to be a double burial. On a single litter, of five poles across, lie two skulls, two pelvises and a number of limb bones. Associated grave goods include large lau hala mat fragments and three distinct types of European cloth. Burials 1/13, 14 and 15 are located further back in the cave. They consist of three skulls and various post- cranial bones lying on a slightly curved plank. A lau hala pillow, lau hala mat fragments ar:d tapa fragments are found in association. Burial #16 is located near the end of the cave amidst rooffall and rubble. It includes scattered post-cranial bones and no skull or fragments. Lau hala and tapa fragments are also present. At the very back of the cave, lying under rooffall, we also found gourd fragments. Tentative Conclusions No formal chronology has been established for disposal caves in the Hawaiian Islands. However, numerous disposal caves contain a significant amount of post-contact materials which indicate extensive use into the 19th century. No absolute dates were obtained from the cave, but grave goods suggest a temporal sequence from the prehistoric into the post-contact, or Christian, period. We estimate an approximate 150 year range for use of the cave, beginning c. A.D. 1750 to c. 1900. These data lend insight into the dynamic acculturation process occurring throughout the Hawaiian Islands in the 19th century. From ethnohistorical data previously sumarized, we would assume the cave to represent exclusive 'ohana, or extended family use, by members of the maka'iiinana social class. The temporal range accounts for the variety of mortuary patterning and associated grave goods. In addition, the botanical remains imply expansive environmental exploitation, i.e., a wide range of microenvironments under utilization. Groups living in these leeward coastal areas not only exploited the rich maritime ecosystem, but also the upland areas, including the native forest. Formal ethnobotanical analysis can lend further insight into environmental exploitation and utilitarian craft. It is our recommendation that further field research be undertaken primarily involving physical anthropology. Age, sex, anthropometric, morphological, and paleo- pathological data can be recorded from these well-preserved skeletal remains. Zt is of utmost importance that the cave be protected from vandalism, theft, and other disturbances.

Acknowledgements We would like to thank the following people for their assistance: Mr. William J. Bonk, University of Hawaii; and Mr. Clyde Wong, former curator of Lyman Museum, and Ms. Virginia Goldstein, Hawaii County Planning Department, for their he1.p during preliminary examination of the site. Thanks also goes to Ms. Beatrice Krauss of Lyon Arboretum, for help with ethnobotanical identification, and to Dr. Patrick McCoy, Ms. Monica Udvardy,and Mr. Larry Tirona for additional assistance.

-References Bowen, R. 1961. Hawaiian disposal of the dead. Unpublished Master's Thesis, University of Hawaii, Honolulu. Buck, P. (Te Rangi Hiroa) 1957. Arts and crafts of Hawaii. B.P. Bishop Museum Special Publication 45, Honolulu. Kooijman, S. 1972. Tapa in Polynesia. B.P. Bishop Museum Bulletin 234, Honolulu. Neal, M.C. 1965. In gardens of Hawaii. B.P. Bishop Museum Special Publication 50, Honolulu. KOA AND LEHUA TIMBER HARVESTING AND PRODUCT UTILIZATION: RELIGIO-ECOLOGICAL RELATIONSHIPS IN HAWAII, A.D. 1778" Russell A. Apple State Director's Office National Park Service Honolulu, Hawaii 96813

Although koa and 'Ehi'a-lehua trees were abundant on all the major islands in the Hawaiian chain during prehistoric and protohistoric times, many oral traditions and written references indicate that their survival in quantity was insured by religious taboos, by the limited products manufactured from wood,and by the limited number of those authorized to utilize these products.

1 Ohi'a-lehua- was reserved almost exclusively for temple structures and residences of high chiefs and temple images, where relatively short, thin, hard, long-lasting wood was essential and where the long, parallel fissures inherent in this wood did not interfere with the use. From selection, harvesting, transportation, and working of a log until its final placement, 'Ehi'a-lehua was treated with religious ceremony and taboos. So greatly was it revered that central temple images made from the wood often required human sacrifices. Unlike '8hi1a,koa could grow to a height of 60 feet before branching. Logs of this straight, fine-grained wood were restricted to use by chiefs for the manufacture of canoes Like 'Ehi'a, it was venerated. Every ritualistic step in the creation of a canoe was handled by canoe-making priests, and these priests were protected by the goddess of canoe builders. Koa and lehua were used in the manufacture of sacred items. Use was limited to chiefs; religious ceremony was followed throughout harvesting and fabrication; color of the woods and names imply spiritual significance. Koa and 'Ohi'a-lehua products were considered suitable for tax payments, gifts, and sacrifices. It is reasonable to assume that sanctions were imposed to conserve and protect these trees. THE OCCURRENCE AND ECOLOGICAL SIGNIFICANCE OF METALLIC SKINKS ON THE ISLANDS OF HAWAI'I AND KAUA'I Ken Baker Mauna Loa Field Station Hawaii Volcanoes National Park Hawaii 96718

In 1972 a lizard was given to me for identification which had been captured by a housecat in Volcano, on the island of Hawai'i, at 3700 feet elevation. It was a metallic skink, Lygosoma (~eiolopisma)metallicum (O'Shaughnessy), which was of interest since the species was new to the island, and to an area previously unoccupied by reptiles. Subsequently, in 1972, I collected two more of the skinks, and another eighteen in 1976. The metallic skink is an alien import from Australia where it occurs around the port cities of Melbourne and Hobart which may offer explanations for its accidental importation to Hawai'i (Barwick, personal communication). A survey of the literature shows that metallic skinks in Hawai'i have been known only from O'ahu where they are considered abundant (Fisher 1948, Oliver and Shaw 1953, Hunsaker and Breese 1967). The first record in the literature for a metallic skink in Hawai'i was a solitary specimen collected in Honolulu in 1919 (Brongersma 1942). Recently, however, I found a specimen in the Bernice P. Bishop Museum collected in 1909. This, so far as is known, is the earliest record for a metallic skink in Hawai'i which indicates the species probably arrived on O'ahu sometime around the turn of the century, likely in shipments of material from Australia. The material in the Bishop Museum also turned up an unreported collection of metallic skinks from the head of Ka-lalau Valley, Kaua'i, at an elevation of 4000 feet. Subsequent to that find I had the opportunity to be on Kaua'i and was able to collect five additional skinks from the head of Ka-lalau Valley, plus an additional fourteen specimens from the adjacent Wai-mea Canyon. Both island records become new to the literature, and both establish altitudinal records for Hawaiian reptiles at 3700 feet and 4000 feet. The other nine species of lizards presently known in Hawai'i inhabit coastal and lowland areas up to about 2000 feet elevation, with specimens of the snake-eyed skink, Ablepharus boutoni poecilopZeurus (Wiegmann), having been taken in the Ka'G Desert at 3200 feet (Fisher 1948). The Volcano population may be less than 10 years old as none of the skinks had been noted, reported on,or collected before 1972. The Kaua'i skinks were first collected in 1963 though they could have escaped the attention of collectors for a number of years prior to that time. It appears, however, that the Hawai'i and Kaua'i populations are much more recent than O'ahu skinks. While population densities for metallic skinks have not been determined in the three island locations, Oliver and Shaw (1953) and Hunsaker and Breese (1967) state that on O'ahu the lizards are abundant and that their rapid multiplication has produced "remarkable" numbers in the lower elevations. Oliver and Shaw (1953) counted 44 skinks beneath a single banyan tree. My observations in Volcano indicate the skinks are only sparsely distributed in an area of a few blocks. In comparison they are much more abundant on Kaua'i where they are common along trails and the shoulders of roads where they venture out into open sunlight. Numbers diminish in denser forested areas adjacent to roads and trails. It appears that on Kaua'i the skinks are dispersing through forested areas by moving along trails and roads and through open areas of forest. This same method of dispersal could be occurring in thevolcano population, which suggests that the skinks will eventually move into Hawaii Volcanoes National Park. Two other skink species, the moth skink, Lygosoma noctua noctua (Lesson), and the snake-eyed skink,already occur within the park, principally in lowland elevations. Scale counts and body measurements of the Volcano and Kaua'i specimens compare favorably with metallic skinks found on O'ahu, and in Australia, except the Hawaiian skinks are slightly smaller than Australian forms. Larger females in Australia reach 50 uun in body length (Cogger 1975) whereas females in Hawai'i reach 46 mm. Within the Hawaiian skinks there are additional variations in body size. Sexually mature females in the Kaua'i and Hawai'i populations are slightly larger than O'ahu females (Table 1). Smallness on O'ahu could be due to population densities causing inter- and intraspecific competition among individuals for available food supplies. This would also inhibit egg production. Numbers of eggs per female of the O'ahu population are lower than numbers of eggs of females on Kaua'i and Hawai'i. Adult females from O'ahu measured by Oliver and Shaw (1953) averaged 39.8 mm in body (snout-vent) length, while those measured from Volcano and Kaua'i averaged 41.2 mm and 41.8 m respectively. These measurements are averages of 1.4 mm and 2.0 mm more than O'ahu females. The numbers of eggs reported by Oliver and Shaw (1953) for 15 females from O'ahu averaged 2.9 eggs per female (Table 1). None of the O'ahu females carried more than 4 eggs each. The numbers of eggs for 8 females from Volcano averaged 4.5 per female (Table 1). Four specimens (50%) had more than 4 eggs each, while none had fewer than 3 eggs. The numbers of eggs for 9 females from Kaua'i (Table 1) averaged 4.1 eggs per female. Three specimens (33%) had more than 4 eggs each, while none had fewer than 3. In animal species the numbers of eggs carried within an oviparous species varies to some extent with individual body size. Larger individuals with more internal capacity lay more eggs than smaller individuals of the same species. This phenomenon is summarized by Lack (1954) which shows that egg production increases with the age of females, when older females are larger. In these studies it was observed that larger females did carry more eggs. In the Volcano population body lengths versus numbers of eggs carried averaged 41.2 mm and 4.5 eggs (Table 1). In the Kaua'i population body lengths versus numbers of eggs averaged 41.8 and 4.1 (Table 1). When body lengths versus numbers of eggs carried from Kaua'i and Volcano females are combined (individual measurements for O'ahu females not available from data in Oliver and Shaw [19533), the number of eggs in metallic skinks is shown to increase relative to body size; and,on an average, body lengths increase 1.6 mm for each additional egg. By comparing egg volume with body size it was found that it takes an increase in body length of at least 1.2 mm to accomodate each additional egg, a measurement within the 1.6 mm figure given above (see also Figure 1) . In animal species there is also a correlation between reproductive capacity and food availability. Simply stated, if an animal is eating only enough to maintain its own tissues, it can either produce eggs and starve to death, or it can stop reproducing and survive. On the other hand, if food is in excess of nutritional needs, and there are no other limiting factors to reproduction, then reproductive capacity can be expected to reach its maximum potential. There will of course be variations as food availability and egg production fluctuate. The Volcano and Kaua'i skinks appear to have sufficient food, feeding principally upon various Coleoptera, Corrodentia, Diptera, Embioptera, Homoptera, Hymenoptera, and Lepidoptera. Invertebrate foods other than insects eaten are Annelids, Arachnids and Symphillids. All of the skinks collected from Volcano and Kaua'i were collected near midday and stomachs of all specimens seemed full, indicating the skinks were able to catch what food was needed within the first half day of foraging. It appears, therefore, that food is sufficiently available in these populations at the present time for optimum growth and reproductive effort, which is 46 mm in body length and 6 eggs. Because of probable space 'and food limitations, the more densely populated O'ahu skinks are smaller in size and carry fewer eggs. The average length of females from O'ahu is 39.8 mm with an average of 2.9 eggs each, as indicated by Oliver and Shaw (1953). However, Oliver and Shaw state that maximum body length for O'ahu skinks is 44 nun, which places females of that size within the ranges in Figure 1 showing a capability of carrying 5 or 6 eggs each. That none of the O'ahu skinks studied by Oliver and Shaw carried more than 4 eggs each--even though they had body size to carry more--indicates that other factors are inhibiting the maximum reproductive potential. These are probably overcrowding of individuals and limited available food supplies. The scenario I would like to present as a conclusion to this paper is that an alien species arrived in the Hawaiian Islands (in this case on O'ahu) and was limited in its distribution at first to a single island. Its accidental introduction was successful; the species was adaptable; and it experienced rapid dispersal. In the relatively short time of a few decades its numbers became exceedingly abundant and the species became a major part of the island fauna. In time, however, the species saturated its habitat and inter-specific competition for diminishing food supplies played its role. Body size decreased and egg production diminished to come into balance with the pressures of limiting factors. Subsequently, the species was imported to other islands, into new situations where the species had adequate space and food while population densities were low. The newly established populations "exploded" and dispersal was rapid. On the second island invaded (in this case Kaua'i), the species spread over a major portion of the island in, perhaps, only two to three decades. Again, it became a major component of the exotic vertebrate fauna of that island. On the third island of introduction (in this case Hawaili),the invasion is recent, but indications are that the species will again run rampant in population growth and dispersal and become another major element of the fauna. Based on knowledge of the species on the previous two islands it can be predicted that the species will become common in the Headquartersl~z-lau-eaMilitary CampIMauna Loa Strip area of Hawaii Volcanoes National Park, through natural dispersal or accidental import by persons commuting daily between Volcano and the Park. The example exhibited by the occurrence of metallic skinks in Hawai'i adds to our existing knowledge of how alien species can become established in new island habitats and niches and flourish or "explode" in the absence of environmental restraints.

Acknowledgements I wish to thank Mrs. LaVieve Forbes for showing me the first skink from Volcano, and Volcano residents Bill Mull, Terry Parman, and Rick Warshauer for information on the occurrences of skinks in vicinities of residences. I wish also to thank Len Anderson for mathematical computations of egg and body sizes; Cliff Davis for determining invertebrate foods; Alan Ziegler of the Bishop Museum for assistance with its collections; and R. Barwick of the Australian National University, Canberra, for information on Australian skinks. BODY LENGTH (mm

FIGURE 1. The average body length of metallic skinks compared with the clutch size. The horizontal bars represent the range of body lengths; the boxes represent one standard deviation on either side of the mean.

TABLE 1. The average body length and clutch size of female metallic skinks from three separate populations in Hawai'i. Population Average body Average Clutch location length (mm) clutch size range

O'ahu 39.8 * 2.9 rt 1.3 1-4

Volcano 41.2 rt 2.9 4.5 rt 1.1 3-6 "Individual body lengths not available for O'ahu. Average of 39.8 mm taken from Oliver and Shaw (1953) who measured 17 females but did not give individual statistics. Literature Cited Brongersma, L.D. 1942. Notes on Sincid lizards. Zool. Mededeelingten 24(1-2):125-152. Cogger, H.G. 1975. Reptiles and amphibians of Australia. A.H. and A.W. Reed, Sydney. 584 pp. Fisher, H.I. 1948. Locality records of Pacific Island reptiles and amphibians. Copeia No. 1, April 15. Hunsaker, D., and P. Breese. 1967. Herptofauna of the Hawaii Islands. Pacific Science XXI (July):423-428. Lack, D. 1954. The natural regulation of animal numbers. Oxford Press. 343 pp. Oliver, J.A., and C.E. Shaw. 1953. The amphibians and reptiles of the Hawaiian Islands. Zoologica 38(2) NOS. 5-8:65-95.

Addendum Subsequent to presentation of this paper it was learned through correspondence with skink specialists in Australia that the species in Hawai'i is not believed by them to be Lygosoma (LeioZopisma) metaZZicum but an undescribed race of LeioZopisma delicata, a common eastern Australian species which encompasses a species complex with several presently undescribed forms. In addition, a recent revision of the lygosomine skinks (Greer, A. 1974. Aust., J. Zool., Suppl. Ser., 31:l-67) now places L. deZicata in the genus LamprophoZis. Furthermore, A. Loveridge ("ANew Skink LeioZopisma hawaiiensis, from Honolulu." Proc. Biol. Soc. Wash. 52:l-2. 1939) described the Hawaiian form nearly 40 years ago as an endemic species. Oliver and Shaw, however, believed the name a misnomer and relegated L. hawaiiensis to synonymy under L. metaZZicum in their 1953 review of the Hawaiian reptiles and amphibians. Other investigators since that time followed Oliver and Shaw. The present taxonomic status of the Australian skink in Hawai'i is, therefore, in a state of confusion. It is apparently an undescribed subspecies known from Hawai'i but whose origin and distribution in Australia are unknown. Researchers have not yet found it in Australia. It appears, however, that the species name hawaiiensis of Loveridge may have precedence and would become the subspecific name, Lampropholis deZicata hawaiiensis, of the Hawaiian populations. Not only is the skink a good study in how alien species become established in island habitats, but it is an excellent example of a displaced animal known only from a foreign habitat, and the confusion which can occur in taxonomic research. STUDIES ON THE ENDEMIC HAWAIIAN GENUS: HIBISCADELPHUS (HAU-KUAHIWI) Ken Baker and Suzy Allen Mauna Loa Field Station Hawaii Volcanoes National Park Hawaii 96718

Two species of the genus Hibiscadelphus (Malvaceae) are endemic to the island of Hawai'i. In 1909 the genus was discovered by Joseph Rock, represented by HibiscadeZpkus huaZa2aiensi.s on the slopes of Hualslai, North Kona. At the time of discovery about a dozen trees were in existence; today only two of these trees survive. In 1910, Rock discovered the species Hibiscadelphus wilderianus on the slopes of Hale-a-ka-15, island of Maui. The species was known from a single tree which died shortly thereafter. Hibiscadelpkus giffardianus, endemic to Kipuka-pua-ulu in Hawaii Volcanoes National Park, was also known from a single tree, found by Rock in 1911. A fourth species, Hibiscadelphus bombycinus is known from a mislabeled herbarium specimen, correctly identified by C. N. Forbes in 1920. Recently, a fifth species was discovered in Wai-mea Canyon on the island of Kaua'i. Hibiscadelpkus distans Bishop and Herbst is represented by six trees. In the 1950's, efforts were made to save both Hawai'i Island species from extinction. Cuttings and seedlings of each were planted in close proximity in ~ipuka-pua-ulu. Seedlings propagated from the Kipuka-pua-ulu H. giffardianus trees were planted in both Kipuka-k? and ~i~uka-pua-uluin the late 1950's or early 1960's. In 1973, while studying rat damage to some of the trees, differences were noted in flower and leaf morphology which led to the discovery of hybrids. These were described by Baker and Allen (1976) in Phytologia as Hibiscadelphus x puakuahiwi.

Briefly, the differences between H. giffardianus and H. kuaZa Zaiensis are as follows : H. gi ffardianus flowers range in length from 48 mm. to 75 nun. H. kualalaiensis flowers range from 33 mm. to 59 m. H. giffardianus flowers have a tendancy towards right-handed whorls of petals. Conversely, H. huaZaZaiensis flowers tend to have left-handed whorls. H. giffardianus blossoms are magenta (rarely splotched with green) throughout their three-to five-day life span. H. hualalaiensis blossoms are yellow-green when one to three days old but may turn purplish in four- to five-day old blossoms. The average number of bracts on a H. giffardianus flower is 6.16 and they are usually more than 20 m, in length. H. kualaZaiensis bracts average 5.02 in number and are usually less than 3 nun. long. H. giffardianus leaves have a tomentum in the angles of the veins on leaf undersides, while on H. hualalaiensis leaves the tomentum is almost non-existent. F1 flowers display characteristics intermediate between both parents and F2 flowers exhibit the typical mendelian range of characteristics. The condition of anthers and pollen may sometimes distinguish F2 hybrids from parents. Anthers of parent and F1 taxa dehisce normally, exposing amber colored pollen, whereas Fz anthers sometimes do not dehisce at all and contain abnormally small pollen grains which are devoid of cytoplasm. Carr and Baker (in press) have determined through meiotic analysis that some F2's are little affected by hybrid breakdown while others exhibit moderate to severe meiotic disturbances. As mentioned before, rat damage to the trees led to the discovery of the hybrid. However, rats are not affecting all the HibiscadeZphus trees in the park. In ~Epuka-kirats are chewing through flower calyxes to reach the nectar. Over a period of four and one-half weeks in April and May, 317 flowers were collected from under KK-HX-F1-1 and KK-HX-F2-5. Two hundred and thirty-nine or 75% were rat damaged. Flowers were again collected in June, July, and August for a seven week period. Of the L967 flowers examined, 538 or 27% were rat damaged. Preliminary observations seem to indicate a correlation between daily rainfall and rat utilization of HibiscadeZphus nectar. There are several HibiscadeZphus trees in the park which do not have rat damage of this particular type. Three relatively isolated trees in ~i~uka-pua-uluhave been damaged by rats stripping the succulent bark, probably in efforts to obtain moisture. Warfarin stations set up in July have sufficiently curtailed rat depredations of this sort. It appears that we have groups of rats displaying varied feeding behaviors. Other researchers report similar differentiated behavior in rats. We hypothesize that a particular learned behavior is being exercised by the rats.

The H. giffardianus flowers produce copious amounts of nectar; up to 5 cc. in a three- to five-day life span. On a daily basis, other Hawaiian flowers we have examined do not contain such quantities of nectar. FI flowers produce slightly less. Today the flowers are frequented by Japanese white-eyes, which usually probe the flowers from the top of the corolla, 'amakihi which enter through the side of the corolla just above the calyx, and more rarely by 'i'iwi, which have long beaks favorably adapted for apical probing. Examination of the now extinct Hawai'i island mamo and 'akialoa show a remarkable similarity between bill curvature and corolla curvature. It ispossible that the two evolved together. The genus HibiseadeZphus offered the aboriginal Hawaiians a number of otential uses. Ms. Beatrice Krauss of Lyon Arboretum hasEb---- een working with us in this respect. Her experiments have shown that H. huazalaiensis produces a cordage of strength and sheen equal or superior to that of hau (Hibiscus tiliaceus--the preferred cordage material of the Hawaiians). Experiments with H. giffardianus are not complete. She is also working with the bark of both species for potential tapa manufacture. As mentioned, the flowers produce copious amounts of nectar. Although we have no historical accounts of such, it is conceivable that the Hawaiians were aware of this. and collected the nectar for consumption, as it is quite tasty. The nectar could have been used as a sweetening agent, as was banana nectar, or as a beverage. There are historical accounts of Hawaiian bird catchers using flowers with long corollas as lures. A blind was constructed of limbs and leaves. The bird catcher would then hide inside the blind and when a bird landed to drink, the catcher would pinch the corolla shut, thereby catching the bird by the bill. It is possible that aboriginal Hawaiians utilized the HibiscadeZphus bark which led to the tree's decline. Another thought is that the trees were already rare and/or were not used for tapa or cordage because of their unique nectar qualities. We are continuing Hibiscadelphus studies concentrating on rat damage and bird utilization. This summer H. giffardianus flowers were bagged and hand-pollinated to obtain pure seed for future propagation of the species both in Hawai'i and world arboreta. We have made an attempt to notify world arboreta and herbaria of the existence of the hybrids and when possible identify misidentified specimens. Literature Cited

Baker, K. and S. Allen. 1976. Hybrid HibiscadeZphus (Malvaceae) from Hawaii. Phytologia 33(4):276.

Baker, K. and S. Allen. Hybrid HibiscadeZphus (Malvaceae) in Hawaii. Manuscript in press.

Carr, G. D. and J. K. Baker. Cytogenetics of HibiscadeZphus (Malvaceae): A Meiotic Analysis of Hybrids in Hawaii Volcanoes National Park. Manuscript in press. SAVING THE 'ALA~(CORVUS TROPICUS)-- SOME PRELIMINARY MANAGEMENT RECOMMENDATIONS Paul C. Banko Hawaii Volcanoes National Park Hawaii 96718

The 'AlalS (Corvus tropicus), the Hawaiian Raven or Crow, is endemic to the island of Hawai'i and is the only representative of its genus in the mid-Pacific region. The 'Alala is the only member of its family, Corvidae, to be considered an endangered species (Vincent, 1966-1971). 'AlalB are traditionally known only from the districts of Kona and Ka'a, although during the last 25 or 30 years sightings have occasionally been reported far outside of this range during the non-breeding season (W. E. Banko, ms). Within this unusually restricted range, 'AlalH were once relatively abundant, as noted by the naturalists of the 1880's and 1890's (Wilson and Evans, 1893; Henshaw, 1902:36; Perkins, 1903:372-3; Munro, 1944:70), but had apparently begun to decrease in some localities before 1900 (Perkins, 1903:372) By the 1930's the population was thought to have become greatly reduced in numbers (Munro, 1944:70) and by the 1940's had disappeared from portions of their known range (Baldwin, 1969:44-45). Since the 1940ts,the population has continued to decline. Interviews and discussions with long-time residents of Kona and Ka'ii reveal that 'Alal2 were relatively common, or at least consistently present, about 25 to 50 years ago in areas where today there are none or only a few still surviving. 'AlalH population surveys were conducted by W. E. Banko from 1969 to 1973 (with P. C. Banko assisting in 1973 under National Parkservice sponsorship) as part of the Endangered Wildlife Research Program of the U. S. Fish and Wildlife Service. Field studies were continued by P. C. Banko with funding by World Wildlife Fund and sponsorship by Hawaii State Division of Fish and Game in 1974, funding by Hawaii Natural History Association in 1975, and funding by World Wildlife Fund and U. S. Forest Service in 1976. The most significant accomplishment of the 1969-1976 'AlalH surveys has been to document the extremely precarious status of the population. Except for 1974, when about 65 adults, immatures, and fledglings were found, the known population in North and South Kona districts has totaled less than 50 individuals during the last eight years. Although 'AlalB have been reported in or near the district of Ka'ii since 1969, there has been no recent evidence of a resident breeding population there despite extensive coverage of Ka'ii Forest Reserve by U. S. Fish and Wildlife Service in 1976 (J. M. Scott, pers. corn.). Not only has the population hovered at a very low level during the past eight years, there is no indication that the population is growing--even slowly--or that new individuals are being recruited into the breeding population at a rate which can perpetuate the species for more than perhaps ten years. The stagnant condition of the populationis apparently not the result of low production of offspring, but may instead be due to low survival of immatures within the first two years of life. The minimum numbers of fledglings which have been produced annually during the last four years are: nine in 1973, eleven in 1974, ten in 1975, and ten in 1976 (total of forty). Of the thirty offspring produced before 1976, sixteen have been banded, but only three banded birds have been seen in succeeding years. If permission can be obtained from the U. S. Fish and Wildlife Service to continue the banding program, sufficient data may be collected to definitively determine the rate of fledgling survival and recruitment into the breeding population. There are many factors which may have had a role, at least temporarily, in the decline of the 'Alal3 population, including habitat alteration and destruction (Perkins, 1903: 372; Baldwin, 1969:45), food competition (W. E. Banko and P. C. Banko, ms.), disease (Baldwin, 1969:44; P. C. Banko, 1974), parasitism (P. C. Banko, 1974), predation (P. C. Banko, 1974) and shooting (Munro, 1944:70). Habitat alteration and destruction probably threaten the existence of the species more than any other factor at the present time, but disease, parasitism,and food competition should also be studied further to determine the extent of their impact. Supporting the claim that habitat quality vitally affects 'AlalZ is the apparently low rate of recruitment of immatures into the breeding population despite a relatively high annual production of fledglings. Having survived the diseases, parasites, and predators to which they should be most vulnerable as nestlings, immatures may be limited by the availability of food once they become independent of parental care. If habitat becomes degraded to the point where inexperienced birds are unable to cope on their own, it should not be surprising that new breeding territories do not become established readily. Two other facts point to food availability as a potentially important factor in limiting the 'AlalB population. First, in nests where there are two offspring (the maximum number which have fledged from any nest found to date) one youngster is usually noticeably weaker than its sibling, indicating that it has been receiving less food from the parents. Second, fledglings are heavily dependent upon the parents for food many weeks after leaving the nest, indicating that self-sufficiency may be slow to develop because of marginal habitat conditions. Since habitat quality may be a crucial factor in the survival of the species, sufficient forest land in Kona and Ka'ii must be protected from continued deterioration so that it may recover as far towards a pristine condition as is possible. To this end, domestic and feral ungulates must be strictly controlled or eliminated throughout large tracts of forest. However, since most of the remaining 'Alalii are resident on private and State lands used primarily for cattle grazing, it may be politically and economically difficult to purchase or withdraw such lands for wildlife sanctuaries in the near future. If in fact sanctuaries cannot be established within the existing range of 'Alalii, efforts must be directed towards three goals: 1) aid the population in their present range by providing supplemental food at feeding stations, 2) take individuals from the wild for captive breeding and release into protected habitats, 3) establish sanctuaries in the most suitable portions of former 'AlalZ range for eventual restoration of 'Alala populations. If 'Alalii can be attracted to feeding stations, higher nestling and fledgling survival may result. Also, newly formed breeding pairs might be able to pioneer additional breeding territories where feeding stations were maintained. Wild and cultivated fruits might be the most suitable items to provide as food, but the omnivorous appetite of 'Alal3 could possibly be capitalized on with a variety of other foods, so long as spoilage and other health hazards could be avoided. For example, if small birds were provided as food items it would be important to insure that their parasites and diseases would not be transmitted to 'Alalii. At best, feeding stations could be expected to provide only temporary relief to the existing wild population, but enough time might be gained to allow some recovery of key habitats where ungulates had been eliminated or reduced and where captive or wild-reared offspring could be released or established profitably. In addition, if more 'Alalii were being supported in the wild with supplemental food, more individuals could be justifiably taken for captive breeding purposes. It is desireable to limit to as few as possible the number of wild birds taken annually for captive breeding stock so that the wild population is not suddenly and drastically reduced. Established breeding pairs should not be taken from the wild, at least until it is shown that individuals taken as juveniles cannot be successfully bred in confinement. 'AlalB are rigorously traditional in maintaining the location of their nest sites within their breeding territories and may not accept a new breeding location or situation (such as an aviary), thereby wasting their reproductive potential. The best sources of captive breeding stock are "excess" nestlings, where one or more youngsters can be taken while leaving one sibling in the nest, and sick or injured individuals which can be restored to health. If providing supplemental food near nest sites increases the number of nestlings which can survive, there would be every justification for taking these "bonus" youngsters. Although it is a less desireable alternative than taking "excess" young, additional birds could be obtained for captive stock by taking up to about one-third or possibly one-half of the annual production of wild fledglings with the justification that there appears to be low recruitment of young into the breeding population under present conditions. The successful release of captive-reared birds may be more difficult than actually breeding 'AlalZ in captivity. One obvious way to facilitate the release of captive-reared fledglings into wild habitats would be to locate the breeding operation within the release area so that from the time that they are hatched, the young are familiar with the immediate area of release. In any case, captive-reared birds should be released into the wilds "gently," that is by gradually reducing their dependency on humanly provided food and shelter. The genesis of a captive breeding program exists at PBhaku-loa where three 'AlalZ, which were taken from the wild in 1973 and held at Hawai'i Volcanoes National Park until 1976, are being maintained by Hawaii State Division of Fish and Game. Pending approval by the U. S. Fish and Wildlife Service, the National Park Service has proposed to initiate a captive breeding Frogram in which the offspring would be released into the Krpuka-pua-ulu/Kipuka-kula-lio area of Hawaii Volcanoes National Park. The Park is the only area on the island presently being managed to reduce or eliminate ungulates. Although the area of former 'AlalB habitat within the Park is relatively small, the protection which is assured under National Park Service policies will make it valuable in the overall effort to save 'AlalE from extinction. By the time of the second Natural Science Conference there will hopefully be news of steady progress being made on behalf of the 'AlalZ by federal, state and other concerned parties. It is also hoped that field research can continue to provide information which may aid the effort to preserve this unique Hawaiian bird. Literature Cited

Baldwin, P.H. 1969. The Alala (Corvus tropicus) of western Hawaii Island. Elepaio 30 (5'):41-45. Banko, P.C. 1974. Report on Alala (Corvus tropicus). Unpublished report submitted to WorldWildlife Fund and Hawaii State Division of Fish and Game. Banko, W.E. Manuscript. Banko, W.E., and P.C. Banko. Manuscript. Henshaw, B.W. 1902. Birds of the Hawaiian Islands: being a compl.ete list of the birds of the Hawaiian possessions, with notes on their habits. Thrum, Honolulu. Munro, G.C. 1944. Birds of Hawaii. Tongg, Honolulu Perkins, R.C.L. 1903. Vertebrata (Aves). Pages 368-465 in D. Sharp, ed. Fauna Hawaiiensis. l(4). Cambridge Univ. Press. Vincent, J., compiler. 1966-1971. Red Data Book 2. Aves. International Council for Bird Preservation, and Survival Service Commission, International Union for Conservation of Nature and Natural Resources. Loose-leaf. Wilson, S.B., and A.H. Evans. 1893. Aves Hawaiiensis: the birds of the Sandwich Islands. Porter, London. ROLE OF FOOD DEPLETION BY FOREIGN ORGANISMS IN HISTORICAL DECLINE OF HAWAIIAN FOREST BIRDS Winston E. and Paul C. Banko U.S. Fish & Wildlife Service Mauna Loa Field Station Hawaii Volcanoes National Park Hawaii 96718

One of my primary assignments during the past 10 years has been the collection of historical literature on Hawaiian birds, the goal being a thorough survey. Over 2,000 notes and articles in The Ele aio (journal of Hawaii Audubon Society) and 800-+ sources o information published elsewhere were assembled, systematically catalogued, annotated and coded for retrieving various kinds of information on data analysis cards. As time and other duties have permitted, further effort was devoted to assimilating the vast amount of collected information and organizing it into an analytical format. As presently envisioned, this broad historical documentation and synthesis task will appear in several parts under the overall title, "Decline and Extinction of the Endemic Ha~aiianAvifauna." The status and distribution of each of the 70 endemic forms will be treated in species accounts. The causes of depopulation--competition,predation, disease, and habitat alteration--will be addressed at the ecosystem level in several major "role" papers. Work on the initial "role" paper dealing with competition as a factor of Hawaiian bird decline started in collaboration with Paul Banko in 1974 and was completed last year. Entitled "Introduction and Part I. The Role of Food Depletion and Competition by Foreign Organisms," and derived entirely from the literature, this work details the diets of the endemic avifauna and correlates the histories of alien competitors, many non-avian, which possess overlapping food requirements. In pursuing this line of investigation we discovered that a great variety of dissimilar organisms appear to have significantly depleted food supplies of many endemic birds. Our interpretation of this finding is new. A greatly abbreviated version of our analyses and tentatrve conclusions follows. It is widely recognized that birds which have evolved on oceanic islands are much more vulnerable to extinction than continental species. Of 52 endemic Hawaiian forest birds, 19 (36%) are known or believed to have become extinct within the last 150 years. An additional 17 forms (33%) are considered endangered--more than in any other biological province on earth. The conservation status of Hawaiian forest birds is shown below.

Endemic Hawaiian Forest Birds and their Current Status Family Species or Races Extinct Endangered Corvidae 1 0 1 Turdidae 6 2 3 Muscicapidae 3 0 0 Meliphagidae 5 4 1 - - - TOTALS : 52 19 17

Habitat destruction, predation and disease have often been suggested as important factors in the decline of Hawaiian birds. The theory that superior competition for primary food resources by a wide array of foreign organisms has been a significant cause of depopulation was developed from a comprehensive analysis of over 200 sources of published and unpublished records of the Hawaiian fauna. Principal references utilized were: R.C.L. Perkins (1903, 1913), E.C. Zimnerman (1948-1970), P.H. Baldwin (1953), and O.H. Swezey (1954). The terrestrial biota of the Hawaiian Archipelago evolved from comparatively few ancestral stocks which arrived over a long period of time. Over 97% of the native flowering plants, land mollusks, insects, and non-migratory land birds are endemic to Hawai'i and unusual adaptations are characteristic of many taxa. With a limited and disharmonic terrestrial resource base available for exploitation, and in the absence of whole groups of ecologically influential plants and animals found on continents, many resident birds became highly specialized in their food habits. Exemplifying this dietary specialization is the great range of highly divergent and extraordinarily adapted bill types of the Drepanididae. Of the 52 forest birds, 31 rely on insects and spiders as primary foods, 12 feed principally on fruits and seeds and the remaining nine subsist largely on nectar. However, species which forage mainly on seed, fruit, or nectar usually i-nclude arthropods in their diets--many predominately --and feed insects and spiders to their young. Therefore, birds with bills highly modified for taking only a select few examples of any particular kind of food, especially if it is seed, fruit or nectar, require an abundant supply of easily obtainable arthropods suitable to the critical demands of growing young. Caterpillars and other large insect larvae are the foods most often recorded in diets of nestling and fledgling forest birds in Hawaii. Lepidoptera larvae also comprises the most prominent element of animal food in the diets of adult Hawaiian forest birds. One genus of geometrid moths, the Scotorythra, is of great importance to the adult and young of most bird species. Coleoptera are also widely and intensively utilized, particularly by species of Hemignathus and Pseudonestor (Drepanididae) which are especially adapted for securing species of certain genera. A few insects in Orthoptera, Corrodentia, Homoptera, Neuroptera and Hymenoptera constitute major proportions in the diets of a few birds while various groups of spiders are extensively utilized by many species. Competition for food becomes a significant regulatory mechanism when the availability of a principal food shared by two or more species is depleted to the point that the population size of the least efficient species is reduced. Since the European discovery of the Hawaiian Islands in 1778, the rate of establishment of foreign organisms, many of which are real or potential food competitors of native birds, has increased by at least five orders of magnitude (lo5 times) over the natural rate for insects, birds and mammals. Perhaps in no other area of the world are so many alien organisms a part of the neo-ecology as in the Hawaiian Islands. Foreign organisms which developed in highly competitive and complex continental ecosystems have considerable potential to reduce food supplies of organisms which evolved in insular ecosystems where the resource base and number of consumers are limited. The following tentative assessments of significance of competing foreign organisms in reducing the endemic Hawaiian avifauna are based on what is known from the literature of apparent dietary overlap and reduction of principal foods.

1. The carnivorous foreign ant PheidoZe rnegacephaZa, probably arriving by 1825 and in the absence of any native ant, is judged to have played an indirect though paramount role in the reduction and extinction of all insectivorous birds. Pheidole was particularly effective in displacing an exceptionally wide variety of insects, even causing species extinctions in Coleoptera and Lepidoptera, the two orders of insects which are most important as food to a majority of the insectivorous birds of Che Hawaiian Islands. PheidoZe is judged to have been most influential in habitats under about 600 meters (2,000 feet) elevation which receive between 100 to 250 centimeters (40-100 inches) of rainfall per year. Birds resident in areas outside this range of elevation and rainfall are judged to have been affected to a somewhat lesser degree. The ecological dominance of PheidoZe on the major forested islands apparently continues to the present day, indicating that this ant may be a potent contemporary competitor of insectivorous birds. 2. The introduced Acridotheres tristis (Common Mynah) is judged to have significantly depopulated a wide variety of endemic Hawaiian forest birds from about 1895 to 1905. Competitive contact of A. tristis with endemic Hawaiian birds is judged insignificant today. Two other introduced birds, Leiothrix Zutea (Red-billed Leiothrix) and Zosterops japonica (Japanese White-eye) are judged to have played potentially significant roles in the reduction of forest birds from about 1940 through the late 1960's and from about 1950 until about the present time, respectively. In recent years, populations of Leiothrix have declined drastically in many areas, reducing their competitive significance. Zosterops populations continue to be abundant and widespread, though they may be on the verge of declining. 3. The immigrant rat, Rattus rattus, is judged to have significantly reduced Psittirostra psittacea (Drepanididae) and possibly Corvus tropicus (Corvidae) by competitive interaction beginning about 1890. Contemporary competitive significance of this rat is unknown. 4. Beginning about 1890, two foreign parasitic flies (Chaet~~aediamonticola and Frontina ~rchip~ivora) and six other foreign parasitic wasps (Casinaria infesfa, Hyposoter exiguae, ZaZeptopygus fZavo-orbitalis, ApanteZes marginiventris, AZZorhogas paZZidiceps, and Doryctodes palbiatus) are judged to have played important roles in reducing populations of insectivorous birds. The ability of these flies and wasps to parasitize whole series of endemic moths and beetles resulted in wholesale depletion of important bird food insects. At least seven species of the geometrid moth genus, Scotorythra, are attacked by one or more members of a four-species complex of these parasites. Contemporary effect of these parasitic flies and wasps as a limiting factor on bird populations is unknown, but it is suspected of being highly significant. The combined impact of all foreign competitors on food supplies of endemic birds should not be overlooked, although it can scarcely be assessed. With at least six foreign predators and parasites attacking Scotorythra, for example, the pressure on this important forest bird food is many times that exerted by natural consumers. Competitive roles between many other foreign organisms and forest birds, including those of bees, lizards and fungi are examined with inconclusive results. The equable climate of Hawaii and comparative freedom from continental predators, parasites, diseases and competitors encourage high survival of competitor species which evolved in more severe climatic and ecological environments. Many foreign competitors have rather generalized dietary requirements and, unlike the more specialized Hawaiian birds, are able to maintain sizable populations on alternate foods when the shared resource becomes temporarily depleted. Some competitors also have greater potential than Hawaiian birds for reproduction and dispersion and may place heavier consumptive demands on the resource. Such advantages enable populations of foreign organisms to gain and maintain ecological dominance over endemic birds, many of which have evolved fragile and inflexible relationships with their food resources. Formulation of valid conclusions about the causes of avifaunal decline and extinction requires that both the disappearance of birds and the agents of depopulation agree chronologically. The sequences of population decline are better known for the islands of O'ahu and Hawai'i than for any of the other large Hawaiian islands. By the 1890's five forest birds found only on O'ahu had vanished, while only one form restricted to the island of Hawai'i could not be found. However, seven birds endemic to Hawai'i were seen for the last time during the 18901s,while another O'ahu species dropped from sight during the same period. Bird extinctions early on O'ahu are correlated with the appearance, probably by 1825, of the carnivorous ant PheidoZe megacephala. On the much larger island of Hawai'i, PheidoZe apparently became ecologically significant over expansive areas much later, effectively delaying initial disappearance of forest birds. The arrival of parasitic flies and wasps on Hawai'i in the 1890's is judged to have played the dominant role in reduction of forest birds on that island just prior to and after the turn of the century. Although population histories of most foreign competitors are sketchy, it appears that species that attained a relatively high level of competitive potential with endemic birds have all done so within about 30 years of their colonization. The fastest to ascend ecologically have been the parasitic flies and wasps (3-5 years), followed by the ant and rat (15 years) and the birds (25-30 years). The bird species most commonly observed in Hawaiian forests today are those which appear to have the most diversified food habits'. Populations of species having highly restricted, specialized diets have most obviously declined over the last 150 years, many to extinction. Many natural-appearing areas within traditional ranges of Hawaiian forest birds are now almost devoid of all birdlife. That these near-vacant habitats may not produce sufficient food suitable to birds is suggested not only by the general absence of endemic forms but by unusually low numbers and diversity of alien species. There is no panacea for conserving Hawaiian birds which are subject to severe competitive pressure from foreign organisms. A new management concept is clearly needed, one focussed on preserving the biological integrity of the natural Hawaiian ecosystem. That foreign food competitors, often in the form of grossly dissimilar organisms, may threaten the existence of unique forms of life on other oceanic islands needs to be recognized. Further research in Hawaii is required to reveal the true dimension of the competitor problem, to evaluate the present status of the principal food species and to critically examine the feeding ecologies of endemic birds which are still extant.

Literature Cited

Baldwin, P.H. 1953. Annual Cycle, Environment and Evolution of the Hawaiian Honeycreepers . Perkins, R.C.L. 1903, 1913. Fauna Hawaiiensis Swezey, O.H. 1954. Forest Entomology in Hawaii. Zimmerman, E.C. 1948-1960. Insects of Hawaii. ECOLOGICAL PARK MANAGEMENT : A MANAGER'S PERSPECTIVE Robert D. Barbee Hawaii Volcanoes National Park Hawaii 96718

The National Park System offers a special contribution to world culture. It has served as a model to be followed by other countries throughout the world. The notion of preserving natural settings for their intrinsic, non-commercial value is a relatively new concept of land use. While the precise origins of the idea for preserving natural environments is somewhat obscure, it did find legislative expression in 1864. In that year, Abraham Lincoln signed into law legislation withdrawing California's Yosemite Valley and the nearby Mariposa Grove of giant sequoias, ceding them to the State of California to be preserved as a public park. The evolution of the preservation movement frcvm that time forward is well documented and continues today on a number of governmental and private levels. It is within the National Park System, however, in which is preserved the greatest diversity of natural ecosystems on earth. As inroads and development of land continues to accelerate and modify the land, it is increasingly becoming the national parks and monuments where a baseline of relatively unmodified land is protected. This fact alone places tremendous responsibility on those whose job it is to manage these lands, and to do so in an ecologically enlightened way. There is an added difficulty in the Hawaii parks in that they are biologically complex and that they were considerably disturbed when taken into the National Park System. As we enter an era of intensive management of national park resources, it is vital to our desire for success that park managers remain in continuous dialog with the scientific community. It is equally important for the scientists working with the Park Service to understand the Service's policies relative to the management of the natural resources within the system. The balance of this paper will attempt to provide this perspective by developing the conceptual framework within which the park manager operates. Most units within the park system were set aside through the political process. Legislation establishing parks was passed because individuals or interest groups perceived a threat of exploitation or development to an area already recognized to be of an unusual or superlative nature. It was only logical that "strict protection" developed as a management philosophy for the early parks: Protection of timber from axe and saw, wildlife from hunters, minerals from miners, and rangeland from grazing livestock. Reinforcing this was the early administration of the new parks .(Yellowstone, Mount Rainier, Sequoia, Yosemite) by the United States Army. After 1916, the newly established National Park Service took over management of the Park System. By 1916 there were 27 natural area parks within the System, including then Hawaii National Park. The best management of park resources was considered no management. "Let nature take its course" became the hallmark of the National Park Service. The natural areas within the System were, and are still today, billed as great living museums of natural history and scenic beauty, with only incidental influences and modifications by man. Keeping these natural wonders essentially unimpaired seems like a reasonable aspiration, and to a large degree it has been achieved; nevertheless, %n recent years a great fallacy has been exposed. Nature was not allowed to take its natural course. Nature was "aided" by man. What was really being practiced was a form of selective protection. Along with successful efforts to protect the parks from exploitation by man himself, they were also "protected" from certain natural depredations and forces. Some resources were considered as "good" and others "bad." The good resources were protected from the bad resources. Deer were thought of as good and predators as bad; trees were all good, but anything that destroyed them was bad; fire was all bad; fish were good and most things that ate them (besides people) were bad; etc., etc. In the words of Lyle H. McDowell, a National Park Service official responsible for consolidating the Service's resource management posture in the 19601s,"protection as a management concept was steeped in emotionalism and sentiment and coated with the best of intentionsbutunfortunately it was misdirected." Viewing all this from the secure vantage point of hindsight, it can be seen that the "naturalness" of the parks was not being guided by ecologically enlightened policies. Had nature truly been allowed to take its natural course, had a11 of the processes been protected, many of our wildland parks would be in better condition today. Only part of the park environment was being protected--only the "things," and in a number of cases, on a selective basis. Some park forests illustrate this fallacy. Forests were protected, but what about the natural processes associated with them? When native forest insects were sprayed, the natural forest rhythm was inhibited. In most western forests natural wildfires were as much a part of their evolutionary development as rain and sunshine. But for years these forests have been "protected" from wildfire. There has been a biological response from the environment induced by withdrawing this natural process--one which has been tremendous and anything but desirable. By mid-twentieth century, ecologists had assembled a basic understanding of many aspects of the "land organism." With the growing sophistication of this budding science, wildland park managers could no longer remain in intellectual isolation from ecological realities. By the early 1960's it had become obvious that an incisive policy of forthright management had to be undertaken if the national parks were to remain the great bastions of naturalness that Congress intended. In 1963, a committee of eminent scientists under the chairmanship of Professor Starker Leopold, ecologist at the University of California, Berkeley, was requested by the Secretary of Interior to submit recommendations for a policy of park management that would be ecologically sound. Their primary task was to advise on wildlife management, but they found that the entire spectrum of park resource management had to be considered. For one example in their analysis, the committee probed the dilemma brought about by selective protection, and they turned their attention to those forests located within the parks of California's Sierra Nevada. In their report they minced no words: "Today much of the west slope is a dog-hair thicket of young pines, white fir, incense- cedar, and mature brush--a direct function of over-protection from natural ground fires. Within the four national parks--Lassen, Yosemite, Sequoia and Kings Canyon--the thickets are even more impenetrable than elsewhere. Not only is this accumulation of fuel dangerous to the giant sequoias and other mature trees but the animal life is meager, wildflowers are sparse, and to some at least the vegetative tangle is depressing, not uplifting." Dr. Leopold and his committee articulated this kind of inconsistency and others as well. They recommended a strong course of action which has formed the nucleus for a new direction in the management of the wildland parks. "As a preliminary goal, we would recommend that the biotic associations within each park be maintained, or where necessary recreated, as nearly as possible in the condition that prevailed when the area was first visited by the white man. A national park should represent a vignette of primitive America." I I ...Restoring the primitive scene is not easily done nor can it be done completely. Yet, if the goal cannot be fully achieved it can be approached. A reasonable illusion of primitive America could be recreated, using the utmost skill, judgment, and ecologic sensitivity. This in our opinion should be the objective of every national park and monument." Contrary to an interpretation by some managers, this does not mean turning back the "ecological clock" to some time in the past and then attempting to stop it. It does mean that park managers must view the total park resource mosaic ecologically. It means finding out what aspect of the ecosystem needs to be rectified (through research) and then doing something about it (through management). Relatively uninfluenced portions of the parks must be closely guarded and maintained in as pristine a state as possible. To build a road, drill a well, or graze a meadow may accomplish one purpose; however, its effect on the naturalness of the park must also be considered. If an improvement has to be made, its disruptive influence must be minimized. The Hawaiian parks offer unique challenges. Modifications to the ecosystem have been extensive. When the Leopold Committee model is applied, there are philosophical problems to overcome regarding the role of Polynesian man. His impact was tremendous, yet the National Park Service view accommodates Polynesian introductions as "natural" and those coming af terwards as "not natural." Restoring and maintaining the pristine scene in the Hawaiian parks will demand increasing sophistication on behalf of the Service's scientific collaborators as well as its own resource managers. Above all, understanding of the long-range objectives regarding the perpetuation of the natural resource mosaic is an absolute necessity. Within Hawaii Volcanoes and Waleakala National Parks are found some of the last vestiges of primitive Polynesia. Their preservation has become, in a sense, a sacred charge--one that can only be successfully fulfilled through ecologically guided management. ECOLOGY OF RARE DROSOPHILA SPECIES IN HAWAII VOLCAIIOES NATIONAL PARK Hampton L. Carson Department of Genetics University of Hawaii at Manoa Honolulu, Hawaii 96822 The Drosophilidae of ~awai'ihave recently been estimated to number nearly 800 endemic species (Carson and Kaneshiro 1976). Roughly 500 of these belong to the genus DrosophiZa. Among them are somewhat more than a hundred unusually large and colorful closely-related forms called the "picture-winged species." At about the time that D. E. Hardy's classic monograph on Hawaiian drosophilids was published (1965) a broad interdisciplinary project to study the evolutionary biology of these flies was launched. At that time the picture-winged species were chosen for special study, primarily because they could be reared in the laboratory, permitting study by genetic techniques. They proved to be phylogenetically very close, to have very favorable giant polytene and metaphase chromosomes and to be amenable to electrophoretic study of soluable proteins. Maintenance of cultures in the laboratory meant that hybridization techniques could be applied to the species problem; furthermore, some of the very complex territorial and mating behaviors could also be studied in the laboratory. Progress in this work has been reviewed by Carson, et a1 (1970) and Carson and Kaneshiro (1976). In the early years of the project, we concentrated on basic systematics (supplemented by genetic data), geographical distribution and phylogeny. Somewhat later, Heed (1968, 1971) and Montgomery (1975) published accounts of the ecology of the flies. For the picture-winged species, at least, more is now known about their natural breeding sites than for any other comparable group of drosophilids anywhere in the world. This has meant that these picture-winged species are prime objects for the study of ecological genetics, a science which attempts to relate the genetic variability carried in a population to ecological parameters. This phase of the work has been concentrated on the Big Island because of the dynamic nature of the biological situation there. Thus, not only is the island itself new, but both the species and the ecosystems are likewise new. Using genetic techniques, it now appears that there are about 27 species of picture-winged DrosophiZa on the Big Island. All are endemic to this island and appear to have arisen from founders arriving from islands to the north and west. Certain chromosome and electrophoretic markers. have been used to trace these past events (Carson 1970, 1976). As a prelude to more intensive studies of ecological genetics, the collectors in our group have been seeking out the breeding sites of as many of these species as possible, adding to the already extensive data of Montgomery (1975). From the very beginning of the project, certain species of picture-wings appeared to be rare and local in their distribution. The reasons for this were not at first apparent. I now believe, however, that the evidence indicates that they are rare and local because their specific plant hosts are also rare and local; in fact, in some cases endangered. Eighteen (two-thirds) of the twenty-seven species of picture-winged DrosophiZa known from the island of Hawai'i have been captured within the borders of Wawai'i Volcanoes National Park. These are listed in Tables 1 and 2. Half of these species have been reared from substrates collected in the Park (Table 1) whereas the other half have not so been reared (Table 2). This latter table lists known host plants from other areas; these may provide a clue to where to look for breeding sites. Nine Big Island species have never been taken in the Park despite extensive efforts over a number of years. Most species of DrosophiZa on all the islands are rare and local in their distributions. In most cases, specimens cannot be located, even using various attrahents at any distance from the host plants mentioned. The rarity of the host plants seems to correlate with the rarity of the species which breed on them. A striking example is afforded-by the-Big Island species DrosophiZa engyochracea of Kipuka-ki and Kipuka-pua-ulu of Hawai'i Volcanoes National Park. This species is found only near large Sapindus saponaria trees. Its oviposition site is quite restricted in that it appears to lay eggs only on the bark of very large Limbs or fallen trunks of this tree. What is more, the fallen branch must at first be green, as in the case of limbs broken off by wind. As the bark decays, the flies will oviposit only at the interface between the green bark and the point of fresh decay as this appears to provide the best, if not the only, micro-niche for growth of the larvae. The specificity of the engyochracea-Sapindus association appears to relate also to the size of the trees. Thus, in Kipuka-pua-ulu the flies become scarce under small-sized trees; possibly when small branches fall they desiccate rather than decay. Attempts to locate either D. engyochracea or Sapindus saponaria elsewhere on the island of Hawai'i have met with little success. A few specimens of the latter were taken from a crater on the northwest rift of HualBlai. Sapindus was not seen. A few specimens of Sapindus (but not the fly) occur along the Ka-'o-iki Pali on ranch land along the southwest 'boundary of the Park. We are left with the tentative conclusion that both tree and fly are virtually confined to Kgpuka-pua-ula and ~Tpuka-kcwithin Hawaii Volcanoes National Park. Even flies utilizing fairly widespread substrates (e.g. Cheirodendron) are not always found wherever the host plant is found. Indeed, there are large areas of 'Cihi'a Cibotium- Cheirodendron forest, especially on newer lava flows, which appear to have very few picture-winged DrosophiZa in them. More data are needed, but the situation here seems to be that most picture-winged DrosophiZa species form stable permanent populations in an area only after the trees have attained a quite large size. On pahoehoe flows, for example, such conditions develop very slowly. A second prominent feature is the patchy distribution of some of these rare trees. I do not refer here to locally en- riched areas such as the kipukas, where soil conditions may permit a rich and diverse forest to develop. Rather, the patchiness of DrosophiZa distribution appears to be related to the chance and patchy nature of the colonization of certain of their host plants.

Good examples are afforded by Charpentiera and Pisonia. As both of these trees are important breeding sites for DrosophiZa, our group has spent a large amount of time seeking out foci of these trees within the Park and elsewhere. Al- though transect information is not available, only three individual large trees of Charpentiera have been located in the '61a'a Tract and one (not including planted specimens) in Kcpuka-pua-ulu.

A few locally dense stands of Pisonia occur in Kipuka- pua-ulu and Kepuka-kc, whereas only a single such clump has been located in the 'bla'a Tract (found by F. R. Warshauer)

CZermontia hawaiiensis is another extraordinarily important host tree. It has recently been listed by the Fish and Wild- life Service on the proposed endangered list. We have worked extensively along the Saddle Road and at Kahuku Ranch where this tree grows. The species is exceedingly rare within the Park and, apparently for this reason, certain species which use it as a primary host are infrequent in the Park. I do not know of a single specimen in 'Fla'a Tract or in ~ipuka-kiand Kzpuka-pua-ulu. A single specimen was found near Thurston Lava Tube and a considerable number are present in the pit within Pu'u-huluhulu. Despite recent slides and volcanic activity near the latter, these specimens are thriving. Over the next five years our group will be seeking out and studying such foci. I would greatly appreciate hearing from naturalists, rangers or anyone who has knowledge of the whereabouts of any patches of rare plants of the kind described above. These preliminary facts and discussion lead to several conclusions. First, these patchy rarities need to be watched carefully in order to see that they do not die out unnoticed. Their loss would not only be deplorable as an erosion of the Park flora but would result in loss of associated fauna as documented here. Chromosomal and electrophoretic data on the DrosophiZa species have indicated that detailed evolutionary patterns can be revealed (Carson 1970, 1976). The genes to some extent seem to work as a clock-like mechanism, recording past history. Both they and their hosts must be preserved if future scientists will be able to capitalize on the start that we in this generation are making. Management of these rare and patchy parts of the ecosystem may require some special procedures. I subscribe to the notion that the native ecosystems should be interfered with as little as possible. These patches must be monitored and should they show decline to the point of disappearance I feel that human intervention in the form of artificial propagation --in situ is justified. An example is the planting, done some years ago, of a modest number of Charpentiera in ~Tpuka-pua-ulu. These trees have now reached the size that small communities of associated insects are present. Another case I would like to mention is Cheirodendron in ~Epuka-pua-ulu. As soon as Cheirodendron was proved to be an important host tree for DrosophiZa, a thorough search for this tree in the ~Epukawas made. We concluded that the only specimen within the ~i~ukawas the very large tree by the trail just 'beyond the spur leading to the giant Koa tree. From 1963 until last year this single tree represented an extraordinary "island within an island." Quite a number of special Cheirodendron-related flies were known only from under that single tree. This tree is now dead. I feel that there is little question that this tree was an original, natural specimen occurring as an outpost in this relatively xeric kipuka. I favor planting several young Cheirodendron in this location within the Klpuka using stock brought from the closest natural area to this site. In the coming years, our group proposes to gather data on the ecology and community characteristics of these rare patches in both Hawaii Volcanoes and Haleakala National Parks. These studies may serve to document some of the special attributes of the delicate island ecosystems and suggest how they may be managed. Public education about these ecosystems, which lack the spectacular large animals and great uniform forests of some of the mainland National Parks, represent a special challenge to administrators of the ~awai'iParks. Table 1. Breeding records for picture-winged Drosophi Za from Hawaii Volcanoes National Park (from Montgomery 1975).

Host Plant Species Locality (breeding site) digressa 'bla'a Tract Charpentieray: heteroneura I Ola'a- Tract Ckeirodendron murphyi I Ola'a- Tract Cheirodendron paucipuncta 'bla'a Tract Cheirodendron setosimentum 'ola'a Tract Cibotium; Cheirodendron

engyochracea KEpuka-k? Sapindus saponaria* kawaiiensis Kzpuka-k? Acacia koa Kzpuka-pua-ulu Osrnanthus* ~Fpuka-pua-ulu Myrsine Poli-o-Keawe Pali Reyno Zdsiay:

rnacrothrix KEpuka-pua-ulu Pisonia* ciZiaticrus Poli-o-Keawe Pali ReynoZdsia"

Jf Rare in Park Table 2. Species of picture-winged DrosophiZa captured within Hawaii Volcanoes National Park but for which breeding sites within the Park are unknown (from Montgomery 1975). Breeding Site outside of Park Species Park Locality (host plant) ochracea '61a'a Tract Freycinetia pro ZaticiZia 'bla'aTract Freycinetia setosifrons '61afa Tract ~etrapZasandra;* Cheirodendron sproati 'bla'a Tract Cheirodendron siZvestris '6la1a Tract Ctermontia hawaiiensist ~arattia;*Cyanaea* silvarentis KZpuka-kE Myoporum;* Osmanthus* Acacia koa puZZipes KZpuka-k: Wikstroemia* basisetae 'bla'a Tract Unknown czaytonae 'bla'a Tract Unknown * Rare in Park Endangered species

Table 3. Big Island species of picture-winged Drosophita unknown in Hawaii Volcanoes National Park (from Montgomery 1975).

*ties Host Plant (breeding site) assita Urera* heedi My oporum* ochrobasis Myrsine; Ctermontia; * Marattia* pisonia Pisonia; * Charpentiera* n. sp. near sobrina Te trap Zasandra" aZsophiZa unknown conspicua unknown forme t Za unknown psi ZotarsaZis unknown

* Rare in Park Literature Cited

Carson, H. L. 1970. Chromosome tracers of the origin of species. Science 168: 1414-1418. Carson, H. L. 1976. Inference of the time of origin of some Drosophila species. Nature 259: 395-396. Carson, H. L., Hardy, D. E., Spieth, H. T. and Stone, W. S. 1970. The Evolutionary Biology of the Hawaiian Droso- philidae. In: Essays in Evolution and Genetics in Honor or Theodosius Dobzhansky. Eds. M. K. Hecht and W. C. Steere. New York: Appleton-Century-Crofts, pp. 437-543. Carson, H. L. and Raneshiro, K. Y. 1976. Drosophila of Hawaii: Systematics and Ecological Genetics. Ann. Rev. Ecol. System 7: (in press). Hardy, D. E. 1965. Insects of Hawaii, Vol. 12, Diptera: Family Drosophilidae, Honolulu: Univ. of Hawaii Press, pp. 814. Heed, W. B. 1968. Ecology of the Hawaiian Drosophilidae. Univ. Texas Publ. 6818: 387-419. Heed W. B. 1971. Host plant specificity and speciation in Hawaiian Drosophila. Taxon 20: 115-121. Montgomery, S. L. 1975. Comparative Breeding Site Ecology and the Adaptive Radiation of Picture-winged Drosophila (Diptera: Drosophilidae) in Hawaii. Proc. Hawaiian Ent. Soc. 22: 65-103.

Acknowledgements

This work was supported by NSF Grants DEB-74-22532 and GB-29288 and by the Cooperative National Park Resources Unit, University of Hawaii. THE PHYSIOGRAPHY AND MARINE FAUNA OF INSHORE AND INTERTIDAL AREAS IN THE PU'U-KOHOLA HEIAU NATIONAL HISTORIC SITE Daniel P. Cheney, Hilo, Hawaii Donald E. Hemmes, Hilo College, Hilo, Hawaii Ronald S. Nolan, O.R.C.A., Kawaihae, Hawaii

This report describes the biotopes and marine fauna of the reefs adjacent to Pu'u-koholii Heiau National Historic Site and offers recommendations for the maintenance of the biota and shoreline development. It is a summary of a longer report prepared for the Cooperative National Park Resources Studies Unit (Contract No. CX 8000 6 0031). The field work was conducted from April through June, 1976 with the assistance of Mr. James MacNeil and Mr. Loren Akaka. Study Site Description The site for the marine survey was the area bounded by Pu'u-kohola, Ka-wai-hae Breakwater and Spencer Beach Park. It extended 150111seaward from the southern boundary of the Historic Site and formed a rough rectangle 150m by 300m, enclosing approximately 45,000m.2 Survey Methods A. Fish. S.C.U.B.A. was used to survey the fish fauna ofhe site. The species list was comprised from six hours of underwater observations throughout the biotope types of the bay. Silt and rubble areas had very limited populations of fishes, therefore no quantitative assessments were made in those regions. At the southern limits of the park boundary there is a rich region of living coral in the form of patch reefs separated by sand plains. In this area two 50m transect lines were installed at a depth of 4 to 5m. An inner line was set 50m out from shore with an east-west orientation. An outer line lay seaward of, and adjacent to, the first line and ran northwest to southeast. Three replicate censuses were performed on each transect (2 April, 1000 hours; 10 April, 1000 hours; and 10 April, 1400 hours). Fish were counted 5m on each side of the line and data recorded in 5m2 quadrats on underwater paper. In addition to offshore observations, intertidal, brackish water and pond fishes were captured with small hand nets and mmi spears for identification. B. Benthic Invertebrates. Quadrats of 100m2 were sampled qualitatively within representative habitats for benthic epi-and infauna. Most of the information concerning molluscs was obtained by a random sampling of patch reefs and sand deposits within the offshore portion of the site and by a detailed search along the intertidal zone. Additional identification of macroinvertebrates was made from samples of living and dead coral heads, coralline and fleshy algae and basalt or limestone rubble. The infauna in sand and mud wgs collected by removing surface samples of 0.5 to lm to a depth of 2 to lOcm with a bucket dredge and filtering the sample through a 2mm square mesh larval net. The line transects employed for the fish censuses were used to ennumerate corals and other benthic invertebrates in the coral-rich fringe zone of the site. The biota and substrate type were recorded in one meter increments along the line and samples were taken at points where an immediate identification could not be made. Discussion The waters adjacent to the Pu'u-kohola Heiau National His- toric Site contain a fauna which is nowhere as diverse as that found in similar areas along the coast of the Island of Hawai'i (Brock and Brock, 1974; Kimmerer and Durbin, 1975). Neverthe- less the area is an important resource and supports a significant population of juvenile fish and a large benthic biomass. Physical Characteristics The dominant factors affecting the distribution of organisms within the study site are substrate and water quality characteristics. Wave activity is, except during periods of strong southwesterly swells, limited to localized wind-driven waves less than one-half meter high. It is usually minimal in the morning and reaches a peak in mid-afternoon. Turbidity generally correlates with increased onshore wind speed and wave height. It is greatest near the inner portions of the site and the waters on the outer margin of the survey site also have a relatively high turbidity when compared with offshore waters. Visiblity is greatest in the morning but never greater than 3 to 10 meters anywhere within the site. Nearshore, the water is extremely turbid (visibility less than lm) and any activity results in a fine opaque suspension. Salinity and temperature are influenced by warm (to 29.5"C) subsurface freshwater springs, intermittent surface runoff, prevailing exchange rates and incident solar radiation. The freshwater springs are most noticeable in the rocky intertidal areas between Spencer Beach Park and the Breakwater. These springs produce a distinct surface layer of warm water which overlies the cooler subsurface water over most of the site. In shallow waters the zone of freshwater influence extends through the water column and results in a more-or-less estuarine environment in these areas. The outer edge of the site has salinities and temperatures approximating oceanic seawater. --The Fish Fauna Although the Ka-wai-hae reef system must have at one time had one of the best developed reefs in the Hawaiian Islands, today its fish fauna is depauperate due perhaps to a combination of environmental disturbances associated with harbor construction and unrestricted resource exploitation at the Pu'u-kohol3 site. Dredge tailings have enveloped much of the reef and siltation from these tailings has spread throughout the survey area. The occurrence of reef fishes is highly dependent upon proper habitat in the form of shelter and feeding sites (Nolan, 1975) and at Pu'u-kohola, the normal habitat has been drastically altered. As expected, the coral- rich area had more individuals and a higher diversity than any other biotope in the site. The reef ecosystem only a few hundred meters seaward of the Park site near the proposed site of a small boat harbor is not so extensively altered. The richness of this reef as documented in pre-and post-detonation surveys during Project Tugboat by the State Division of Fish and Game (Day, -et al., 1972) contrasts markedly with that of the park site. Fisrand Game biologists observed or collected 111 different species of fishes in thirty-four familes while only about half as many species were seen in the park site (63 total-- two of which occurred in brackish pools and five were restricted to tidepools). Common species observed during Project Tug- boat were ThaZassoma duperreyi, Chromis ovalis, and MuZZoidichthys samoensis. These were also among the most common individual species found on the Park transects. The most abundant fish seen during Project Tugboat, Scarus dubius, had a low level of abundance during the Pu'u-koholg survey. Sharks are characteristic inhabitants of the Park site. Grey reef sharks, Carcharhinus menisorrah, black tip sharks, C. meZanopterus, and white tip reef sharks, Triaenodon obesus, were seen at most high tides during calm, still air conditions Feeding bouts on mullet have been observed and it may be that the sharks have a predatory advantage over prey in the murky, turbid water. Local fishermen report that the sharks enter the bay to mate and give birth to young. Prior to the survey a female grey reef shark was taken and found to be bearing fetuses. An additional attractant may be the elevated water temperature and/or reduced salinity of the inner part of the survey site. -The Invertebrates--Excluding Mollusks The distribution of invertebrate species and particularly corals and crustaceans is distinct from thac of fishes and is indicative of the prevailing conditions of the area. Living corals are absent in silt laden rubble areas or limited to only a few species where silt and fresh water intrusion predominate. Many coral colonies are in what appears to be a stress-state and exhibit extensive dead zones, usually in basal areas, or reduced pigmentation. A considerable fraction of the coral-based substrate, i.e. not a sand or silt bottom, is made up of dead and moribund colonies. Crustaceans are especially numerous in and under dead coral heads and beneath algal crusts and zooanthid mats in the intertidal zone. These habitats support a diverse population of xanthid crabs and alpheid shrimp and probably represent optimal environments for these organisms. Eryozoans are common throughout the survey site. Turbidity and sedimentation do not appear to be a problem of survival of these animals. Most are found attached to the undersurfaces of rocks and coral colonies where they would not be affected by sediments. Polychaete annelids, which were not identified to the genus level, were an important group of invertebrates in all habitats except those areas with a high percentage of live coral or bare pavement. The extensive porous coral rubble combined with minimal water movement offer favorable conditions for the development of large populations of serpulids in intertidal regions, sabellariids, in the silted rubble, and sabellids. The polychaetes are a major infauna in living and dead coral heads and occupy small habitats in association with sponges, bryozoans, crustaceans, gastropods and echinoderms. -The Mollusks There appears to be a remarkably smaller population of macro-mollusks in the highly silted portion of the Historic Site than in the clearer waters further offshore and to the north and south of the site (personal observation). Cowrie species such as Cyprea caputserpentis, Linn., C. isabeZZa Linn., and C. heZvola Linn. were not seen after extensive searches under rocks and coral rubble in the area. Early morning searches of sand pockets turned up some Conus quercinus Solander, C. pulicarius Hwass, and Terebra crenuZata Linn. on the Spencer Beach side of the Historic Site, but nothing comparable to the wide variety of species of sand dwellers which can be observed in sand pockets several hundred meters seaward of the Site at approx. 5-10 meters depth (e.g. T. macutata Linn., T. guttata Roeding, r. strigiZata Chem., T. areoZata Link, T. affinis Gray, HastuZa Zanceata Linn., and H. Zauta Pease) . The intertidal zone was dominated by Nerita picea Recluz and Littorina scabra Linn., the latter being found in abundance on Kiawe (Prosopsis paZZida HBK) branches hanging in the ocean. Only small populations of crabbed mollusks were observed along the coastline from Spencer Beach Park to the beach near Make-Bhua Gulch during the night or day. The beach and the shoreline along the artificial breakwater bordering the northern end of the marine portion of the Historic Site were not surveyed for crabbed specimens because of the large number of empty shells available in the breakwater rubble which could become crabbed and not be representative of the living mollusks of this region. Man-Related- Site Features A. Sharks, Hale-o-Kapuni and toxic organisms. Sharks are common in the site, nevertheless, local fishermen report that they have never been molested by these large carnivores and they are rarely seen swimming inside the bathing area of Spencer Beach Park (which should attract the attention of the swimmers!). Be- cause the presence of these animals reinforces the historical significance of the shark heiau, Hale-o- kapuni, they should be protected, or at least not disturbed. With the exception of the concentration of sharks within the apparent site of Hale-o-Kapuni, positive evidence for the existence of the structure is lacking. The area is presently heavily silted and no rock out- croppings or artificial structures are visible. The site is identified (not clearly) on Jackson's map of 1870 (Soehren, 1964) where it appears to rest about 30m offshore in line with Pu'u-koholH and Maile-kini. Soehren (1964) writes of an informant speaking of a channel leading into a structure ("heiau") where bodies were placed for sharks. Whether the bodies were human or animal was not stated. A recent archaeological survey of the site (Carol Link, personal communication) failed to reveal the presence of any subsurface structure. Thus the true significance of Hale-o-Kapuni remains unclear and it cannot be stated if it was an actual heiau or merely a site of concentrated shark activity. Except for the sharks, few other toxic or dangerous organisms are present at Pu'u-kohola. The toxic echinoids, Echinothrix caZamaris and E. diadema, otherwise known as wana, are very common in shallow waters facing the rocky natural shoreline and could be a hazard to swimrncrs or waders in the area. A few small (1 to 3 foot) moray eels are present inshore among the coral heads and boulders. B. Human Waste. An abundance of litter in the form of momboats, engine blocks, tires, bottles, cans and wire may be seen on the bottom and along the shoreline. Much could be done to improve the aesthetic qualities of the site both in the water and on the land by removing the litter and educating the public as to the nature of the park in an effort to maintain a relatively natural and litter-free environment. The Ka-wai-hae breakwall and fill is a major visual obtrusion but cannot be easily modified. Vegetation cornpatable with the park site should be placed along the breakwall facing the park to improve the visual setting, at least from sea level. An additional benefit to the plantings may be reduction in sand and silt erosion into the water from the fill area. C. Recreational Impact. Public use of the park site can be expected to increase once access and aesthetic improvements are made. Presently gill nets are set in the northern portions of the survey site and hook and line fishermen sometimes work through the kiawe trees to fish from the shoreline. The fish species taken most frequently by these fishermen are pZpio (young of several species of ulua or jacks), mullet and moana (~ara~eneusmultifaciatus). Spearfishermen are common but their catch rates appear to be very low comparedwith the catch in areas north or south of the site. Greater use of the intertidal zone is expected with increased public awareness of the park and heavier use of the neighboring county park. The organisms in this zone and subtidal areas are relatively resistant to non-consumptive use and should be protected to the limits of the survey site although some fishing could be permitted D. Present and Future Develo ment. The greatest impact on the wzraualitv--YP--- an marine biota of Pu'u-koholg will come from the construction and maintenance of the proposed Ka-wai-hae small boat harbor.and industrial developments in and around Ka-wai-hae harbor. For example, turbidity resulting from dredging operations can extend into adjacent areas and possibly affect the type and rate of colonization of marine organisms. Secondary impacts of these and park developments will be the release of pollutants from boats, industrial sources and cesspools. The influence of future development on the Pu'u-kohols marine fauna will probably be overshadowed by the ongoing impact of the Ka-wai-hae breakwater. This structure undoubtedly effectively blocks longshore currents and facilitates the deposition and resuspension of detrital and organic sediments. Unless the nearshore circulation can be reestablished, the marine fauna of Pu'u-kohols will continue to undergo change to silt tolerant species. Literature Cited

Brock, J. H. and R. E. Brock. 1974. The marine fauna of the coast of northern Kona, Hawaii. UNIHI-SEAGRANT-AR-74-02. 30~~.

Day, W. C., --et al. 1972. Project Tugboat: Explosive excavation of a harbor in coral. U. S. Army Engineer Waterways Experiment Station, Livermore, California, Explosive Excavation Research Lab (390123).

Kimmerer, W. J. and W. W. Durbin, Jr. 1975. The potential for additional marine conservation districts on Oahu and Hawaii. Sea Grant Technical Report, UNIHI-SEAGRANT- TR-76-03.

Nolan, R. S. 1975. Ecology of patch reef fishes. Ph.D. Dissertation, Univ. Calif., San Diego, Scripps Institution of Oceanography. Fishery Bulletin, In Press.

Soehren, L. J. 1964. An archaeological survey of the shores of Ouli and Kawaihae, South Kohala, Hawaii. Dept. of Anthro., B. P. Bishop Museum, Honolulu, Hawaii. SUMMARY OF RESULTS OF PRELIMINARY INVES'TIGATIONS AT THE HILINA PAL1 PETROGLYPH CAVE AND ASSOCIATED SITES" Paul L. Cleghorn Department of Anthropology Bernice P. Bishop Museum Honolulu, Hawaii 96818

Introduction Archaeological investigations at the Hilina Pali Petroglyph Cave and associated sites were conducted by members of the Department of Anthropology, B. P. Bishop Museum, under contract with the National Park Service in March 1975. The primary goal of this work was to record the Hilina Pali Petroglyph Cave, a newly discovered site. It appeared that this site might provide the unique opportunity of placing Hawaiian petroglyphs in direct stratigraphic context, in that a number of khe petroglyphs extended below the floor deposit. One of the objectives, then, was to conduct test excavations to uncover buried petroglyphs and to recover datable material with which to assess their age. In addition to the work at the main petroglyph cave, the immediate area was to be surveyed and any sites encountered were to be recorded. Location and Setting The survey area is located within the Hawaii Volcanoes National Park, in the Ka'G District, on the island of Hawaii. Specifically, it lies on the plain below Hilina Pali, approximately 3 km from the coast. The main petroglyph cave, centrally located in the cluster of recorded sites, is at the 1,000-ft. elevation on a gently sloping, grassy plain. Rainfall for this region is less than 40 inches per year (Carlquist 1970:77). Although there are no specific data for the Ka'E Desert, Carlquist thinks that this area is probably drier than the rainfall figure suggests, due to excessive evaporation and low humidity (1970:79). Another factor affecting the availability of moisture is the seasonality of rainfall. Kona (southerly) storms, occurring primarily in the winter months, account for most of the rainfall in this region. They can be heavy over a short period of time, or fail to appear for one or more entire winter seasons.

*"Phase I Archaeological Survey of the Hilina Pali Petroglyph Cave and Associated Sites, Hawaii Volcanoes National Park" (Ms. 051576) is on file at the Dept. of Anthropology, B. P. Bishop Museum, Honolulu, Hawaii. As a result of this aridity, the area's vegetation comprises xerophytic scrub and forest, giving way to open, dry grasslands and barren, recent, lava fields. The grasslands are made primarily of the indigenous pili grass (Heteropogon contortus [L.] Beauv.) and various Andropogon grass species, with scattered '6hi'a-lehua trees (Metrosideros coZZina [Forst.] Gray). It is probable that the introduction of goats into the area in the early 1800s resulted in decima- tion of certain native plant species. The endemic and indigenous flora of the area may thus have been somewhat more diverse and abundant prior to European contact. The Hilina Pali Petroelv~hCave (Site 50-HV-383) Description The Hilina Pali Petroglyph Cave consists of a N-S-extending lava tube with a collapsed section of roof permitting access into both the N and S portions of the tube. The S section extends for 145 meters from the collapse and shows little evidence of use and few petroglyphs. The N section contains abundant petroglyphs for about 17 meters from the collapse. The collapsed area is roughly circular, c. 9.6 meters in diameter, with a maximum depth of 2.7 meters. Located within the collapsed area is an irregularly shaped platform, that is roughly paved and contains a square, stone-lined fireplace (FP-2), constructed by placing five stones on edge. Just off the N edge of the platform is another fireplace (FP-3), marked by a circle of five stones. A flaked basalt cobble was found on the surface within the circle of stones. On the E side of the collapsed area a 1-meter-wide rampway leads from the plain into the N section of the tube. This rampway is partially bounded on the E by a retaining wall, and on the W by a free-standing, multiple-stacked wall. Across the 7.8-meter opening of the N extent of the tube is a free-standing, multiple-stacked wall that rests upon a pavement of small, angular, basalt cobbles. The pavement extends approximately 9.0-meters into the tube; its exact extent could not be determined because of extensive roof-fall. A probable fireplace (FP-I), consisting of four stones placed in a square, and a fifth in the center, is located within the tube atop this pavement. Petroglyphs The N section of the tube abounds in petroglyphs. The majority of the petroglyphs are relatively near to the opening, although a few occur as far as 17 meters into the tube, which is the approximate limit of daylight. Approximately 550 were recorded in detail, and the number still to be recorded on the ledges to the E and W of the opening may be even greater. The tube has numerous natural ledges running along the sides for almost its entire length. These ledges are shiny, smooth, and almost glazed. These smooth surfaces are thin and cover a relatively porous and soft basalt. The production of petroglyphs in this surface must have been relatively easy and the glyphs exhibit a delicate, or fine-lined, light-handed style that is unusual for Hawaiian petroglyphs. This fine technique made smaller figures possible. Many of the same forms that are found on the surface sites in the vicinity are reproduced in miniature here. The largest petroglyph recorded in the cave was 52 cm long. A few others are over 40 cm in length, but the majority are less than 30 cm, and many are less than 10 cm. Even the smallest glyphs are executed in surprising detail--toes, fingers, and hair or headdresses are common. The most common petroglyphs recorded are non-human forms, the majority of which are dots and lines. Some of the lines may be incomplete human figures. Seven turtles, nine human feet, and one probable dog were also recorded. Human forms compose the bulk of the remaining petroglyphs. Almost the full range of body types are represented: open- body, linear, and triangular-body forms. Only the columnar- body form was not recorded. The open-body type of glyph has the lower part of the torso left open. The feet, legs, and torso are formed by one continuous incised line. The head is often a part of this line and when it is, it is usually of the hawk-headed type. Alternatively, the head is depicted by a bruised dot above and separate from the torso. Arms and hands are incised lines appended onto the torso. Fingers and toes are often depicted. The male sex is shown on one figure with an arrow in the opening of the torso. Another open-body figure has a dot at the base of the torso and may represent a female. Several of these figures are depicted in a squatting position with the knees upraised. Linear figures are depicted by incised lines, with one vertical line representing the torso. The arms and legs are shown in two ways--with elbows and knees bent at right angles, or with arms and legs formed by curved lines. A combination of these two styles is also present. Arms are depicted as being either upraised or down or a combination of the two. The majority of the linear figures have fingers and toes; the number of appendages ranges from two to four. The male sex is often shown simply by an extension of the torso line below the juncture of the legs, sometimes with a swelling or widening at the end. The head of the figure is always depicted as a dot at the upper end of the torso line. Triangular-body figures are those in which the torso is formed by a triangle, apex down. The triangle usually has its entire center bruised, although there are some figures in which the triangle is formed by an incised line. The arms, attached to the base of the triangle, and the legs, attached to the apex, are usually curvilinear. The head is depicted as a dot, either separate above the torso, or connected to it by a line. Fingers and toes are often represented and some figures have hair or headdresses. A few glyphs even depict muscles on the arms and legs. Sex is not usually depicted on the triangular-body figures. Pictorial groupings or multiple units are generally considered rare in Hawaii (Cox with Stasack 1970:43-51). At least two such groupings were found at the petroglyph cave. One group consists of two tiny, detailed, fighting cocks inside a circle, with three small human figures positioned as spectators. The other group is a cluster of six, almost identical, triangular-body figures. No petroglyphs of European motif, such as letters, dates, or sailing vessels, were observed at the petroglyph cave. -Test Excavations Four test pits were excavated at the Hilina Pali Petro- glyph Cave. Three were excavated to reveal partially buried petroglyphs, while the fourth excavated FP-3 in order to obtain a carbon sample for dating the terminal occupation of the site. The test excavations uncovered 107 petroglyphs, some of which extended down to 62 cm below the surface. All forms were represented; 37.4 percent were human forms, with linear figures predominating. In addition, test pits 2 and 3 revealed the modifications that had been made in the petroglyph cave, The modifications are characterized as the building up of a raised and paved floor throughout much of the inner portion of the N section of the tube, which increased the total usable floor space in both the sheltered area and in the open sink. The sequence of modification involved the utilization of large roof-fall blocks, some up to the size of small boulders as a base layer. In some areas smaller boulders and cobbles were laid so as to provide a relatively level surface. The final layer was a paving of soil and small cobbles. Atop .. this pavement, across the mouth of the cave, the free-standing wall was built. Interestingly, and perhaps uniquely, some of this building activity resulted in the burial of petroglyphs on the lower cave walls. Radiocarbon Age Determinations Two carbon samples from the Hilina Pali Petroglyph Cave were submitted in order to determine the temporal span for the occupation of the site and, more importantly, to obtain absolute dates for the sub-deposit petroglyphs. This is the first time that direct archaeological dating of petroglyphs in Hawaii has been possible. One carbon sample, from the base of test pit 2, yielded a corrected date of A.D. 1540-1720 at 1 standard deviation (corrected dates based on Ralph et al. 1973). The carbon sample from Fireplace 3 yielded a corrected age of more recent than A.D. 1655. The absence of historic-period artifacts and petroglyphs indicates that this site was abandoned prior to significant European contact. Therefore, a temporal span for the occupation of the petroglyph cave can be approximated as A.D. 1600-1800. Artifacts Seventeen portable artifacts were recovered from the surface and from test excavations at the petroglyph cave. These artifacts can be separated into four functional classes (fishing gear, domestic implements, tools, and miscellaneous) that represent both extraction and maintenance types of activites (Binford and Binford 1967:71). The artifacts found consist of one finished and one unfinished fishhook, a tapa beater, two patellid scrapers, a birdbone pick, abraders of basalt, coral, and echinoid-spine, a grinding stone, a flaked basalt cobble, a utilized basalt flake, and drilled shell and worked bone. Contact with the coast is indicated by the artifact collection. The patellid shells were gathered at the coast, and fishing and fishhook manufacture are also indicated by the complete and unfinished fishhooks, the pieces of cut bone, and the abraders. Midden Analysis Quantitative analysis was performed on the midden material recovered from the excavations. The majority of this material consists of limpets, CeZZana exarata Nuthall (Hawaiian 'opihi). Small quantities of other shellfish species--Conus, Drupa, Cypraea, Nerita, etc.--were also recovered. Bone. of dog, pig, bird, and fish was sparse. The inhabitants were probably also utilizing food that does,not appear in the remains, such as poi, bananas, or breadfruit. The inhabitants of the site evidently subsisted primarily on marine resources that could only have been gathered some 3km. away. The results of the midden analysis suggest intermittent use of the site, or at least frequent contact with people from some coastal area. Associated Sites In addition to the petroglyph cave, sixteen additional sites were recorded in the area. These included surface petroglyph fields, shelters, trail segments, and three additional lava-tube sites. One of the lava-tube sites (HV-393) contained 50 concentrations of gourd remains in addition to 78 concentrations of sticks, grass, and charcoal, and two ash deposits. The remnants of a wooden bucket with a metal bail and half of a coconut shell (endocarp) were also found. All of this material was probably associated with water catchment. The gourds most likely had been placed under drips fromthe roof and the wooden bucket and coconut shell presumably served the same Eunct.ion. The sticks, grass, and charcoal could be the remains of torches. and the ash deposits are probably the result of small fires used for light. A carbon sample from this site yielded a corrected date of more recent than A.D. 1655. It seems likely that this site was utilized in the late prehistoric period, with probable sporadic use into early historic times. Conclusion Based on the results of the survey, excavations, and the radiocarbon dates, some tentative conclusions regarding chronology of sites in this area and the settlement pattern may be attempted. Two alternative hypotheses concerning the nature of occupation of the Hilina Pali Petroglyph Cave can be formulated on the basis of the stratigraphy and the variety of artifact types and midden material recovered: (1) the site was used intermittenly, but frequently, for temporary shelter during trips between coastal fishing sites and the dryland forest or the more distant upland forest of the Kilauea area; or (2) the site was used for longer periods, during the wet winter season, when the water-catchment gourds in the damper caves would have needed regular tending. It is also conceivable, of course, that the site was used in both ways. The latter hypothesis appears preferable, however, especially when the gourd materials in the neighboring cave sites, particularly Site HV-393, are considered. The petroglyph cave (HV-383) was occupied and a variety of modifications made in it, over a time span of approximately 200 years, c. A.D. 1600-1800. The modifications are characterized by the building-up of a raised, paved floor, which considerably increased the total usable floor space. This building activity resulted in the burial of petroglyphs on the lower cave walls. Utilization of HV-383 evidently terminated prior to European contact. No historic artifacts were found at the site, either on the surface or from test excavations. Furthermore, none of the glyphs recorded or observed were of European motif. We thus feel justified in concluding that this site and its petroglyphs date entirely to the indigenous Hawaiian period. The water-catchment site (HV-393) was utilized in late prehistoric times and into early historic periods as evidenced by the historic artifacts. Contact with the coast is indicated by the midden material and by the artifacts associated with fishing and fishhook manufacture. The fresh water collected in the tubes may have been transported to the coastal settlements along the arid coastline. In summary, the survey area was utilized in late prehistoric and early historic periods, and at least two activities account for the seventeen sites: (1) travel between the coast and the upland forest; and (2) collection of fresh water for coastal settlements. Literature Cited

Binford, Sally R., and Lewis R. Binford. 1967. "Stone Tools and Human Behavior." Scientific American 220(4): Carlquist, Sherwin. 1970. Hawaii: A Natural History. New York: The Natural History Press. Cox, J. Halley, with Edward Stasack. 1970. Hawaiian Petroglyphs. B. P. Bishop Museum Spec. Pub. 60. Honolulu. Ralph, E. K., H. N. Michael, and M. C. Han. 1973. "Radiocarbon Dates and Reality." MASCA Newsletter -9(1):1-20. University Museum, University of Pennsylvania. BIRD DISTRIBUTION AND ABUNDANCE ABOVE 3000 FEET IN HAWAII VOLCANOES NATIONAL PARK

Sheila Conant Department of General Science University of Hawaii at Manoa Honolulu, Hawaii 96822

I conducted studies of bird distribution and abundance in Hawaii Volcanoes National Park as part of a subproject of the US International Biological Program (IBP) Island Ecosystems Integrated Research Program. The results of part of this work were published in the program's technical report series (Conant,1975). This paper presents a brief summary of the work with emphasis on descriptions of spatial distribution patterns of the more abundant bird species found above 3000 ft in Hawaii Volcanoes National Park. The objective of the study was to examine patterns of bird distribution and abundance on an altitudinal gradient from 3,000 to 10,000 ft on the east flank of Mauna Loa. A transect (called the Mauna Loa Transect) spanning the ecosystems along this gradient was established by IBP researchers, and individual IBP study sites identified at intervals along the transect (see Fig. 1 for site locations). Eight one-mile bird census transects were established in seven different ecosystems that lay along the Mauna Loa Transect 1. Metrosideros tree-line ecosystem at 7,500 to 8,500 ft 2. Open Metrosideros-Sophora subalpine scrub at 6,500 to 7,500 ft 3. Acacia mountain parkland in a matrix of grassland at 4,500 to 6,500 ft (2 transects) 4. Acacia-Sapindus savannahat4,ZOO ft 5. Closed mesic, kFpuka forest at 4,000 ft 6. Open, dry Metrosideros forest at 4,000 ft. 7. Closed Metrosideros-Cibotium montane rain forest at 3.900 ft Censi were conducted monthly from December 1972 to July 1973, and from August 1974 to March 1975. One ecosystem on the altitudinal gradient (open wet Metrosideros-GZeichenia forest) was not sampled because it did not cover a sufficiently large area in which to sample bird populations. No census transect was established between 8,500 and 10,000 ft (alpine stone desert). Rather, I visited the area four times to assess the status of the single bird species found there. I used Emlen's (1971) "count x detectability" method for censusing. Values for bird species frequency and density for each site were calculated from the census data. Frequency and density data were subjected to two types of community ordinations to identify patterns of spatial distribution, and to determine if distributionally associated bird species groups existed. The first type of ordination was a community ordination and was based on a qualitative similarity index. Results of this analysis allowed evaluation of the degree of similarity among census transects. The method was described by McCammon (1968) and adapted by Mueller-Dombois and Ellenberg (1974). Census sites were "lumped" into zones on the basis of similarity. The second type of ordination was the two-way synthesis table technique (Mueller-Dombois and Bridges, 1975), for which I used the Ceska-Roemer (1971) computer program. This is a species ordination that allows recognition of distributionally associated species groups. It is based on presence-absence criteria for each bird species (Conant, 1975) . Similar analyses of plant communities were conducted previously (Mueller-Dombois and Bridges, 19751, and provides a basis for evaluating the influence of plant community structure on bird distribution.

Results Community ordinations. During censi on IBP study sites 22 bird species were recorded (Table 1). Community ordinations with these species resulted in the definition of six zones within the Mauna Loa Transect. Zone I -- Rhacomitrium-moss desert (8,500-10,000 ft) and Metrosideros tree-line ecosystem

Zone I1 -- Open Metrosideros-Sophora subalpine scrub Zone 111 -- Acacia koa mountain parkland Zone IV -- Acacia-Sapindus savannah(inc1uding a closed mesic kfpuka forest) Zone V -- Open dry Metrosideros-forest Zone VI -- Closed Metrosideros-Cibotium montane rain forest These zones are very similar to those derived from community ordinations of vegetation data (Mueller-Dombois and Bridges, 1975) . Species ordinations and distribution patterns. Figure 1 shows Datterns of nine native birds and Fieure 2 shows that of eight exotic birds. These 17 species were the most abundant, and it is possible to identify well-defined distribution trends for them. In the figures, bird species abundance (birds per 40 ha [lo0 acres1)i.s indicated by the height of the curves as explained in the captions. Species are grouped according to distributional similarities with high altitude species shown at the top, ubiquitous species at the bottom. The six transect zones established by community ordinations are shown at the top of the figure. Two-way table analyses resulted in the definition of two distributionally associated species groups of native birds (3 and 4). Qualitative assessment of spatial distribution patterns allowed the establishment of two additional groups (1 and 2). The 'oma'o (Phaeornis obscurus obscurus) was considered a single group (group 1) because of its distinct and unusual distribution. This bird occurred in low densities in the alpine stone desert, tree-line ecosystem and subalpine scrub (Zones I and 11). It was observed very rarely in the upper reaches of mountain parkland (Zone 111), and did not reappear on the transect again except in closed Metrosideros-Cibotium montane rain forest (Zone VI), where it occurred at its highest density. Group 2, including Nene (Branta sandvieensis), Golden Plover (Pluvialis dominica)and Pueo (Asio fZammeus sandwichensis), was found characteristically in open habitats. All three species occurred in the mountain parkland (Zone 111). Only the Nene occurred at tree-line (Zone I) and in the subalpine scrub (Zone 11), and of the three, only the Pueo was observed in the Acacia-Sapindus savannah (Zone IV). Neither Pueo nor NEne were observed in open dry Metrosideros forest (Zone V) where the Plover occurred. TABLE 1. Summary of the means of bird species densities (birds140 ha) in selected IBP sites on the Mauna Loa Transect. (P = present at a density of less than m01 1 bird per 40 ha. )

IBP Census Site Number Hean total Species no. of 1 3 16 4 5 7 9 12 individuals

N&; 'I0 California Quail Chukar Erckel's Francolin pheasants Golden Plover Spotted Dove Barred Dove Barn Owl Pueo Leiothrix Hawai'i 'Cfma'o Hawai'i 'Elepaio Mynah White-eye Hawai'i 'Amakihi Table 1. (Continued)

IBP Census Site Number Mean total Species 1 3 16 4 5 7 9 12 no. of individuals

' Apapane 511 65 267 335 88 243 112 19 1641 'I'iwi 16 - 40 15 28 127 13 - 239 House Finch 1 32 138 93 53 32 37 1 570 Spotted Munia - P 3 5 3 1 - - 32 Cardinal 1 1 46 7 5 1 - - 61

Mean number of individuals per 924 229 1110 931 629 1041 455 53 site Total number of species' per site Total number of native species 6 4 5 6 8 8 6 3

Total number of exotic species 3 8 11 8 8 7 6 3

h he two pheasants are added as one species - LIFE. FORM ;ROUI

'hoeornis o. obscurui --- Vanto sondvicensls I 'luviolis dominico I I \sio flammeus sondwichensis

~ ~.~......

lestiaria coccinea

-.--~ ~ ~~~~ :hasiempis s. sondwichensist

. ~ - .

luteo solitorius

~ Loxops v, virens I- ~- @ I-I N Hlmotione songuinea,

I I,00C

9,OK

7,mc

5.m

3.00C

(feel

ALTITUDE DISTANCE (km) Figure 1. Spatial distribution of endemic bird species on the east flank of Mauna Loa. Transect zones are those generated by the presence/absence Smrensen similarity index. N.S. = not sampled. Life-form groups : Amplitude scale: F = fruit-insect eaters (Birds140 hectares I100 acres]) B = browsers broken' line-< 1 R = raptorial predators 1 mn = 2- 5- 5 mm = 71-120 N = nectar-insect eaters 2 mm = 6-15 6 mrn = 121-160 I = insect eaters 3 m = 16-30 7 mm = 161-300 4 m = 31-70 8 mm > 301 LIFE- TRANSECT ZONESLI m FORM - iROUP -- I

~ ~ . ~~~..-

Zosterops joponico

I I I I I I I (feet) (meters) 0 4 8 12 16 20 24

ALTITUDE DISTANCE ( km) Figure 2. Spatial distribution of exotic bird species on the east flank of Mauna Loa. Transect zones are those generated by the presencefabsence S@rensen similarity index. N.S. = not sampled. Life-form groups: Amplitude scale: B = browsers (birds140 hectares 1100 acres1 ) F = fruit-insect eaters broken line 5 1 S = seed eaters 1mm= 2-5 5 mm = 71-120 I = insect eaters 2 mm = 6-15 6 mm = 121-160 3 mm = 16-30 7 mrn = 161-300 4 mm = 31-70 8 mm > 301 Group 3, a quantitatively defined group, had a wider distribution than group 2, occurring from the lower subalpine scrub (Zone 11) to the rain forest (Zone IV). Although the ' 10 (Buteo solitnrius) was observed in open dry Me trosid~ros forest (Zone V) , 'Elepaio (Chas iempis sandw ichcnsis sandwichensis) and 'I'iwi (vestiaria coccinea) were absent. This suggests that the latter two species are to some extent dependent on tree colonies, within which they were observed on censi. Group 4 includes the 'Amakihi (~0x0~svirens virens) and the 'Apapane (Himatione sanguinea), the two most abundant native birds. The birds were almost ubiquitous on the transect. Al- though they share a common range, the 'Apapane appeared to be less dependent on closed tree colonies as suggested by its higher densities in open dry Metrosideros forest (Zone V). Five spatial groups were established for the exotic birds, and two of these (3 and 5) were quantitatively defined. The Chukar (AZectoris chukar), in group 1, had the most restricted distribution of any species, being found only above the mountain parkland. The Red-billed Leiothrix (Leiothrix lutea) in group 2, had a rather unique distribution, although it was quite similar to that of the 'I'iwi, 'Elepaio, and '10, except that the Leiothrix was not observed in rain forest (Zone VI). However, Baldwin (1953) found this species in rain forest, so its disappearance from that zone is historically recent. Group 3 consisted of California Quail (Lophortyx caZifornicus), the pheasants (Phasianus spp.) and the Spotted ?.lunia (Lonchura punctuZata). These birds were found in the mountain parkland, savannahandopen dry Metrosideros forest (Zones 111, IV, and V), and can be considered "open area" birds. This group is distributionally- - similar, but not identical to, the group formed by Nene, Plover and Pueo. The last three exotic birds extended into the closed rain forest (Zone VI). The Cardinal (Cardinalis cardinaZis) was considered unique, but its distribution is almost identical to that of 'I'iwi, 'Elepaio, and '10. Thus, two exotic birds, Leiothrix and Cardinal, can be associated with group 3 or the native birds. These two exotics probably do not offer serious competition to the endemic birds because their food preferences are so different.

The last group (group 5) includes the House Finch (Carpo dacus mexicanus SrontaZis) and the Japanese White-eye (Zosterops japonica), which occupy the same range as the 'Amakihi and the 'Apapane. Although the House Finch is a seed eater, and thus does not compete with the two honeycreepers for food, the White- eye takes insects and nectar, just as the honeycreepers. However, at the extremes of the transect gradient, neither exotic species is as common as the native honeycreepers. It appears that along most of the transect the White-eye may well compete with 'Amakihi and 'Apapane for some habihat resources.

Discussion Censi conducted along the Mauna Loa Transect were important in an attempt to correlate the relationship of vegetation, elevation and climatic factors to bird distribution along the environmental gradients of the transect. Interpretation of distribution patterns can be made using Figures 1 and 2. Any species whose distributions began or ended abruptly in a sharp transition between different plant communities was considered to be limited by abrupt changes in vegetation structure. For example, densities of the Red-billed Leiothrix drop off abruptly at the lower limit of the Acacia-Sapindus savannahandat the upper limit of the mountain parkland, indicating that the species is limited by vegetation type rather than some physical environmental factor. Those species whose distributions did not terminate abruptly at a sharp boundary between plant communities were judged to be limited by environmental factors varying along the transect. For example, Spotted Munia densities gradually drop off at the upper and lower limits of its distribution, suggesting that altitude, temperature or rainfall are limiting factors for the species on the Mauna Loa Transect. In some cases, one end of a species' range appears to be limited by vegetation structure, while the other end appears to be limited by an environmental factor. The 'I'iwi is a good example of this because its density in Zone V (open dry Metrosideros forest) drops to zero from about 25 birds/ 40 ha (100 acres) in Zone IV (Acacia-Sapindus savannah), increasing again abruptly to about 16 birds140 ha (100 acres) in Zone VI (closed Metrosideros-Cibotium rain forest). At the upper end of its range 'I'iwi density falls off gradually in Zones I11 and I1 (subalpine scrub and tree line ecosystems). Using the above criteria, and considering upper and lower distribution limits of species separately, I judged that vegetation structure accounts for about 60 percent of the distributional limits of birds on Mauna Loa Transect, and continuously varying environmental factors for about 40 percent. The definition of six zones in the Mauna Loa Transect with distinct avian cornunities that closely correspond with vegetation transect zones lends further support to the conclusion that vegetation structure has a significant influence on the patterns of spatial distribution on this transect. Other studies of avian communities (MacArthur and MacArthur, 1971; Karr and Roth, 1971; Terborgh, 1971; Willson, 1974; Terborgh and Weske, 1975) have also demonstrated that the structure of vegetation influences avian cornunity structure in other parts of the world. In a study of the distribution of Peruvian forest birds along an environmental gradient (elevation) Terborgh and Weske (1975) found competition and continuously varying environmental factors (they included gradual changes in vegetation structure here), rather than sharp vegetation transitions to be the most important limiting factors for bird distribution. It is difficult to assess why abrupt changes in vegetation structure play a more important role in limiting species distributions on the Mauna Loa Transect than in Peruvian habitats, especially in the absence of comparative data on vegetation structure. Perhaps boundaries between plant communities are more distinct here than in the Peruvian study area because the Mauna Loa Transect crosses a number of rather drastic changes in substrate within an area that is geologically quite young (less than 1 million years). Although it is not an important factor determining species distribution limits, diffuse competition (i.e., competition among species exploiting a common pool of resources) may influence species densities on the Mauna Loa Transect where competitors occur together. For example, the 'Apapane and the Hawai'i 'Amakihi compete for food (Conant, unpublished). Although these two species occur together all along the transect, the peak density of the 'Apapane occurs in Zone VI (closed Metrosideros-Cibotium rain forest), that of the Hawai'i 'Amakihi in Zone 111 (Mountain parkland). One interpretation of these differences in density is that 'Apapane obtain food more successfully where nectar is the principal resource, as it is in Zones IV (Acacia-Sapindus savannah), V (open dry Metrosideros forest), and VI (closed Metrosideros-Cibotium rain forest). 'Apapane densities exceed those of Hawai'i 'Amakihi in these zones. Another interpretation of the data is that areas of high rainfall provide optimum habitat for 'Apapane, and that Hawai'i 'Amakihi are more successful in drier areas. This interpretation is also supported by the data, which show that Hawai'i 'Amakihi densities exceed 'Apapane densities in the drier transect zones. The two interpretations are compatible, and both may'be invoked to explain the distribution patterns of these species. Perhaps the most interesting distribution pattern is that of the 'Dma'o which is characteristically a rain forest species (Berger, 1972). Only on Mauna Loa has it been found in subalpine scrub and at tree line. Food and competition for food from the Red-billed Leiothrix, the only other fru- givore on the transect, may be limiting factors in the distribution of the 'Dma'o (Baldwin, 1953), and, at present, both species occur in rain forests elsewhere (Conant, 1975). However, plant species with suitable fruit are more abundant in closed rain forests than khey are in mountain parkland or Acacia-Sapindus savannah. The fact that Red-billed Leiothrix are present in dry forests where 'Oma'o are absent (Conant, 1975; van Riper and Scott, in review) suggests that it is more tolerant of dry forest conditions than the 9 -Oma'o. This hypothesis needs further investigation. Compared to continental tropical ecosystems (Karr, 1971, 1975; Karr and Roth, 1971; Terborgh and Weske, 1975) the avian communities in this study area have very low species diversities. Recent extinctions have increased the number of "empty niches" that existed in the original avifauna, few, if any, of which are now occupied by exotic species. Competition among endemic species is thought to be minimized by differences in anatomy (Lack, l97l), diet, habitat segre- gation, foraging strategies or behavioral interactions (Conant, unpublished). Thus, one of the principal ways in which spatial distribution of birds on the Mauna Loa Transect differs from that in continental ecosystems is that competition apparently does not play an important role in determining distributional limits.

Literature Cited Baldwin, P. H. 1953. Annual cycle, environment, and evolution in the Hawaiian honeycreepers (Aves:Drepaniidae). Univ. Calif. Publ. Zool. 54:285-398. Berger, A. J. 1972: Hawaiian Birdlife. Univ. Press of Hawaii, Honolulu. xxxi + 270 pp. Ceska, A., and H. Roemer. 1971. A computer program for identifying species-relev6 groups in vegetation studies. Vegetatio 23: 255-277. Conant, S. 1975. Spatial distribution of bird species on the east flank of Mauna Loa. US/IBP Integrated Ecosystems IF3 Tech. Rept. 74. 98 pp. Emlen, J. T. 1971. Population densities of birds derived from transect counts. Auk 88: 323-341. Karr, J. R. 1971. Structure of avian communities in selected Panama and Illinois habitats. Ecol. Monogr. 41: 207-233. Karr, J. R. 1975. Production, energy pathways, and community diversity in forest birds. In: Golley, F. B., and E. Medina (eds.) 1975. Tropica~Eco16gicalSystems: Trends in Terrestrial and Aquatic Research. Ecological Studies. Vol. 11, pp. 161-176. Springer-Verlag. New York. Karr. J. R.. and R. R. Roth. 1971. Vegetation structure 'and avian diversity in several ~ew-worldareas. Amer. Natur. 105:423-435. Lack, D. 1971. Ecological isolation in birds. Harvard University Press, Cambridge, Mass. xi + 404 pp. MacArthur, R. H.,and J. W. MacArthur. 1969. On bird species diversity. Ecology 42:594-598. McCammon, R. B. 1968. The dendrograph: a new tool for correlation. Geol. Soc. Am. Bull. 79:1663-1670. Mueller-Dombois, D. and K. W. Bridges. 1975. Integrated island ecosystem ecology in Hawaii. Spatial distribution of island biota: Introduction. US/IBP Island Ecosystems IRP Tech. Rept. 66. 52 pp. Mueller-Dombois, D.,and J. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley and Sons, New York. xx f 547 pp. Terborgh, J. 1971. Distribution on environmental gradients: Theory and a preliminary interpretation of distributional patterns in the avifauna of the Cordillera Vilcabamba, Peru. Ecology 52:23-40. Terborgh, J. and,J. S. Weske. 1975. The role of competition in the distribution of Andean birds. Ecology 56:562- 576. Willson, M: F. 1974. Avian community organization and habitat structure. Ecology 55:1017-1029. van Riper, C. and J. M. Scott. Patterns in the distribution and abundance of the Palila on Mauna Kea. Auk (in review) VARIATION OF HAWAIIAN METROSIDEROS ALONG THE SOUTH AND EAST FLANKS OF MAUNA LOA, HAWAII Carolyn A. Corn Botany Department University of Hawaii at Manoa Honolulu, Hawaii 96822

The genus Metrosideros (Myrtaceae) grows as a tree on well-drained soils, or as a shrub in bogs and windy mountain ridges. It occurs on various soil substrates from sea-level to 2500 meters under annual rainfalls ranging from 750 to 11,250 mm per year. The flower color is variable, ranging from pink, red, orange, yellow, and very rarely white. Red is the most common color. Many insect and bird species are attracted to the open flower clusters that supply abundant pollen and nectar. Ample opportunity for transmittal of pollen between trees has been found on the feathers around birds' beaks (Corn, 1972). The amount of genetic exchange between plants is dependent upon the range of the pollinating animals visiting the flowers. since the pollen is sticky and would only be blown under very strong wind conditions. Genetic exchange can also take place through dispersal of the small seeds which may sprinkle like rain when gusty trade winds are blowing. Throughout this paper the taxonomic classification will follow Rock's monograph descriptions (1971). All Metrosideros on the island of Hawai'i are classified M. coZZina subspecies po Zymorpha. Varieties poZymorpha, incana, imbricata, macro- phyZZa, newetlii, gZaberrima £. sericea, and a variety nuda, described subsequently by Skottsberg (1944) are found on Mauna Loa, HawaiY. Metrosideros rarely has distinctive characteristics that can be used to define the species, varieties, and forms. The intergradation of plant characters has caused taxonomic chaos and may be best described as a species complex. Variation is found in such plant characteristics as tree form and size; flower color; seed size and color; leaf shape, size, thickness, and pubescence. Overlapping distribution of taxa is common on the six largest Hawaiian Islands, and the chances of hybrids forming between morphologically different individuals may be great. My discussion of Metrosideros will be limited to its vegetative morphology and leaf anatomy as it is found on the south and east flanks of Mauna Loa, Hawai'i. Sastrapradja and Lamoureux (1969) report IW dist.ir~ct differences in wood anatomy among the taxa. Yet the external appearance of plants growing next to each other may be distinctive. Corn and Hiesey (1973) found that seeds gathered from Mauna Loa, Hawai'i, and Maui at different elevations and grown in a common environment near sea level retained differences that were indicative of the habitat from where they collected the seed. Seedlings from high elevations grew least while seedlings near sea level grew the largest in 23 months of experimentation. These results demonstrated differences among the plants from different elevations that may be an ecotypic formation that forms a cline with changing elevation. Since no comparable quantitative data of trees in their natural habitat had been collected,a preliminary survey along an altitudinal transect was made in 1973 to verify if certain patterns occur that may be attributable to changes in climatic parameters. METHODS Thesurvey was conducted at twelve sites--eleven of them in Hawaii Volcanoes National Park and one from the Saddle Road (Figure 1). Most of these were located on lava flows, either piihoehoe or aa, and from drier habitats within the park. Sites were chosen over an altitudinal range from 46 to 2438 meters without particular concern for rainfall or soil substrate (Table I). One additional site was chosen (site 12) to obtain a comparison of low and high rainfall sites (site 2 and 12). Both sites were located on pzhoehoe lava at the same elevation. Fifty trees per site were sampled by gathering one branch at random from each tree. From each branch stratified leaf samples were gathered by collecting five leaves--the largest, smallest and three additional leaves--and computing the average size for each branch. Leaf and branch characters measured included leaf length, width, area, thickness, petiole length, and internode length. For each characteristic the fifty average values per site were measured and recorded on graphs according to the site's altitudinal and rainfall parameters. Yearly rainfall was computed from U. S. Weather Service data, Inter- national Biological Program Reports No. 38, 59, and 70, and isoheit maps. Tree height, trunk diameters and tree spacing werealsoobtained for fifty trees at each of the sites. Since very high correlations were obtained between the climatic parameters and leaf characteristics, a further study was begun using six selected sites A through F (figure 2). These six sites were carefully chosen to vary in only one climatic variable--altitude or rainfall. All six sites are on pahoehoe lava with a southern exposure. Four sites (A through D) compose the altitudinal transect along the south flank of Mauna Loa. Three sites (B, E, and F) compose the rainfall transect from a low to high rainfall regime. The two transects cross each other in a T fashion at site B. Table I1 summarizes the climatic and topographic features at these sites. Rainfall for one year was collected at or near each of the six sites starting 18 July 1974, which happened to be a wet year. To assure that leaf sections were obtained from mature leaves, leaves for microscopic analysis were picked at random from the middle of the second flush back from the end of the branch. These leaves were then sectioned by Hooker Microtome at twelve microns thickness and stained with toluidin blue 0. RESULTS In the preliminary survey tree heights and trunk diameters were measured for fifty trees per site (figure 3). All twelve sites are illustrated with each dot representing a tree height and trunk diameter. A curve is drawn that gives a representative growth curve for each site. The eleven sites along the altitudinal transect, excluding the Saddle Road site at 700 meters elevation,illustxate a series of growth curves. The trees are rather stunted and do not grow very tall at the lower and upper extremes of the altitudinal transect. Tree heights are greatest at middle elevations, especially where there is higher annual rainfall and ash soil. This trend is also apparent when comparing the low and high rainfall sites at 700 meters in the lower middle portion of this diagram. The drier Hilina Pali site has trees that do not grow very tall, while the wet Saddle Road site has trees that appear to be growing tall at a faster rate. However, the pshoehoe lava at the Saddle Road site is of recent origin, so that no trees larger than nine meters tall were sampled. Measurements of sites (A through F) on pghoehoe lava give similar results as the preliminary survey. Tree height on the drier south side of Mauna Loa is greatest at middle elevations decreasing at both high and low elevations. Tree height also is greater at sites receiving greater rainfall. Although most trees appear intermediate between two or more varieties, trees in sites A through F were classified according to the variety that they seemed most alike. The six varieties present in each of the six sites along the two transects are diagramed in figure 4. Variety imbricata is most numerous in lower, dry sites. Variety poZymorpha is found throughout the dry altitudinal transect, but decreases in number until it is virtually nonexistent in wet localities. Variety incana is most common at lower and middle elevations where rainfall is less. Variety gzaberrima forma sericea is virtually nonexistant in dry localities but becomes dominant in higher rainfall sites. Likewise, varieties macrophyZZa and newcZZii are present where rainfall is highest. These latter three varieties, that are blackened in, in the diagram, may be call-ed rainforest varieties, since they occur in greatest frequency in high rainfall areas. Likewise those varieties with pubescent leaves may be called dry-land varieties, although one should realize that the transition from dry to wet varieties is gradual. Figures 5 through 10 diagram results of leaf and branch morphology from the preliminary survey. In each of these diagrams, the horizontal line is the range per site, the vertical line is the mean per site, and the blackened rectangular box portion is the standard deviation. The lowest two sites and the upper four sites in each of these diagrams are on si-milar soils in a low rainfall area, and five trees on each of these sites have been correlated for changes with elevation. The middle six sites are on differing soils and rainfall regimes. Therefore, they have not been included in the following correlations. Figures 5, 6, and 7, representing leaf length, width, and area measurements, all give similar trends along the altitudinal transect. Theranges and standard deviations are less at the extreme high and low elevations and greater at middle elevations where the rainfall is higher. The means of the sites also have a similar trend. Leaf length and leaf area data (figures 5 and 7) have a similar appearance. Figure 6 (leaf width) also resembles this pattern., but the differences between the sites are not as pronounced. If one eliminates the six middle sites that are on different soi.1~and/or receiving more rainfall, a very significant linear correlation is found (P=0.001) for the six sites on pahoehoe lava for figures 5, 6, and 7. This linear correlation is also being found in data being analyzed from sites A through D that include two sites from middle elevations. The two sites diagramed in figures 5 to 7 at 700 meters represent the seasonally dry IIilina Pali site to the left and the high rainfall Saddle Road site to the right. The mean, standard deviation and range of leaf, length, width, and area is greater in the wet site. Data from sites B, E, and F also support this trend. Leaf thickness in tenths of a millimeter is shown by site elevation (figure 8). Leaf thickness is greater from trees at high elevations with a major increase in leaf thickness approximately at frost line. A trend toward shorter petiole length from trees at higher elevations might be suggested in figure 9, but this trend is not dramatic, Internode length does not show any particular trend w-Lth changes in elevation (figure 10); however, it does increase with increasing rainfall. The external leaf morphology which has been discussed is also reflected in the laminar leaf anatomy. Leaves collected from trees at high elevations have thicker leaf sections than leaves collected from trees at low elevation. Leaf portions that are thicker include the upper cuticle, upper epidermis, hypodermis and mesophyll. The number of cell layers in the hypodermis and mesophyll also increase. The hairs on the lower portion of the leaf become longer. (Extra epiphytic fungus is found associated with leaves that have long hairs on the under side of the leaf.) The following are some of the major changes in leaf anatomy when leaves are sampled from low, medium, and high rainfall sites. Pubescence and upper cuticle thickness are less from trees at high rainfall sites. At the same time the hypodermis and mesophyll within the leaf show an increase in their thickness. In surnmary,the results of the anatomical evidence along both transects are as follows: from trees at a higher elevation all parts of the leaf cross section are thicker, except for the lower epidermis and cuticle; from the trees at high rainfall sites, the upper cuticle thickness is thinner, the leaf pubescence basically is absent, and the hypodermis and mesophyll portions are thicker. CONCLUSIONS Differences in laminar leaf measurements are greatest at extreme altitudinal and rainfall sites. The leaf measurements many times follow a cline. As one samples plants along an altitudinal transect from sea-level to timberline, Leaf size measurements become smaller, while leaf thickness and hair length measurements increase. Petiole length also decreases, but internode length remains unchanged. As one samples plants along a rainfall transect from a low to high rainfall, leaf and internode measurements change. The leaf size increases, the pubescence under the leaves disappears, and petiole and internode length increase. These morphological features are manifested as anatomical differences within the leaf. How many of these differences are genetically influenced and how many of them are environmentally influenced is open to question. Clones of different varieties are being grown in a greenhouse at the University of Hawaii (Manoa) that can be compared to the plants in the field that may help answer this question. Also some of the varieties from Mauna Loa were hybridized and the F1 progeny are being grown. ACKNOWLEDGEMENTS Support for research of this paper was made possible by NSF GB 23230 to Island Ecosystems IRP/US IBP Hawaii and Pacific Tropical Botanical Garden Research Grant. TABLE 1. Description of sites* of sampled Metrosideros at different elevations, in Hawai'i Volcanoes National Park (except Site 12). (See also Figure 1.) o) 0 Average Average annual annual Site Elevation rainfall temp. No. Locality (m) (mm) (00 Soil Exposure Lower Chain of pshoehoe flow south Craters Road Hilina Pali 700 pzhoehoe flow with south scattered surface ash Thurston Lava Tube 1158 pahoehoe flow with south ash and humus Tree molds 1220 pahoehoe flow with north ash and humus ~ipuka-ki 1280 ash and humus south Ke-'a-muku flow 1570 'a'3 flow south Ke- ' Z-muku flow 1722 'a's flow west End of Strip Road 2012 weathered pzhoehoe flow south Mauna Loa Trail 2133 pahoehoe flow south Mauna Loa Trail 2286 pshoehoe flow south Mauna Loa Trail 2438 pahoehoe flow south Saddle Road 700 pahoehoe flow east *from preliminary survey TABLE 2. Description of sites (A through F) of sampled Metrosideros at different elevations and rainfalls, on Mauna Loa, Hawaii. (See also Figure 2.) Sites A, B, and D are in Hawaii Volcanoes National Park. Average" Annua 1 annual Elevation rainfall temp . Site Locality (d (m) ("C) A Lower Chain of Craters Road (same as Site ijl, Table 1)

B 300 meters northwest of Hilina Pali Road, 4 kilometers past ~Z~uka-nCnC 808 2062 (1902) 18.3

C Kala-pana Forest Reserve, approximately 1570 1891 (1388) 14.1 1 kilometer west of 1880 flow

D 500 meters east of Mauna Loa Trail 2332 1906(1452) 10.7

E 1.6 kilometers southeast of Highway 11, 808 5185 (4478) 18.0 beside gravel road

F Saddle Road 808 5875 (5212) 18.0

'One year rainfall measurements (18 July 1974 to 18 July 1975) 'Median annual rainfall interpolated from U.S. Weather Service data and International Biological Program rainfall data (IBP Tech. Rep. Nos. 38, 59, and 70). 3Average annual temperature interpolated from U.S. Weather Service data. Fi ure 1. Location and elevation Fi ure 2 Location and elevation +in meters) of Metrosideros sites +In meters) of Metrosideros sites sampled in the preliminary survey sampled in later study on Mauna Loa, on Mauna Loa, Hawai'i. See also Hawai'i. See also Table 2. Table 1. TRUNK DIAMETER (m)

Figure 3. Metrosideros tree spacing and fitted growth curves at the 12 preliminary survey sites. 1000 mm. 6000 mm. LOW : HIGH RAINFALL

Figure 4. Frequency by variety of Metrosideros along altitudinal and rainfall transects on Mauna Loa, Hawaii.

-.._-_- SADDLE ROAO

LEAF LENGTH (cm).

Figure 5. Mean,range, and standard deviation of Metrosideros leaf lengths from the 12 preliminary survey sites. SADDLE ROAD

LEAF WIDTH (cm) Figure 6. Mean, range, and standard deviation of Metrosideros leaf widths from the 12 preliminary survey sites.

LEAF AREA (cm2) Figure 7. Mean, range, and standard deviation of Metrosideros leaf areas from the 12 preliminary survey sites. LEAF THICKNESS (mm) Fi ure 8. Mean, range, and standard deviation of Metrosideros kcknesses from the 12 preliminary survey sites.

-7-- --- 'A 'A i i 3 4-i6 i a I PETIOLE LENGTH (mm)

-Figure 9. Mean, range, and standard deviation of Metrosideros petiole lengths from the 12 preliminary survey sites. INTERNODE LENGTH (mm)

Figure 10. Mean, range, and standard deviation of Metrosideros internode lengths from the 12 preliminary survey sites. Literature Cited

Bridges, K. W. & G. V. Carey. 1974-75. Climate data for the IBP sites on Mauna Loa, Hawaii. Island Ecosystems IRP, U. S. International Biological Program Technical Report No. 38, 59, 70.

Corn, C. A. 1972. Genecological studies of Metrosideros. In Island Ecosystems Stability and Evolution Subprogram Technical Report No. 2. Second Progress Report and Third-year Budget. P. 88-95.

Corn, C. A. & W. M. Hiesey. 1973. Altitudinal variation in Hawaiian Metrosideros. Arner. 3. Bot. 60(10):991-1002. Rock, J. A. 1917. The Ohia Lehua trees of Hawaii. Botanical Bull. No. 4. Division of Forestry, Honolulu.

Sastrapradja, D. S. & C. Lamoureux. 1969. Variations in wood anatomy of Hawaiian Metrosideros (Myrtaceae). Annales Bogorienses 5(1):1-83. Skottsberg, C. 1944. Hawaiian vascular plants. IV. Meddelanden fran Goteborgs Botaniska Tradgard 15:402-209. Taliaferro, W. J. 1959. Rainfall of the Hawaiian Islands. Hawaii Water Authority, State of Hawaii. Honolulu, Hawaii. U. S. Department of Commerce. 1974-75. Climatological Data, Hawaii and Pacific: Annual Summary. 70(13) and 71(13). National Oceanic and Atmospheric Administration, Environmental Data Service, Asheville, N. C. MITES ASSOCIATED WITH A FUMAROLE IN HAWAII VOLCANOES NATIONAL PARK* M. Lee Goff Entomologv De~artment ~ishoc'~us'eum Honolulu, Hawaii 96818

During May 1971, several steam vents or fumaroles in the Hawaii Volcanoes National Park were sampled as part of the soil arthropod studies connected with the Hawaii International Biological Program. Only one of these fumaroles will be considered in this presentation. The fumarole in question is located lkm west of the park headquarters along Steaming Bluffs Trail and has been designated Vent I1 in the series of fumaroles investigated. This fumarole measures 8.3m in length and has a maximum width of 3.2m. The actual steam emitting opening is located 1.5m from the east end of the fissure at 1.7m below the surface of the ground. A total of 81 samples were removed (Table I). Each of the samples was lOcm square and lcm thick. Surface temperature and relative humidity were recorded for each sample at the time of removal. Samples were examined with a dissecting microscope and them processed with a Berlese funnel to extract the arthropods. Recorded temperatures for the sample sites ranged from 25OC to 72'C. Subsequent remeasurement of temperatures over a 3 month period did not show any substantial fluctuation of temperature at individual sites. Relative humidity (RI-I) in the area around the fumarole at the time of sampling was 62 percent at a temperature of 25°C. Sample sites at the periphery of the fumarole ranged from 65 to 76 percent RII. The only exceptions to this were sites 38, 43, 50 and 58 which were located in a dense growth of fern, mosses and grasses. Relative humidity values at these sites ranged from 92 to 100 percent. Sites within the fumarole ranged from 92 to 100 percent RH. An extensive analysis of the vegetational component of the fumarole has not yet been completed. A preliminary assortment of plants into 12 associations has been made and related to temperature (Table 11). This association is purely of a preliminary nature. "Studies upon which this report is based were supported by NSF Grants GB23075 and BMS70-00697 to the Bishop Museum, Honolulu, Hawaii 96819. Twelve species of the superfamily Parasitiodea (Mesostigmata: Gamasina) have been recovered from this fumarole. These species represent seven families of free-living mites: four species of Ascidae, three of Phytoseiidae and one each of Laelapidae, Paraholaspidae, Parasitidae, Podocinidae and Rhodacaridae. The vast majority of collections containing these mites were made at temperatures between 31°C and 50°C (Table I). These collections comprise 95 percent of the total number of mites collected. A single specimen of Phytoseiidae, AmbZyseius sp., was recovered at 40°C. Only members of the Ascidae were found in samples taken at temperatures greater than 40°C. Two species, Cheiroseius sp. 1 and Cheiroseius sp. cf. necorniger, were present at temperatures greater than 46OC; a single individual of Cheiroseius sp. 1 was found in a sample taken at 58°C. It may be significant that none of the species of Ascidae were recovered from samples taken at temperatures less than 31°C in the fumarole. Three species of Eupodidae (Prostigmata: Eupodostigmatina) were found in the fumarole. Eupodes sp. 1 was found in a sample containing mosses and grass at 25OC. Eupodes voxencoZZinus Sig Thor, 1934, was also collected at 25OC from a sample of grasses. Ten specimens of a species representing an undescribed genus were recovered from samples of mosses at temperatures of 40°C and 41°C. Discovery of mites within a fumarole represents another instance of arthropods occuring in stress environments on the island of Hawaii. Howarth (l973), Schultz (l973), Gertsch (19731, Fennah (1973), Bellinger and Christiansen (1974) and Gap6 and Howarth (1975 a & b) have described various arthropods from lava tubes in the Hawaiian Islands. Fumaroles, like caves, are effectively isolated as microenvironments from the surrounding areas. In the case of the fumarole the isolating factors are temperature and high humidity. The Gamasina associated with the fumarole are from soil- and litter-inhabiting groups which are widely distributed throughout the Hawaiian Islands. The family Ascidae, which contains the species occurring at the highest temperatures within the fumarole, is well represented in the Hawaiian Islands and along the Mauna Loa transect of the island of Hawaii. The same is true of the family Eupodidae. It is noteworthy that of a total of fifteen species of Acari involved, seven species (including those found at high temperatures) and one genus represent undescribed taxa. TABLE 1. Distribution of Mesostigmata related to temperature. Numbers in parentheses indicate numbers of individuals represented. 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61+ No Temp. OC 0 C O C C OC C OC C OC Data No. of samples 2 0 10 25 2 4 9 2 2 5 2

No. of samples 1 0 7 12 11 3 0 1 0 1 with Mesos tigmata

Cheiroseius sp. (undescribed)

Cheiroseius nr. necorniger

Asca spicata 1(1)

Proctogamase ZZus sp. (undescribed)

TyphZodromus neobakeri 1(1)

HoZaspuZus tenuipes l(1)

Hypoaspis queens Zandicus TABLE 2. Vegetation related to temperature w

Plant 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61+ Notenp. Totalno. N Associations C O C0 CO - c C O C C O C" C" data of samples No visible 1 4 1 2 plant material 8 Algae Algae, Hoss Algae, Moss, Me trosideros Algae, Fern Algae, Grasses Moss Moss, Fern 1 Moss, Grasses Moss, Fern, 1 Grasses Moss, Metrosideros Grasses No vegetation data

Total number 2 0 10 25 24 9 2 2 5 1 81 of samples Literature Cited

Bellinger, P. F., and K. A. Christiansen. 1974 The cavernicolous fauna of Hawaiian lava tubes. 5. Collembola. Pacific Insects 16: 31-40.

Fennah, R. G. 1973. The cavernicolous fauna of Hawaiian lava tubes. 4. Two new blind Oiarus (Fulgoroidea: Cixiidae). Pacific Insects 15: 181-184

Gagng, W. C., and F. G. Howarth. 1975a. The cavernicolous fauna of Hawaiian lava tubes. 6. Mesoveliidae or water treaders (Heteroptera). Pacific Insects 15: 399-413

and . 1975b. The cavernicolous faunxHawaiian lava tuFes. 7. Emesinae or thread legged bugs (Heteroptera: Reduviidae). Pacific Insects 16: 415-426.

Gertsch, W. J. 1973. The cavernicolous fauna of Hawaiian lava tubes. 3. Araneae (Spiders). Pacific Insects 15: 163-180.

Schultz, G. A. 1973. The cavernicolous fauna of Hawaiian lava tubes. 2. Two new genera and species of blind isopodscrustaceans (Oniscoidea: Philosciidae). Pacific Insects 15: 153-162. HYDRATION-RIND DATES FROM ARCHAEOLOGICAL SITES IN THE SOUTH POINT AREA: A CONTRIBUTION TO HAWAIIAN PREHISTORY Terry L. Hunt Department of Anthropology University of Hawaii at Hilo Hilo, Hawaii 96720

INTRODUCTION Over the past two decades stratigraphic excavation in the South Point area on the Island of Hawai'i, has produced a wealth of archaeological data. Archaeological data from the South Point area have been the basis for seriation and relative chronology of fishhook types and subsequent theories concerning the origin and migration of Polynesians to Hawai'i. However, while several archaeological components were successfully dated by means of radiocarbon analysis, other important sites remained undated. Problems with the radiocarbon dating for site H-1, and a reassessment of available dates was published by Emory and Sinoto (1969) followed by a publication by Green (1971). Both publications assigned approximate temporal chronologies, but they were not in total agreement. Sites at Mahana Bay, near South Point, remained undated; these sites contained artifact types indicative of a significant time-depth, and absolute dates were crucial in any synthesis of a prehistoric sequence. The advent of hydration-rind dating of basaltic-glass artifacts in the Hawaiian Islands has provided an efficient, inexpensive, precise dating method. This method was applied by the author, who saw the beginnings of this project. THEORY AND METHODOLGY Glass is essentially a super-cooled liquid which, for the most part, is not in equilibrium with its environment. As a consequence, glass changes in state toward a more stable substance through a geochemical process known as hydration. The resulting product of the weathering of basaltic-glass is an amorphous, clay-like substance, palagonite. With regard to the progressive palagonitization of basaltic- glass, Morgenstein and Riley (1974) have concluded that following occurs : 1. Water enters the glass structure via small capillary channels developing the "mist zone,,I 2. an immobile product layer then develops from the mist zone, and 3. when there is an increased and sufficient quantity of water in the immobile layer it converts to a band of palagonite of the same thickness as the original mist zone. Hydration is not interrupted during the process. A new mist zone enters the fresh glass, even as the original mist zone develops into an immobile product layer. Therefore, the palagonite band thickness is directly related to the period of hydration. The hydration rate of Hawaiian basaltic-glass has been established at 1.177 micrometers per 100 years (Morgenstein and Riley, 1974). A hypothetical rate was confirmed by measurement of palagonite bands of glass specimens from historic lava flows on the Island of Hawai'i. In order to obtain archaeological dates, i.e. dates of human modification, basaltic-glass artifacts are cut perpen- dicular to the flaked surface and thin-sectioned so that the palagonite layer can be observed and measured under the microscope. All dates reported in this paper represent one standard deviation with no less than five measurements to no more than ten. Measurements were taken on more than one hydration layer when present. Basaltic-glass artifacts prepared and dated by this author were selected from significant levels of H-1 and the Mahana Bay sites. BASALTIC-GLASS ARTIFACTS Flaked basaltic-glass provides a sharp edge which could have served to cut hair, plant and fibrous materials. Barrera and Kirch (1973) suggest that these core and flake tool assemblages could have been utilized in food preparation, fine woodworking, or simply as a prehistoric "pocket-knife," to use a modern analogy. Further research of edge damage patterns on basaltic-glass artifacts must be undertaken in order to more fully understand the function of these ubiquitous tools. EARLY PREIiISTORIC CHRONOLOGY IN HAWAII A tentative summary of prehistoric Hawaiian chronology must begin with a mention of the Bellows Beach site (018), on Windward O'ahu, currently the earliest known site in the Hawaiian Islands (see Kirch, 1974:lll). The available dates, established by hydration-rind dating and radiocarbon age- determinations, place 018 within a temporal range of c. A.D. 400 to A.D. 960. The artifact assemblage recovered from 018 provides important implications concerning the origin of Polynesian settlers to Hawai'i, and incipient cultural adaptations to the Hawaiian environment. Although 018, most likely, does not represent the earliest Hawaiian site, it is indicative of an assemblage closely related with the "archaic" Eastern Polynesian artifact types (see Sinoto and McCoy, 1975.143-186 and Kirch, 1975:65). Data from HZlawa Valley, Moloka'i, site Al-3 have provided further evidence for the origin and cultural adaptations of early Polynesian settlers in Hawai'i, with earliest settlement at c. A.D. 600 in an ecologically-focal location (Kirch, 1971a, 1971b, 1974, and 1975). Significant discoveries at site A1-3 include a round-ended house and early or adaptational artifact types for the Hawaiian archipelago. 'ho very important sites have been mentioned concerning early Hawaiian prehistory. Stratigraphic excavation has revealed relatively few early components, i.e. A.D. 400 to A.D. 1000. Kirch (1975:64) suggests that "under normal intrinsic rates of population growth, population size (and hence the total number of occupation sites) would remain low for 400 or 500 years following colonization." This demographic consideration explains the relatively numerous excavated sites with components dating c. A.D. 1000 to the post-contact period. Significant sites with early components include: Wai-'Ahukini Site H-8, c. A.D. 750 (Sinoto and Kelly, 1975), Kuli-'ou'ou Shelter Site 01, c. A.D. 1000 (Emory and Sinoto, 1961:14), Nihoa Island, A.D. 890?90, Hale-a-ka-la Crater, c. A.D. 800 (exploitation of andesitic basalt for adz manufacture) (Kirch, 1974:117), and 'Anae-ho'omaluc. A.D. 900 (Barrera, 1971:lOZ). The remainder of this paper provides new data which may be included in the correlation of early, or other wise significant archaeological components. THE PU'U-ALI'I SAND DUNE SITE, H-1 Site H-1 is located approximately 0.4 kilometers from South Point proper. Excavation of H-1 was undertaken throu hout the 1950's and inter reted as a fisherman's workshop T~mory,Bonk, and Sinoto 1868). The cultural deposit yielded over 1,710 fishhooks, most of which were early types for the Island of Hawai'i, Emory, Bonk, and Sinoto (1968) provide an analysis of these recoveries. Radiocarbon analysis of organic samples from H-1 produced inconsistent dates, which seem to indicate contamina- tion. However, Emory and Sinoto (1969:13) established, from the radiocarbon dates, an estimated duration of occupation from c. A.D. 1250 to c. A.D. 1600. Emory and Sinoto conclude the following: It is not possible to determine accurately the duration of the fisherman's establishment at H-1 from the radiocarbon dates alone. They do, however, when viewed along with the cultural evidence, and when compared with radiocarbon dates from site H-8, allow for a minimum estimate of 200 years and a maximum of 400 years for its occupation. It seems that the best we can do in the present state of our knowledge is to think of the fishermen at South Point as having set up their establishment not earlier than A.D. 1000, and having continued its occupation to about A.D. 1350 Emory and Sinoto's suggested temporal chronology of H-1 is comparable to initial hydration-rind age-determinations, which provide evidence of occupation beginning c. A.D. 967563 to abandonment, c. A.D. 1395227 (Table 1.).

Table I. Hydration-Rind Dates for South Point H-1 Provenience Depth (inches) Date (A.D.) + o H 9 lowest 6 967+63 I3 lowest 5 1023i19

IUHANA BAY SITES The Mahana Bay sites, 4.5 kilometers northeast of South Point proper, are located in an environment adjacent to a rich maritime ecosystem. These sites are discussed in order of absolute chronology. Mahana Bay site MB-1 comprises c. 47 square meters enclosed by a single course stone alignment, somewhat square in shape. Vertical excavation of c. 30 square meters has revealed coral pebble paving, hearths, postmolds, and stone working surfaces. The cultural deposit is between 20 to 35 cm in depth. The major midden component is that of edible molluscs, recovered at all levels of the site. Portable artifacts recovered are basically associated with exploitation of marine resources. In addition, post-European artifacts of metal and glass were recovered in the upper levels of MB-1. The earliest date for bB-1 was obtained from a square located directly outside of the enclosing stone alignment, and probably represents the locus of an earlier structure, which has been eroded away by weathering and tidal action.

Table 2. Hydration-Rind Dates for Site MB-1

Provenience Depth (cm) Date (A.D.) i o

Mahana Bay site MB-4 comprises a rectangular area of c. 110 square meters enclosed by a stone wall ranging from a ground-level alignment to 1.50 meters in height. Vertical excavation has revealed a cultural deposit between 20 to 60 cm in depth, including features such as hearths, stone-slab paving, pebble paving, postmolds, wall storage vaults, and a feature best interpreted as a stone "anvil" working surface. Portable artifacts are primarily associated with procurement of marine resources. Diagnostic material-culture traits from MB-4 may be compared with other early sites in the Hawaiian Islands. These include: fishhooks with HT1 head-types, some of pearlshell, a type 1 pig-tusk pendant (Cox, 1967), and intrusive postmolds revealed at the lowest sector (in sterile volcanic-ash) of MB-4 could be interpreted as evidence of a round-ended house form, comparable to those discovered by Kirch (1971a, 1971b, 1975) at HBlawa Valley, Moloka'i. Post-European artifacts, such as bent- nail fishhooks and glass trade beads provide evidence of occupation into the historic period.

Table 3. Hydration-Rind Dates for Site MB-4

Provenience Depth (cm) Date (A.D.) i o NW49 32-24 981+105 NW48 24-26 1276+33 NW4 7 20-22 14192 27

Mahana Bay site MB-6 consists of a collapsed lava tube with a crude stone terrace at the entrance of the cave. Midden remains cover an area of about 10 square meters. Cultural deposits have filtered down through loose stones of the terrace, making stratigraphic excavation impossible. How- ever, the artifact assemblage from MB-6 indicates emphasis on procurement of marine resources. The salient attribute of MB-6 is water seepage in the interior chamber; here gourd fragments were recovered, which indicate that MB-6 served as a source of potable water for inhabitants of Mahana Bay. Currently, the earliest date for MB-6 is A.D. 1460+27, but MB-6 was undoubtedly exploited much earlier as a water source. A basaltic-glass artifact, dated A.D. 1649t24 was recovered with pig and chicken remains and, by inference, is associated with agricultural systems.

Table 4. Hydration-Rind Dates for Site MB-6

Provenience Depth (cm) Date (A.D.) 5 o Midden, cave entrance Midden, cave entrance

Mahana Bay site MB-3 consists of a roughly square enclosure of c. 41 square meters with stone walls approximately 1 meter in height. Vertical and horizontal excavation has revealed large stone-slab paving, basalt pebble paving, postmolds, and a large square stone hearth near the center of the walled enclosure. Unfortunately, a complete quantitative midden analysis is not yet available for MB-3. However, a significant midden component is pig remains recovered from several levels throughout the site. The presence of pig, by inference, provides evidence of dependence upon agricultural systems. The artifact assemblage is indicative of late prehistoric types. Several historic artifacts were also recovered, including large quantities of glass and metal fragments.

Table 5. Hydration-Rind Dates from Site MB-3

Provenience Depth (cm) Date (A.D.) t a

NE42 38-40 1725231 - 9: lowest sector of cultural deposit THE TEMPORAL SEQUENCE: TENTATIVE CONCLUSIONS The establishment of an absolute chronology, correlated with archaeological data, provides a framework for a preliminary discussion of prehistoric subsistence, residence patterns, and cultural change. Earliest settlement in the Hawaiian Islands has been characterized as permanent habitation restricted to choice locations focal to several microenvironments. It has been conjectured that these small permanent settlements were involved in a wide range of economic strategies. Wai-'Ahukini Shelter site B-21-6 (H8), with an early radiocarbon date of c. A.D. 750, is within the period characterized as "early." But, in all probability, this date only represents initial sporadic occupation in the South Point area. The primary economic motive was exploitation of rich marine resources, most likely, no horticultural activities were undertaken within this earliest period. The period of initial permanent or semi-permanent settlement within the area seems to have begun c. A.D. 900. Archaeological data clearly provide evidence for maritime exploitation throughout the South Point area sequence. From c. A.D. 900 to c. A.D. 1600, horticultural activities were marginal in nature>k, possibly such as those mentioned by Handy and Randy (1972 : 562) : Formerly in this whole seaward area .... there were coconut trees, taro, and sweet potatoe and gourd gardens maintained by the fishing families that dwelt there. Conjecturally, early inhabitants at Mahana Bay were involved in maritime exploitation associated with marginal supple- mentary horticultural endeavor. From the early to the late prehistoric periods, the subsistence pattern at Mahana Bay seems to shift from a maritime to a terrestrial resource emphasis. This is the predominant pattern among several Polynesian settlements within physically diversified insular environments (Yen 1971; Tuggle and Griffin 1973; Kirch 1973b). The earliest evidence for agricultural activities and related animal husbandry associated with the Mahana Bay sites is A.D. 1649+24, i.e. using pig bones as indirect evidence of agriculture. Under the conditions of intrinsic population growth, more sites (including those in coastal areas) should be * Based on Mahana Bay data, synthesis of H-1 horticultural associations is not possible at this time; a complete report of excavation of H-1 is not yet available. occupied in the late prehistoric periods. Although at Mahana Bay this may not be the case. By about A.D. 1600 some sites may have been abandoned as places of semi-permanent habitation, with only sporadic or seasonal occupation continuing. This abandonment, in all probability, could have resulted from dispersed residency to the upland areas of intensive cultivation and animal husbandry. Groups which continued to occupy coastal sites into the late prehistoric period had probably established trade relations with upland settlements. .The ethnographic record suggests trade relations with the same non-unilineal descent group ('8hana). The dispersed residency trend is documented not only at Mahana Bay, but also with the absolute date of A.D. 1523'15 (the earliest date currently known) from a habitation site in Wai-'6hinu Valley (Kelly and Crozier, 1972). In addition, the ethnohistorical accounts generally describe Wai-'5hinu Valley as the locus of dense population and intensive cultivation. It is not argued, however, that Wai-'5hinu is the specific location of agricultural associations for the South Point area, these data only serve to exemplify the model presented. The plains of Kama'oa were, in all probability, under dryland cultivati.on, as were several upland slopes where temperature, soil, and precipitation allowed. Site ME-3 (A.D. 1680'41 to c. A.D. 1900) was probably the permanent residence of a local high-ranking individual. The predominance of pig bones recovered from most levels throughout the site suggest economic differentiation and, by inference political status, or social ranking. Artifactual material supports this conclusion; e.g. a lei niho palaoa pendant (made of pig-tooth) was recovered. In light of the data presented, three phenomena, subsequent to c. A.D. 1600, are correlated for Mahana Bay: 1. Advanced cultural adaptations to the terrestrial ecosystem with increasing emphasis on intensive cultivation and animal husbandry systems. 2. The development of economic and social differentiation, 3. Dispersed residency from maritime to terrestrial ecosystem according to specialization of labor, i.e. a portion of the population shifting from coastal to upland habitation. This postulated model seeks only to synthesize the prehistoric sequence of Southern Ka'6, specifically Mahana Ray; it is not an attempt at an all-embracing scheme for the Hawaiian Islands. As Kirch (1973a) reminds us, it is only through increased low-level synthesis that we will be able to discern the more general trends of cultural development in the Hawaiian archipelago. Acknowledgements

It is indeed a pleasure to thank those individuals who made this project possible. I would especially like to thank Dr. H. David Tuggle for initial guidance and encouragement. My sincere thanks also go to Dr. Joseph B. Halbig and Mr. Larry Olson whose expertise allowed my work to commenceand toMr. William Bonk who kindly allowed access to archaeological material. I am also sincerely grateful to Drs. Patrick V. Kirch, Patrick C. McCoy, and Yosihiko H. Sinoto for their valuable suggestions, guidance, and encouragement. I also wish to thank Dr. James F. Downs, Ms. Sabra W. Downs, Dr. Carol A. B. Link, and Dr. Michael McCrae for their valuable suggestions. In addition, I wish to acknowledge the excavation of site H-1 carried out under the auspices of the Bernice Pauahi Bishop Museum. References

Barrera, William, Jr. 1971. Anaehoomalu: a Hawaiian oasis. Preliminary report of salvage research in South Kohala, Hawaii. Pacific Anthropological Records, no. 15. B. P. Bishop Mus., Honolulu. Barrera, William, Jr. and Patrick Vinton Kirch. 1973. "Basaltic-glass artifacts from Hawaii: their dating and prehistoric uses." J. Polynesian Soc. 82(2) Bellwood, P. 1975. "The Prehistory of Oceania." Current Anthropology. 16 (1) Cox, J. Halley. 1967. "The Lei Niho Palaoa." In Highland et. al. (eds.). Polynesian Culture History. B. P. Bishop Mus. Spec. Pub. 56. Honolulu. Emory, Kenneth P., Y. H. Sinoto. 1969. the South Point Area, Ka'u, Hawaii. pological Records 8, B. P. Bishop Mus. Honolulu. Emory, Kenneth P., Y. H. Sinoto. 1961. Hawaiian Archaeology: Oahu Excavations. B. P. Bishop Mus. Spec. Pub. 49. ~uUl.u. Emory, Kenneth P., W. J. Bonk, and Y. H. Sinoto. 1968. --Hawaiian Archaeology: Fishhooks. B. P. Bishop Mus. Spec .-. 47. fIonolulu. Friedman, I., R. L. Smith and W. D. Long. 1966. "Hydration of Natural Glass and Formation of Perlite." Bulletin --of Geol&cal - Soc. of America, 77. Green, Roger C. 1962. "Obsidian, its Application to Archae- ology." New Zealand Archaeological Assn. Newsletter, 5 (1). 1964. "Sources, Ages and Exploitation of New Zealand Obsidian." New Zealand Archaeological Assn. Newsletter, 7(3). --- (ed.) 1969. Makaha Valley Historical Project: Interim Report No. 1. Pacific Anthropological Records, 4. B. P. Bishop Mus. Honolulu. (ed.) 1971. Makaha Valley Historical Project: Interim Report No. 2. Pacific Anthropological Records, 10. B. P. Bishop Mus. Honolulu. 1971. "The Chronology and Age of Sites at South Point, Hawaii." Archaeology and Physical Anthropology in Oceania 6(2). Handy, E. S., Craighill and Elizabeth Green Handy. 1972. Native Planters in Old Hawaii, Their Life, Lore and Environment. B. P. Bishop Mus. Bulletin 233. Honolulu. Hunt, T. L., W. J. Bonk, and J. B. Halbig. 1976. "Preliminary Results of Hydration-rind Dating of Basaltic Glass Artifacts from Archaeological Sites in the South Point Area, Island of Hawaii." Paper read at The Big Island Science Conference, Hilo, Hawaii. Kirch, Patrick Vinton. 1971a. "Halawa Valley Site (Hawaiian Islands): A Preliminary Report." J. Polynesian Soc. 80 (2) . 1971b. "The Halawa Valley Project: Two Field Seasons in Retrospect." New Zealand Archaeological Assn. Newsletter, 14(2). 1971~. "Archaeological Excavations at Palauea, Southeast Maui, Hawaiian Islands." Archaeolovy and Physical Anthropology in Oceania, 6(1). 1973a. Archaeological Excavations at KahaluTu, North Kona, Island of Hawaii. Report 73-1 B. P. Bishop Mus. Honolulu. 1973b. "Prehistoric Subsistence Patterns in the Northern Marquesas Islands, French Polynesia.11 Archaeology and Physical Anthropology in Oceania, 8(1). 1974. "The Chronology of Early Hawaiian Settlement." Archaeology and Physical Anthropology in Oceania, 9 (2) . Kirch. Patrick Vinton and Marion Kelly (eds). 1975. ~irchin Prehistory and Ecology in a Windward Hawaiian Valley: Halawa Valley Molokai. Pacific Anthro~olo~icalRecords, 24 B. P. Bishop Mus.

Kelly, Marion and S. Neal Crozier. 1972. Archaeological Survey and Excavations at Waiohinu Drainage Improve- ment Project, Ka'u, Island of Hawaii. Report 72-6, B. P. Bishop Mus. Honolulu. Morgenstein, Maury and Thomas J. Riley. 1974. "Hydration- Rind Dating of Basaltic Glass: A New Method for Archaeological Chronologies." Asian Perspectives, 17(2). Pearson, Richard, P. V. Kirch, and M. Pietrusewsky. 1971. "An Early Prehistoric Site at Bellows Beach, Waimanalo, Oahu, Hawaiian Islands." Archaeology and Physical Anthropology in Oceania, 673). Sinoto, Yosihiko H. 1967. "Artifacts from Excavated Sites in the Hawaiian Marquesas, and Society Islands." In G. A. Highland (ed.) Polynesian Culture History, B. P. Bishop Mus. Spec. Pub. 56. Honolulu. Sinoto. Yosihiko H. and Patrick C. McCov. 1975. "Report on the Preliminary Excavations of an Early ~abitation Site, Huahine, Society Islands." -Journal de la Soci6t6 des Ocganistes, 31(47). Sinoto, Yosihiko H. and Marion Kelly. 1975. Archaeological and Historical Survey of Pakini-Nui and Pakini-Iki Coastal Sites; Waiahukini, Kailikii, and Hawena, Ka'u, Hawaii. Report 75-1 B. P. Bishop Mus. Honolulu. Tuggle, H. D. and P. Bion Griffin (eds). 1973. La akahi Hawaii: Archaeological Studies. Asian and *Paci ic Archaeological Series, 5. SSRI. University of Hawaii, H&olulu Yen, D. E. 1971. "The Development of Agriculture in Oceania.I, In Green and Kelly (eds.) Studies in Oceanic Culture I-Iistror . Vol. 2, pp.1-12 Pacific Anthropological Records, dP. Bishop Mus. Honolulu. THE INFLUENCE OF FERAL PIGS ON A NATIVE ALPINE GRASSLAND IN HALEAKALA NATIONAL PARK James D. Jacobi Department of Botany University of Hawaii at Manoa Honolulu, Hawaii 96822

During this meeting there has been considerable discussion concerning the impact feral pigs have on the Hawaiian biota, and what steps may be taken to control their influence. Where pigs are present, they do pose a serious threat to the stability of the native ecosystems. This threat is both in terms of physical damage to the vegetation by their rooting, and secondly because areas which have been uprooted by pigs may be invaded by vigorously competing exotic species of plants which could subsequently become firmly established in the community. In this paper I will discuss a study I have been conducting for the past three years in Haleakala National Park to determine the impact feral pigs have on the native alpine grassland of Ka-lapa-wili Ridge. This work was initiated during the summer of 1973 when members of the HZna Rain Forest Project spent several months studying various aspects of the ecology of the outer northeast slopes of Hale-a-ka-la. Since that time, additional support has come from both Haleakala National Park and the Cooperative National Park Resources Studies Unit, University of Hawaii (CPSUIUH) to continue this work. Three objectives are being addressed in this study: 1) to quantitatively assess the degree of pig damage in this community; 2) to determine the role pigs play in the introduction and establishment of exotic plant species in this grassland; and 3) to determine to what extent the native vegetation may be expected to recover if feral pigs were entirely removed from this area. This is an ongoing project which will be continued for several more years. The purpose of this paper is to briefly describe the study area and then summarize the results of the study to date. The Ka-lapa-wili Ridge grassland is located on the outer northeast slope of Hale-a-ka-18 volcano, on the island of Maui. It is the most extensive native alpine grassland found in Hawai'i, encompassing an area of approximately 120 ha, between the elevations of 7,400 and 8,000 ft (2255 and 2438 m) Its eastern and southern boundaries are sharply delineated by KT-pahulu Valley and the Hale-a-ka-lE summit depression respectively. To the north and the west however, the grassland diffuses into an alpine scrub habitat. Below 7,000 ft (2134 m), the scrub community gives way to a dense native rainforest, which then continues downslope to at least 1,600 ft (488 m). This grassland is an important habitat for nGnG geese, (Branta sandwicensis), which have recently been introduced to Maui by the State of Hawaii Division of Fish and Game and Haleakala National Park. Eighteen species of vascular plants are found in the grassland, fourteen of which are native species, while only four are "exotic," or introduced species. Most dominant by far is the endemic bunchgrass, Deschampsia austraZis f. haleakalensis, which forms a very dense ground cover in all areas, except those uprooted by pigs. The other native species found here are generally of widespread distribution, but occur in low numbers. These include a unique form of the sedge Uncinia, the native rush LuzuZa hawaiiensis, Pteridium aquizinium, and a small species of 'Ghelo (Vaccinium berberidifolium). The larger shrubs: Stypkelia tameiameiae, Vacciniurn reticulaturn, Sadleria cyatheoides, and Hubus hawaiiensis grow in more well-drained sites, both in the gulches which run through the grassland, and on rock outcrops with shallow soil. Henrickson (1972) reports that the large Maui greensword (Argyroxiphium virescens) was once common along the upper portions of Ka-lapa-wili Ridge. This species is no longer found in the area, and Henrickson speculated that it was eliminated by feral pigs and goats. One small patch of silverswords (Argyroxiphium sandwicense) has been found growing in the upper portion of the grassland near the peak, Pshaku-pslaha. Another large patch is found at a lower elevation on the cliffs of Pu'u-'alaea. Three of the four exotic species present here are both wide-spread and abundant. These species are the grass Holcus Zanutus, and two herbaceous species: Hypochaeris radicata and Kumex acetocetla. Rumex and Hypochaeris are generally found growing in open situations, particularly where the grass cover has been disturbed by pigs. HoZcus is found growing in small dense patches, but also invades the opened areas. Both feral pigs and goats are found within the grassland. The greatest damage here however, is caused by the pigs, primarily because they uproot extensive areas of the vegetation while feeding. The pigs are not resident in the grassland, but come up in the late afternoon and evening from the forest and scrub communities below, to feed. Generally they feed singly or in small family groups, moving slowly through the area in search of suitable forage. Their two preferred foods here are presumably Hypochaeris leaves and roots, and rhizomes of the bracken fern (Pteridium aquilinium). The stomach of a large boar collected in the grassland in July 1973 contained almost exclusively Hypochaeris leaves. Assessment of Plant Species Composition and Pig Damage In the Grassland The first phase of this study was to determine the distribution and abundance of both the plant species, and of pig rooting, throughout the grassland. For this purpose, three transects, each approximately 1000 m long, were laid out through the area. Two- by two-meter plots were sampled at 20 m intervals along each transect. In these plots, the cover for each species was estimated. Also recorded was the precent of the ground area which had been recognizably uprooted by pigs. Pig rooting was distinguished in three categories corresponding to the age of the disturbance. "Fresh" rooting was defined as ground which had obviously been very recently uprooted, the overturned vegetation still green; "recent" rooting: the ground still noticeably disturbed, but the litter all dead; and "old" rooting: most of the upturned litter having disappeared from the ground surface, some invasion of plant seedlings noted. Of the total area sampled, a mean of 23.3 percent was found to have been recognizably uprooted by pigs. Fourteen percent of this total was in the categories t, recent" or "fresh." Given this amount of disturbance, there is little doubt that the pigs do have a major impact on this ecosystem. Each transect was also analyzed in three segments: the upper, middle, and lower thirds, corresponding to an elevational gradient through the grassland. Here it was seen that a majority of the total rooting was found in the lower portion. This is most likely due to the fact that this is the area first encountered by pigs coming up from the lower forests. Of the plant species found here, only Deschampsia, HoZcus, and Hypochaeris showed cover values of greater than 5 percent. Deschampsia was by far the most dominant species with a mean cover value of 57.8 percent. However, the other native species were found to cover less than 1 percent, giving a total of less than 60 percent cover for all of the native species in the grassland. This is compared with 18 percent for the three exotic species: Hotcus, Hypochaeris, and Rumez . --The Exclosure Study The ultimate objective of this study is to determine what changes might occur in the vegetation if pigs were entirely removed from the area. The results of this phase of the study are still preliminary, and no conclusions have as yet been drawn. After two years of sampling however, some trends have developed, which are discussed below. A 10 x 40 m fenced exclosure was constructed in a representative portion of the grassland to exclude pigs. By periodically resampling the vegetation both inside and outside the fence, changes without the presence of pigs can be compared to changes in the vegetation with the continued presence of pigs. This same technique has been used to monitor changes in grassland vegetation in Hawaii Volcanoes National Park on the island of Hawai'i, following exclusion of feral goats and pigs (Mueller-Dombois and Spatz 1972, Spatz and Mueller-Dombois 1972). The Ka-lapa-wili exclosure is located below Pohaku-palaha at an elevation of 7,900 ft (2408 m). It was built in February 1974 with materials supplied by Haleakala National Park. The area enclosed by the fence includes a section which was recently heavily uprooted by pigs and an area relatively undisturbed when it was constructed. The exclosure is revisited, and the vegetation resampled, approximately every six months. Sampling is conducted using the point-frequency technique, which employs an aluminum frame one meter long, holding five upright, pointed steel rods, spaced 20 cm apart. The rods are lowered to the vegetation; whatever is first touched by the tip of the rod is recorded for that sample point. By this means, the vegetation can be characterized in terms of percent cover by the number of points intercepted by each species. Sampling is conducted along ten transects laid out through the exclosure. Each transect is twenty meters long, of which ten meters is inside the fence, and ten meters is outside the fence (five meters on either side).

The exclosure has been sampled five times: in March and October 1974, March and October 1975, and in March 1976. TO date, no significant difference has been found in the vegetation inside and outside the fence. Such differences should start to become more apparent as the study continues. The most interesting aspect of this study at this point however, is following the revegetation of the areas inside the fence which had been heavily uprooted when the exclosure was constructed. These areas are covered by Transects 1-4. Large changes in cover values were found for all sample categories along Transects 1-4 in the first year (1974-1975). Much smaller changes were found in the vegetation along Transects 5-10 during the same time. In the first four transects, the largest cover change came in the "litter" and "barren soil" categories. These categories reflect the amount of uprooting by pigs at the beginning of the study. The amount of litter and barren soil decreased as the rooted areas were invaded by various plant species. The primary invaders in this case were the native grass Deschampsia, and the three exotic species: HoZcus, Hypochaeris, and Xumex. In the second year, the overall changes in cover for the different species was less than the first year along all ten transects. The smaller changes indicate a somewhat more stable situation at this time. The herbaceous species reached their greatest abundance during the first year along the disturbed transects (1-4). In the second year however, the cover values for each of these species decreased as a whole throughout the exclosure, presumably as they were crowded- out by the denser-growing grass species. Over the two year sampling period, the native grass Descharnpsia showed an increase in cover on seven of the ten transects, while the exotic species Hotcus increased on six of the ten. Both species increased on the disturbed transects, indicating their ability to invade pig-disturbed areas.

Holeus showed an increase of at least 10 percent on Transects 1, 3, 5, and 9, and an 8 percent increase on Transect 8. On the other hand, there were small decreases in its cover on Transects 6, 7, and 10. The question now is, will HoZcus continue to increase and eventually maintain a stable position in this community, or will it eventually be crowded out by the native species, in particular by Deschampsia? If Hotcus can increase to the point of displacing Deschampsia, it will remain a serious threat to the native grassland, even after the pigs have been removed. Summary In summary, considerable portions of the Ka-lapa-wili Ridge grassland were found to be heavily uprooted by feral pigs. To put this disturbance into vivid perspective, if the three sampling transects through the grassland give a reasonable description of situation there, then 23.3 percent of the entire grassland, or 24 of the total of 120 hectares of the grassland were uprooted at that time. These pig-disturbed areas are initially invaded by three exotic species: HoZcus, Rurnex, and Hypochaeris; and the native Deschampsia. None of the other native species were found to be active colonizers of these situations. The herbs are abundant until the grass cover gets so dense that they are crowded out. The big question now is whether the native species are able to recover the areas in which Holcus has become established, or if the exotic species will be maintained in the grassland. Hopefully this question will be answered in the next few years. In the meantime, based on the high degree to which the grassland is being disturbed by pigs, I would recommend that a rigorous pig control plan be immediately put into effect to reduce and eventually eliminate the influence of feral pigs from this area.

Literature Cited

Henrickson, J. 1972. Vascular flora of the northeast outer slopes of Haleakala Crater, East Maui, Hawaii Contributions, Nature Conservancy, No. 7. Mueller-Dombois, D., and G. Spatz. 1972. The influence of feral goats on the lowland vegetation in Hawaii Volcanoes National Park. Phytocoenologia 3(1):1-29 Spatz, G., and D. Mueller-Dombois. 1972. Succession patterns after pig digging in grassland communities on Mauna Loa, Hawaii. Phytocoenologia 3(2/3):346-373. TOPOCLIMATIC GRADIENTS IN HAWAII VOLCANOES NATIONAL PARK^,' J. 0. Juvik and G. G. Clarke Department of Geography University of Hawaii at Hilo Hilo, Hawaii 96720

Hawaii Volcanoes National Park, occupying the southeastern slopes of Mauna Loa, encompasses an altitudinal range extending from sealevel to a summit elevation of 4169 m. The area is characterized by a complex local climatology, which results from the interaction of a large-scale marine trade wind environment with a variety of localized climatic controls generated by the mountain itself. Climatic parameters were monitored along an altitudinal gradient with the objective of developing a general model to characterize the energy and water balance climatology of the windward slope. To the extent that climatic factors exert a strong control over the development and differentiation of biotic communities, the study provides useful data for interpreting the spatial organization of these communities, and the development of distinctive altitudinal eco-zones on the mountain. Data collection included the monitoring (over a 2 year period) of diurnal wind patterns, evapo-transpiration, precipitation (including canopy throughfall and fog drip), leaf wetness, and soil moisture. Integration of the climatic data has provided a quantitative base for preliminary assessment of the energy and water balance gradients on windward Mauna Loa. The results also facilitate an evaluation of the role of fog drip and dew formation in the overall water economy, as well as the importance of subtle topographic features in intensifying mid-altitude summer drought.

'Abstract '~esearch supported in part by a grant from the Hawaii Natural History Association. EVOLUTIONARY STUDIES OF TWO DROSOPHILA SPECIES FROM HAWAII: A MULTIDISCIPLINARY APPROACH Kenneth Y. Kaneshiro Department of Entomology University of Hawaii at Manoa Honolulu, Hawaii 96822

Various techniques of evolutionary biology are being utilized in an investigation of the phylogenetic relationships of two sympatric Drosophila species from the Island of Hawai'i and their counterparts on Maui Nui. The two species from Hawai'i are Drosophila heteroneura and D. silvestris while D. planitibia is from Maui and D. differens is from Moloka'i. Patterns of speciation and the evolutionary history of the four species in relation to one another can be inferred by pooling the data from various studies including morphology, hybridization experiments, ethology, cytology, electrophoresis, and ecology. Preliminary analyses of these data indicated that a single introduction into the Big Island of Eawai'i resulted in the speciation of both D. heteroneura and D. silvestris. However, further investigation into the morphology, electrophoresis, and behavior of the four species revealed a possibility that two separate introductions may have been involved in the speciation of the two Big Island species. Back~round The chromosomal relationships of all four species have been reported by Carson and Stalker (1968) and Craddock (1974). The two Maui Nui species are monomorphic and homosequential in the banding patterns of polytene chromosomes while D. heteroneura and D. sihestris share a unique polymorphic inversion in Chromosome 111 (3m/+). In addition, D. silvestris carries six other polymorphic inversions not found in the other three species. Thus, there are no fixed inversion differences in the karyotype of the four species. Equally intriguing results have been obtained by measurements of genetic similarity based on allozymic variation (Johnson et al. 1975). Calculations of similarity indices among the four species revealed an unusually high level of genetic similarity ranging from 0.71 - 0.96. Ayala et al. (1974) described anobservedrelationship between genetic similarity indices and the categorization of taxa at and below the species level. If identity is taken at 1.0, local geographic populations of a species are shown to have mean similarity values of 0.97, subspecies and semispecies about 0.8, sibling species 0.52, and closely related but morphologically differentiated species 0.35. Despite striking morphological differences between D. heteroneura and D. silvestris, the similarity value of 0.96 reported by Johnson et al. (1975) indicate an extremely close phylogenetic affinity between these two species. Various programs of interspecific hybridization between these four species were undertaken to test the existence of postmating isolation barriers (e.g. hybrid inviability, sterility, etc.). Crosses involving D. pZanitib$.a and the other three species produced fertile female hybrids but sterile males. However, reciprocal crosses between the sympatric species D. heteroneura and D. silvestris produced vigorously fertile F hybrid progeny of both sexes (Craddock, 1974 and Ahearn an Ii Val, 1975). Furthermore, there appeared to be no obvious breakdown in hybrid fertility in the F2 and back- cross generations. Preliminary data on crosses involving D. differens and the two Big Island species indicafe that F1 hybrids are also fully fertile in both sexes. Observations of the courtship behavior patterns of D. heteroneura and D. sitvestris (Spieth, personal communication) indicate that there are no qualitative differences in the courtship behavior of these two species. However, Ahearn et al. (1974) and Kaneshiro (in press) report that sexual (behavioral) isolation between the two sympatric species is nearly complete which indicates that there may be quantitative differences in their courtship patterns. Ecologically, Heed (1968) and Montgomery (1975) report that all four species utilize the decaying bark of CZerrnontia spp. (Lobeliaceae) as the primary substrate for the larval breeding site. In a recent collection, separate rotting branches of Clerniontia yielded nearly equal proportions of both D. heteroneura and D. sitvestris adults. Thus, despite the possibility of differences in larval nutritional requirements, ovipositional behavior of gravid females of both species is apparently stimulated by the rotting bark of Ctermontia. Phylogenetic Interpretation By assessing the data accumulated from the various studies discussed above, Craddock (1974) proposed three possible patterns of speciation in the evolution of D. heteroneura and D. sitvestris. Nevertheless, on the basis of the shared polymorphic inversion in Chromosome 111, she favored the hypothesis that the two sympatric species arose from a single founder event from one of the Maui Nui populations. Since neither D. differens from Moloka'i nor D. ptanitibia from Maui carries the 3m inversion in their populations, derivationof the two Big Island species from two independent founders was not a parsimonious interpretation. However, subsequent morphological comparisons revealed an interesting phenotypic resemblance between D. differens and D. heteroneura and between D. plantibia and D. situestris which provided conflicting evidence for Craddock's single introduction hypothesis. The most obvious similarity between D. planitibia and D. siZvestris is the velvety-black coloration of the face while those of D. heteroneura and D. differens are light brown to bright yellow. The marking in the costal margin of the wing also appears to relate D. pZanitibia with D. siZvestris and D. differens with D. heteroneura. Zanitibia and D. silvestris, this costal marking usually tapersIn D.zan disappears at a out the level of the posteriorcrossveinwhile in the latter pair, the marking usually extends beyond the level of the posterior crossvein to the markings near the apex of the wing. On the basis of these phenotypic criteria, a second possible interpretation of the origin of the two Hawai'i species must be seriously considered. That is, two independent founders, one from Moloka'i (D. differens as progenitor) and one from Maui (D. pzantibia as progenitor), may have given rise to D. heteroneura and D. silvestris respectively. Further analyses of the behavioral relationships among the four species by Kaneshiro (in press) appear to support the "two-founder" hypothesis. Preliminary observations (F. Sene, personal communication) of one of the Esterase locus as determined by starch-gel electrophoresis indicate that D. heteroneura may be more closely related to D. differens than to D. silvestris or D. planitibia and conversely, D. si2vestris may be more closely related to D. pZanitibia than toD. heteroneura or D. differens. These studies are continuing. The shared polymorphic inversion in Chromosome I11 of the two Hawai'i species poses a major stumbling block to the two-founder hypothesis. The origin of an identical chromosomal polymorphism from two independent founders is highly improbable. Alternatively, the shared inversion may be explained by interspecific hybridization and a limited amount of genetic introgression between the two populations short of complete gene exchange. Natural hybridization between a pair of "good" biological species is, at best, a rare phenomenon. However, the discovery of fertile interspecific hybrids of D. heteroneura and D. silvestris (Kaneshiro and Val, in press) in the Kahuku Ranch area (S. Kona) permits invoking natural hybridization as a very possible mechanism by which the two species now share a polymorphic inversion. Thus, inferences of phylogenetic relationships and evolutionary patterns of speciation is made possible by analyses of data from these various tools of evolutionary interpretation. We are continuing such studies by comparing the natural history of three sympatric populations of D. heteroneura and D. silvestris. One such population occurs at Pau-ahi, Kona,on the Greenwell Ranch where both species coexist with no evidence of interspecific hybridization. Another population being scrutinized occurs at Ka-huku Ranch, S. Kona,where approximately 2% of the specimens collected thus far showed morphological properties reflecting inheri- tance from both species. The third population occurs in Gla'a Tract within the Hawaii Volcanoes National Park. Here, both species coexist with some evidence that genetic introgression may be occurring although not a single bona fide hybrid specimen has been collected. In the Pau-ahi and Ka-huku Ranch populations of D. silvestris, the abdominal tergites are jet-black while D. heteroneura displays a distinctive black-yellow contrasting pattern. However, in the 61a'a Tract population, the black-yellow pattern of D. heteroneura appear to be segregating in the D. silvestris population. It is hoped that comparisons of the behavioral, morphological, ecological, and electrophoretic analyses of these three populations will enable us to better understand the genetic mechanisms involved in species divergence (e.g. "character displacement"). These studies will make possible inferences of phylogenetic relationships among four species of Hawaiian Drosophila. More important, it is hoped that these basic concepts of evolutionary theory can be applied to the understanding of the evolutionary strategies of other organisms in the Hawaiian biota. Literature Cited

Ahearn, J.N., H.L. Carson, T. Dobzhansky, and K.Y. Kaneshiro. 1974. Ethological isolation among three species of the pZanCtibia subgroup of Hawaiian DrosophiZa. Proc. Nat. Acad. Sci. 71:901-903. Ahearn, J.N., and F.C. Val. 1975. Fertile interspecific hybrids of two sympatric Hawaiian Drosophila. Genetics 80:s9. (Abstr.) Ayala, F.J., M.L. Tracey, D. Hedgecock, and R.C. Richmond. 1974. Genetic differentiation during the speciation process in Drosophiia. Evolution 28:576-592. Carson, H.L., and H.D. Stalker. 1968. Polytene chromosome relationships in Hawaiian species of DrosophiZa. 11. The D. pZanitibia subgroup. Univ. Texas Publ. 6818:355-366. Craddock, E.M. 1974. Reproductive relationships between homosequential species of Hawaiian DrosophiZa. Evolution 28:593-606. Heed, W.B. 1968. Ecology of the Hawaiian Drosophilidae Univ. Texas Publ. 6818:387-419. Johnson, W.E., H.L. Carson, K.Y. Kaneshiro, W.W.M. Steiner, and M.M. Cooper. 1975. Genetic variation in Hawaiian Drosophila. 11. Allozymic differentiation in the D. planitibia subgroup. Pages 563-584 in C.L. Markert, ed. Isozymes IV genetics and evolution. Academic Press Inc., New York. Kaneshiro, K.Y. 1976. Ethological isolation and phylogeny in the planitibia subgroup of Hawaiian Drosophi Za. Evolution (in press). Kaneshiro, K.Y., and F.C. Val. 1976. Natural hybridization between a sympatric pair of Hawaiian Drosophila. Am. Naturalist (in press). Montgomery, S.L. 1975. Comparative breeding site ecology and the adaptive radiation of picture-winged Drosophila. Proc. Haw. Ento. Soc. 22:65-102. MANAGEMENT OF SCIENCE RESERVE AREASJC John I. Kjargaard Haleakala National Park Makawao, Maui, Hawaii 96768

I will present-some current views on the establishment and management of Ki-pahulu Valley and the HHna Rain Forest Science Reserves from the point of view of the visitor, the scientist, and the park. Also covered will be a proposal to better unite the objectives of these three groups so that improved working relationships may be established.

"Abstract ENDANGERED PLANTS IN HAWAII VOLCANOES NATIONAL PARK C. H. Lamoureux Botany Department University of Hawaii at Manoa Honolulu, Hawaii 96822

During the summer of 1976 a field survey of Hawaii Volcanoes National Park was initiated to determine which plants on various lists of endangered species were present in the park. All specimens in the park herbarium were also examined as part of this study. In the past two years a number of lists have appeared proposing candidates for inclusion on a list of endangered taxa which would be entitled to legal protection of some sort. The most comprehensive are: (a) "Rare and endangered species of Hawaiian plants" by F. R. Fosberg and Derral Herbst, Allertonia 1:l-72, 1975. (b) a list in the Federal Register, 40 FR 27823-27924, July 1, 1975. (c) a list included in a letter from Governor George R. Ariyoshi to F. Eugene Hester, August 22, 1975. (d) a list in the Federal Register, 41 FR 24524-24572, June 16, 1976. Since list (d) is the one currently being considered for adoption under the terms of the Endangered Species Act of 1973, the following interim list for HAVO is based on that list, although information for the park has been obtained for the taxa on all four lists.

The plants in this list are those which now grow in Hawaii Volcanoes National Park, or which have been collected in the past from sites now included within tWO, and which were listed in the Federal Register of June 16, 1976, pp. 24524-24572 as among those proposed for determination as endangered taxa pursuant to section 4 of the Endangered Species Act of 1973. The Federal Register list includes only taxa proposed for endangered status, it does not include any proposed for threatened status, so all plants listed below are to be considered endangered. Family Scientific Name Common Name

FERNS : Ophioglossaceae OphiogZossum concinnum Brack. pololei Polypodiaceae Adenophorus periens L.E. Bish.

MONOCOTYLEDONS:

Liliaceae Dracaena hawaiiensis (Deg.) hala-pepe Fosb . Poaceae Ischaemum byrone (Trin.) Hitch.

DICOTYLEDONS:

Apocynaceae Ochrosia compta K. Schum. RauvoZf ia remotif zora hao Deg. & Sherff

Araliaceae Reyno Zdsia hi ZZebrandii 'ohe m5ka1i Sherf f

Asteraceae 9cArgyroriphium kauense 'Hhinahina; (Rock & Neal) Deg. & Deg. Ka'E silversword

*A. sanndwicense DC. 'iihinahina; Hale-a-ka-18 silversword; Mauna Kea silversword

Bidens skottsbergii Sherff kobko'olau var. skottsbergii

B. skottsbergii var. ko'oko'o lau conglutinata (Deg. & Sherff) Sherff

Campanulaceae Clermontia hawaiiensis (Hbd.) '8hFi-k5pau Rock var. hawaiiensis

C. peZeana Rock Fabaceae CanavaZia kauensis St. John

Sesbania tomentosa H. & A. ' 6hai * There is no evidence that these taxa occurred naturally in KAVO and present populations are those planted by man. 125

Family Scientific Name Common Name

DICOTYLEDONS :

Goodeniaceae ScaevoZa kiZaueae Deg. naupaka ; var. kiZaueae huahekili-uka

Loganiaceae Labordia baiZZonii St. John kzmakahala

L. hedyosmifotia Baill. kEmakahala var. kiZaueana Sherff

L. hedyosrnifoZia var. magnifotia Deg. & Sherff Malvaceae Hibiscade Zphus giffardianus hau-kuahiwi Rock

*H. huaZaZaiensis Rock hau-kuahiwi

*Kokia drynarioides (Seem.) hau-hele-'ula Lewt . koki'o Pittosporaceae *Pittosoorum hosmeri Rock 'a'awa- hua-kukui var. saint-johnii Sherff

Portulacaceae Portutaca scterocarpa 'ihi-mzkole; Gray po'e

Rhamnaceae Alphitonia ponderosa Hbd. kauila

Rubiaceae Bobea timonioides ' ahakea (Hook. f . ) Hbd . Rutaceae Pe Zea zahZbruckneri Rock Zanthoxy Zum dipetaturn Mann var. geminicarpum Rock Sapindaceae Dodonaea eriocarpa Smith var. skottsbergii Sherff

Solanaceae Nothoeestrum breviflorum 'aiea Gray

N. ZongifoZium Gray var. 'aiea rufipitosum Stone Urticaceae Reraudia ovata Gaud.

* There is no evidence that these taxa occurred naturally in HAVO and present populations are those planted by man. PHENOLOGICAL STUDIES IN HAWAII VOLCANOES NATIONAL PARK* C. H. Lamoureux Department of Botany University of Hawaii at Manoa Honolulu, Hawaii 96822

Studies of phenology of 85 plant species have been made from 1971 through 1975, in a number of ecosystems between 8 and 2040 m altitude on the island of Hawai'i, mostly within Hawaii Volcanoes National Park. Most species exhibit pronounced annual peaks for such phenophases as leaf production, flowering, and fruit ripening. Usually, however, not all individuals in a population are in close synchrony, and careful search is likely to reveal a few individuals in any phenophase at any month of the year. In some species different branches of the same tree exhibit different phenophases at the same time. Growth rates of woody stems vary seasonally, but some diameter growth occurs throughout the year. Phenology of several species will be discussed as examples of different uatterns of seasonal activity in onlyLslightly seasonal' climates.

"Abstract CURRENT STUDIES OF KT-LAU-EA AND MAUNA KEA VOLCANOES BY THE HAWAIIAN VOLCANO OBSERVATORY OF THE U.S. GEOLOGICAL SURVEY* P. W. Lipman, L. A. Anderson, R. Y. Koyanagi, and J. P. Lockwood Hawaiian Volcano Observatory U.S. Geological Survey Volcano, Hawaii 96718

The prime responsibility of the Hawaiian Volcano Observatory (HVO) of the U.S. Geological Survey is to advise the State and County of Hawaii and the National Park Service of potential volcanic and other natural hazards on the Island of Hawai'i. HVO, with a permanent staff of 17 scientists and technicians, also carries on a diversified program of basic scientific research on seismicity, eruptive processes, and cooling and crystallization of basalt. Over the last 25 years, attention has focused on Ki-lau-eavolcano, but since its reawakening in 1974, Mauna Loa Volcano has occupied an increasing proportion of HVO's efforts. In addition to volcanic hazards, the 1975 Kala-panaM=7.2 earthquake and related tsunami and catastrophic ground movements provided a recent reminder of other natural hazards of recurrent concern. Readiness of a Hawaiian volcano to erupt is evaluated primarily on the basis of seismic ac-tivity and changes in shape (inflation) of the volcano. Ki-lau-ea andMauna Loa inflate gradually prior to an eruption, as magma accumulates in shallow reservoirs within the volcanoes; abrupt deflation accompanies and follows many eruptions, especially flank activity. Primary monitoring efforts include a 37-station island-wide seismic net to study patterns of earthquakes related to volcanism, a trilateration net of several hundred lines to detect horizontal-distance changes, more than 100 km of level lines to measure changes in elevation, approximately 70 tilt stations to detect changes in slope, and electrical surveys to detect conductivity changes caused by the subsurface movement of magma within the volcano. As of early June, 1976, KT-lau-ea is relatively quiet at a low degree of inflation; it is only beginning to reestablish a stable pattern of inflation after the large deflation and collapse associated with the November, 1975, earthquake. No KT-lau-ea eruption seems likely in the immediate future. In contrast, Mauna Loa has been inflating steadily since the July, 1975, summit eruption--the first eruption of this volcano in 25 years. This pattern of deformation andthe historic pattern of Mauna Loa eruptions, in which a flank outbreak typically follows a summit eruptionwithin about 3 years, suggest that Mauna Loa will likely erupt again within a few years. The seismic and deformation data indicate that this eruption may occur along the northeast rift zone, from which lava flows can threaten Hilo, the principal population center on the island. PU'U-KOHOLA HEIAU NATIONAL HISTORIC SITE PLANT SURVEY James D. Macneil, Jr., Lisa K. Croft, Don E. Hemmes Biology Department University of Hawaii at Hilo Hilo, Hawaii 96720

High on a hill overlooking the Pacific Ocean, near the village of Ka-wai-hae on the island of Hawai'i, sits Pub-kohola Heiau, the last major religous structure of the ancient Hawaiian culture to be built in the islands. This Heiau dominates what is now the location of a National Historic Site. Three temple sites are found in the vicinity: the ruins of Pub-koholg Heiau, Maile-kini Heiau, and the traditional site of Hale-o-kapuni Heiau. The park was authorized by Congress on August 17, 1972. Other features of historical interest located within the park include Queen Emma's home and the housesite of Ka-mehameha the First's English confidante, John Young. Pdu-koholg Heiau, the dominating structure in the park, was built by Ka-mehameha in 1790 - 1791 on Pub-kohola, the "hill of the whale," and was dedicated to his family war god, Kii-E'ili-moku. Pub-koholB Heiau National Historic Site is located approximately one mile south of Ka-wai-hae at the base of the western slopes of the Kohala Mountains, on the leeward side of Hawai'i. The 77 acre site consists of a gently rolling landscape with thehighestelevation of 135 ft. and running down to sea level. It is bordered on the west by Samuel Spencer County Park and undeveloped savannah. The eastern boundary transects State Highway #27. T'nis arid coastal zone has the lowest annual precipitation in the State of Hawai'i. The lack of rain is in the most part due to the dramatic rain shadow produced by the Kohala Mountains which rise 5,500 ft. above sea level to the north and the north east, thus blocking the predominant tradewinds. Mauna Kea and HualHlai are situated to the southeast and south, respectively, and contribute in producing the arid environment. A definite rainy season occurs during the months of January and February when Kona storms converge over the area on the leeward slopes. The mean annual temperature at Pub-koholg is 760F., with the highest temperature recorded during the month of September. The plant survey of the area was conducted during the "dry" season, May 1975, and the "wet" season, after a heavy rain, in March 1976. A total of 53 species representing 21 flowering plant families were recorded along with representatives from the Pteridaceae and Ophioglossaceae, and Basidiomycetes (mushrooms, puffballs, etc.). Thirty- eight of the species collected were introduced, nine were native, and five were endemic. Eleven species were seen only during the wet season. Four distinct plant communities were recognized within the Pu'u-kohols Heiau National Historic Site. Aerial photographs show a sharp demarcation between the open xeric grassland surrounding the Pu'u-koholZi Heiau and the lower elevation coastal forest. To the north this forest forms a protective canopy over the third community: a brackish pond and an extensive salt marsh which extends from the back berm of the sandy beach several hundred yards inland. The roadside plants were recognized as the fourth vegetation community.

Pennisetum setosum, (Sw.) L. C. Rich (feathery pennisetum) and scattered Prosopsis paZZida HBK. (kiawe) predominate in the scrub grassland. Within this community there are also scattered populations of Desmanthus virgatus (L.) Willd. (slender mimosa), Pennisetum setaceum (Forsk.) Chiov. (fountain grass), Leucaena ZatisiZiqua (L.) Gillis (koa haole), Sida faZZax Walp. ('ilima), ipomoea congesta R. Br. (blue morning glory), and Merremia aegytia (L.) Urban (hairy merrmia). The habit and cover of these species varies greatly according to the time of the year. During the summer, vegetative die-back is extensive. However, during the few months of the "wet" season a dramatic greening takes place along with a burst of vegetative growth. The xerophytic Basidiomycete Battaurrea phaZloides Pers. and the endemic Jacquemontia sandwicensis Gray, though encountered rarely, were also collected in this region. The closed forest community can be divided into two subcommunities: a closed coastal forest of tall Prosopsis paZZida HBK. bordering the ocean from the south end of the Hale-o-kapuni Heiau Beach to the Spencer Beach Park. Because of its thick canopy only sparse populations of Pennisetum setosum (Sw.) L. C. Rich, Sida faZZax Walp., and AbutiZon grandifolium (Willd.) Sweet occur under the trees. A closed fluvial forest community borders the Makeahua Gulch and extends through the northern portions of the site. It also consists of a forest of mixed species, including Prosopsis paZlida HBK., Setaria uertici2Zata (L.) Beauv., Panicum maximum Jacq., and other introduced plants. Around the brackish pond and the beach area at the northern eastern end of the historic site there are scattered populations of AtripZex semibaccata R. Br. and Atriplez mueZZeri Benth. of relatively low stature. Scattered seedlings of Prosopsis paZZida HBK., and Thespesia popuzena (L.) Soland ex Correa grow not more than 25 feet from the sea on the beach during the winter months. Toward the berm of the beach stands of Ipomoea pes caprae R. Br. and ScaevoZa taccada (Gaertn.) Roxb. These immediately give way to marsh bordered on the south by a closed forest of well-established Prosopsis paZZida HBK. interspersed with tall representatives of Cocos nucifera L. and Messerschmidia argenta (L.F.) Johnston (heliotrope tree). The marsh area between the pond and the beach contains large populations of Sesuvium portutacastrum (L.) L. and Batis maritima L., as well as many introduced and one native species, HeZiotropium curassavicurn L. Near the brackish pond a young date palm (Phoenix canariensis Hort. ex Chadbaud) and Chenopodium oahuense (Meyen) Aellen were observed. Vigorously growing populations of AtripZex sernibaccata R. Br. and A. mueZZeri Benth. are found around the periphery of the pond and are more robust than those encountered on the beach. A community of vigorously growing species is also found along the roadways that transect the Historic Site. These species include Sida faZZax Walp, Amaranthus dubius Mart., Portulaca cyanosperma Egler, WaZtheria indica var. americana Voxb., and AbutiZon grandifotium Sweet. Several endemic species located in the Park are noteworthy. Chenopodium oahuense (Meyen) Aellen was encountered near the marsh area and consisted of two shrubs approximately seven feet tall. Jacquemontia sandwicensis Gray was observed near Park Headquarters, but is uncommon. PortuZaca cyanosperma Egler is relatively common along the roadside of the county road connecting State Highway 1/27 to the Samuel Spencer County Park road. Two endemic ferns, OphiogZossum concinnum Brack., which is considered rare and endangered, and Doryopteris decora Brack., are ephemeral species found growing in the shade of rocks and were encountered only during the wet season.

An undetermined species of AtripZex was discovered on the sand beach. Its habit and fruit structure are unlike any of the four AtripZex species described from the Hawaiian Islands. THE MAUNA KEA ADZ QUARRY COMPLEX, HAWAII: A FIRST ANALYSIS Patrick C. McCoy Bishop Museum Honolulu, Hawaii 96818

Abstract Distributed over an area of roughly 74 square miles, at the 8,600 t~ 12,400 ft. elevation on a landscape that today can be described as an alpine stony desert, the Mauna Kea Adz Quarry Complex is the largest known adz-making locality in Polynesia. The quarries assume further significance in their remoteness from coastal settlements and the inferred organization and preparatory arrangements needed to undertake work at a high altitude. Results of fieldwork conducted in the summers of 1975-76 at the Mauna Kea Adz Quarry Complex are described in relation to a set of primary objectives aimed at developing a technological model of Hawaiian adz manufacture. Sites are defined in terms of constellations of activity areas, using topography to delimit boundaries. The site distribution pattern is examined in terms of bedrock exposures of a single basalt flow and downslope glacial outwash deposits. Lithic, vegetal, and faunal remains from two stratified rock-shelters are briefly discussed in a temporal context of short term, intermittent adz production. INTRODUCTION The first systematic archaeological investigation of the Mauna Kea Adz Quarry, a National Historic Landmark, was undertaken by the Bernice P. Bishop Museum in the summers of 1975-76, comprising seven months of fieldwork. A detailed summary of the 1975 fieldwork is presented elsewhere (McCoy, in press) and the reader is referred to that paper for a fuller explication of the project's background, research design, field procedures, and preliminary results, including the first radiocarbon dates obtained for the quarry. At this stage of the project, with the field research just completed, it is possible to do little more than present a general description of the quarry--its location, environmental setting, and makeup (types and spatial relationships of activity components). Beyond this, the limited aims of this paper are to examine briefly the site distribution pattern in terms of differing levels of exploitation, and to discuss the implications of excavated material and deposits for posited intermittent, short-term, seasonal adz manufacture WSEARCB OBJECTIVES AND METHODOLOGY The primary objective of the research is the development of a technological model of Hawaiian adz manufacture that details manufacturing techniques, core-reduction processes, and the patterning of activities in the quarr . The Mauna Kea Adz Quarry is the largest, most complex, and t est-preserved adz-making locality in the Hawaiian Islands, and thus is most likely to yield the types and quantity of data needed to achieve this goal. The research is also addressing questions such as the possible cultural constraints influencing the demand for adzes, and the possibilities of trade in adzes outside the island of Kawai'i. The size of the quarry and the volume of rejected adz preforms and waste flakes has led to the formulation of a hypothesis that adz production on Mauna Kea exceeded local needs, thereby representing a commercial venture in a commodity that, probably, would have had a high trade value. It is expected that the completed research will also provide important new data on change, or the lack of it, in Hawaiian adz forms through time. Field investigations have included reconnaisance survey, intensive site survey and recording (detailed mapping and description), excavation, and surface collection. The broad- based organization of field activities reflects the ultimate goals of the project and paucity of knowledge about the quarry prior to 1975. Emphasis was placed on recording and sampling of surface remains over the entire quarry area in order to document the range and possible areal variability in adz manufacture, and to aid in developing a general preservation plan. Excavation of selected rock-shelters was equally important in recovering: (1) datable materials for developing a quarry chronology; (2) faunal and floral materials to interpret dietary patterns; and (3) artifacts utilized by the adz makers in maintaining themselves during the period of work. The program of excavation was expanded to include soil-profile trenches outside rock-shelters. Analysis of sediments from rock-shelter environments and open slopes is being undertaken as part of a preliminary effort at reconstruction of environmental conditions immediately prior to, during, and after the abandonment of the quarry. An integral Dart of the ~roiectis lithic experimentation, an approach That is becoming hc;easingly popular in the study of stone tool assemblages. The rationale for this goes beyond the satisfaction gained in successfully replicating an adz, or simple familiarization with the raw material--its inherent properties and constraints. Such experiments provide information on the mechanics of fracture and probable behavior of reducing and shaping a piece of material to an intended final form and size. Controlled experimentation will provide quantified data to formulate a model of manufacturing stages against which field data can be compared. LOCATION AND ENVIRONMENTAL SETTING The quarry is an area of roughly 7% square miles on the south slope of Mauna Kea (Fig. 1). The main activity was concentrated in a zone that is 1-to-1% miles wide between the 11,000 and 12,400 ft. elevation. Below 11,000 ft. activity was confined to two narrow strips along Wai-ka-halulu and Pohaku-loa gulches. The gulches are the effective east-west boundaries of the quarry at the lower elevation extremity. Adz manufacturing sites have been found down to 8,600 ft. at Liloe Spring on the west bank of Pohaku-loa Gulch. Archaeo- logical remains (shrines), believed to be related to the operation of the quarry, have been located as high as Lake Wai-au (13,020 ft . elevation) . A large part of the quarry is located on a broad smit plateau encircling the mountain between 11,000 and 12,000 ft. The landscape is dotted with numerous cinder cones, the principal one of which in the quarry area is Pu'u Ko'oko'olau. The upper slopes of Mauna Kea have been described as a stony alpine desert (Ugolini, 1974). There is little vegetation and the ground surface has the appearance of a desert pavement. The modern climate is both dry and cold. It is sufficiently cold that periglacial features, such as patterned ground, are actively maintained. Evidence of formerly colder conditions can be seen in the summit region in the form of glacially scoured bedrock and surficial glacial drift deposits. This brief summarization of location and environment is presented as background for a further consideration of the level of organization and environmental adaptation required of Hawaiians to undertake work at this remote locality. In prehistoric times the nearest permanent settlements on the Hamakua coast would have been at least 25 miles distant by foot. Because of the quarry's location, it would have been necessary to transport food and warmer clothing, remnants of which we have found in pieces of tapa cloth cloaks and possible ti leaf rain capes. These inferred preparatory arrangements and subjection to an alien and frequently inhospitable climate are testimony to the importance of adzes in the Hawaiian tool kit. There are numerous sources of adz-quality basalt on the island, but obviously the choice was made in favor of exploiting more fully an isolated locality instead of more accessible coastal ones. The determining factor was undoubtedly the quality and quantity of basalt at this one location on Mauna Kea. THE QUARRY COMPLEX: AN OVERVIEW The Mauna Kea Adz Quarry Complex is an unusually large complex of clustered and dispersed adz-making localities, designated as sites and constituting a site complex. With several unique exceptions, sites are isolated workshops or, more commonly, constellations of several functionally linked activity loci, termed site components. These include habitation rock-shelters, overhang shelters that were probably used for sleeping, open-air walled shelters, religious shrines, and workshops. To this list of commonplace components our 1976 fieldwork added single, unique occurrences of petroglyphs and pictographs (rock paintings), and a workshop where small tools were made from a nearby source of basaltic glass. Rock-shelters are small, natural overhangs utilized for habitation. All but two, at lower elevations on P6haku-loa and Wai-ka-halulu gulches, exhibit stacked-stone enclosing walls across the mouth. Four rock-shelters were excavated; initial radiocarbon dates for two of these, in the upper quarry area, ranged between A.D. 1424 and 1657 (McCoy, in press). The floor areas of rock-shelters are small, suggesting that other nearby overhangs, lacking midden deposits and fire hearths, were additional resting places. Open-air shelters are low, walled enclosures occurring on or at the base of talus slopes that were workshop locations, with one special exception. A group of 25 such shelters, occurring in three discrete clusters, was found in an area where virtually no evidence exists for adz manufacture. Workshops vary considerably in size and internal complexity. Some are thin scatters of flakes, cores, adz rejects, and hammerstones, while others are massive accumulations of these items, forming piles 20 to 30 meters across and perhaps 3 to 4 meters deep. Shrines are simple constructions, usually of long, narrow, thin, and flat slabs set in an upright position on a high point on or very near a workshop. A few non-workshop-associated shrines were found near the open-air shelter-complex, at Lake Wai-au, and near the western flank of Pu'u-Lilinoe. THE SITE DISTRIBUTION PATTERN Completion of an extensive and intensive survey provides the opportunity to examine and interpret the distribution of sites in terms of levels of exploitation. The discussion is limited to geological parameters and lacks a consideration of the temporal factor, since a quarry chronology is not yet fully developed. Quarry location and site distribution have a clearly defined, multifaceted geological basis. Site survey data indicate that exploitation of raw material was both extensive and intensive. Intensification, characterized by site clustering, large site size and internal complexity, subsurface mining, and the largest accumulations of debitage (waste by-products), is a function of the localization of a fine-grained vitreous basalt along a flow margin at the 12,200 to 12,400 ft. elevation. Along this general contour level the best quality basalt is found in the largest quantity. Maximal geographical limits of the adz quarry and locations of smaller, more dispersed, and less complex sites below 12,200 ft. are related to the effective lateral (roughly east-west) and downslope limits of surficial glacial drift deposits. Selective utilization of scattered boulder erratics characterizes the extensive exploitation pattern, which only in rare instances at the lower elevations also became a somewhat intensified and presumably longer term pattern of repeated adz making. Sites down to c. 8,600 ft. on Pijhaku-loa Gulch and c. 10,000 ft. on Wai-ka-halulu Gulch are exemplary of this broad ranging search for adz-quality basalt. The massiveness of debitage deposits in the upper quarry area suggests that sites there are the oldest. Lower elevation sites also might be early, but the limited quantity of usable boulders, and general absence of adz-quality basalt exposures on ridges and walls of the two gulches, would have prevented long-term exploitation in these areas. SEASONALITY AND INTERMITTENT USE One of the questions of greatest concern in comprehending the magnitude of the quarry is the length of time that it was in use, the regularity of exploitation, and time and duration of the work period in a year. Answers to each of these questions are not yet available and may never be as precise as one would wish. The first radiocarbon age-determinations indicate that the quarry dates to at least the early 15th century A.D. and probably 200 to 300 years earlier. Climatic conditions and data from rock-shelter excavations are used in a preliminary discussion of seasonality and intemittent use. Assuming for the moment that the general climatic conditions of Mauna Kea's upper slopes have been moderately uniform since A.D. 1000, it is reasonable to expect that coastal-adapted Hawai-ians would have worked in the area only in the warmer, snowfree, summer months. Annual temperature and snowfall fluctuations undoubtedly occurred, but in all probability the usual period of use would have been restricted to the months between July and October. Bones of a few immature dark-rumped petrel ('ua'u--Pterodroma phaeopygia sandwichensis) in rock-shelter middens is supporting evidence for summer habitation and adz production. William Bryan (1914:156) reported Hawaiians taking young ua'u from burrows always in September and October. Fish bones and scales and plant remains have not been analyzed, but additional data on seasonality might be forthcoming from such studies. Duration of the seasonal work period is difficult to establish with any degree of certainty because of our inability to isolate seasonal occupation floors within major stratigraphic units and the lack of precision in archaeological dating methods. Stratigraphic layers are homogeneous deposits of sequential occupations spanning a number of years. The best we can do at the present time is to make estimates based on quantity of food remains and a general assessment of maintenance activites. Weights for shellfi-sh, fish, and bird midden can be used as a rough index of length of occupation, using floor area to estimate numbers of people that a rock-shelter could accomodate. There are limitations to this approach, however, since it is likely that fish, birds, and shellfish constituted only a fraction of the diet. It is conjectured that prepared foods, such as poi and dried bananas, were quantitatively more important. Reco-very of birdbone awls and basaltic-glass flake tools, probably used for cutting and scraping wood and fiber, -indicates performance of maintenance activities. These material items and quantities of unprepared pandanus leaves suggest that items such as mats and carrying baskets were made or repaired, thereby hinting at residency greater than a week or two. There are additional considerations such as the length of time required to: (1) acclimatize to the high altitude; (2) sharpen manufacturing skills which may have been used infrequently or not at all during the remainder of the year; and (3) procure raw material, which at some time extended to the more involved process of mining. The quantity of food that could be transported from the coast would have effectively limited the length of stay, unless, of course, there were other people who formed task groups responsible for supplying the craftsmen and also, for carrying down selected preforms for finishing by grinding and polishing. The combination of these factors with archaeological reasoning based on excavated evidence suggests seasons no shorter than two to three weeks. Gaps in the cultural sequences of stratified rock-shelters, represented by sterile (non-cultural) soils, are evidence for intermittent exploitation. Periods of abandonment were sufficiently long for the development of B2 horizon soils. In two of the upper-elevation rock-shelters, there are three or more such soil developments suggesting general quarry abandonment rather than quitting of just one locality. Two B2 horizon soils in the upper part of one rock-shelter are bracketed by radiocarbon dates of A.D. 1424 and 1657. Reasons for abandonment are not certain, but we suspect that climate is a likely factor. Year-round snowcaps on Mauna Kea are reported in historic times, and we can expect short term cycles of increased cold, and thus snow, in the past. CONCLUDING REMARKS The Mauna Kea Adz Quarry Complex is probably one of the nation's least known but most important National Historic Landmarks, from both a research and interpretive point of view. It is the only Landmark of its kind in the United States. Moreover, it is probably one of the largest and most complex stone tool quarries in the world. Prior to the inception of the current project, the quarry was imperfectly known archaeologically; the boundaries had not even been established. As a consequence little had been done to insure its protection. As a result of the 1975-76 research we are now in a position to evaluate the full significance of the quarry and to recommend measures to be taken to maintain its integrity. A preservation plan is being prepared for the Department of Land and Natural Resources and the National Advisory Council on Historic Preservation. ACKNOWLEDGEMENTS The 1975 research was funded exclusively by the National Sci-enceFoundation (Grant No, SOC75-13421). Continued investigations in 1976 were supported by the NSF grant, a grant from the Cooke Foundation of Honolulu and matching Federal monies provided through a program established by the National Historic Preservation Act of 1966. The 1975 field crew consisted of Paul Cleghorn (field foreman), Jim McDowell, Holly McEldowney, Warren Osako, Chuck Streck, and my wife Judy McCoy. The 1976 crew included Paul Cleghorn and Tim Lui-Kwan (field foremen), Jay Aiu, Toni Han, Terry Hunt, Beverly Maekawa, Judy McCoy, Holly McEldowney, Monica Udvardy and Jamie Young. Others who participated in the fieldwork were: Dr. Richard Gould, Steve Hynson, Jennie Peterson, Marilyn Plott, Tom Manabe, and Dr. Alan Ziegler. To all of these people I am extremely thankful for their fine efforts in successfully completing what was intended and more. Special thanks are due to Patience Bacon and Bonnie Clause for their administering the grants, typing, and editorial work.

LITERATURE CITED Bryan, William Alanson. 1914. "Hunting the Uau on Molokai." The Mid-Pacific Magazine. 8(2):153-157 McCoy, Patrick C. In press The Mauna Kea Adz Quarry Project: A Summary of the 1975 Field Investigations. Journal of the Polynesian Society. Ugolini, F. C. 1974. "Hydrothermal Origin of the Clays from the Upper Slopes of Mauna Kea, Hawaii." Clay and Clay Minerals, 22:189-194. CATERPILLARS (EUPITHECIA SPP.) AS OBLIGATORY 'AMBUSH PREDATORS : A UNIQUE ADAPTIVE SHIFT IN THE HAWAIIAN ISLANDS S. L. Montgomery Entomology Department University of Hawaii at Manoa Honolulu, Hawaii 96822

Ambush predation, a previously unreported and bizarre feeding habit for the insect order Lepidoptera, has been discovered among geometrid caterpillars of the genus E'upithecia in Hawai'i. Certain endemic species consume no plant matter whatsoever; instead, they perch cryptically along leaf edges or plant stems and await passing insects. When prey individuals touch their caudal end, the caterpillars thrust the anterior part of the body backwards, seize the prey, and quickly return to the forward position. The entire strike, illustrated in Fig. 1, lasts about one-tenth of a second. While grasping the prey, sharp tarsal claws and spinelike setae of the thoracic legs often pierce the victim's exoskeleton (see Fig. 2). Unlike the praying mantid, vision is not utilized; one species is particularly alert in darkness. Prey observed being eaten in the field include calliphorid and dolichopodid flies, a cockroach, a gryllid cricket, a cosmopterygid moth, and an ichneumonid wasp. The geometrid genus Eupithecia has a worldwide distribution and contains about 1000 described species, with nine described species endemic to the Hawaiian Archipelago. None have been recorded previously as obligatorily predaceous (Balduf, 1939; Clausen, 1940) nor have I found any report of specialized behavioral or structural adaptations for capturing active prey by the larvae of any other moth or butterfly species. Those few predators recorded among the overwhelmingly plant-feeding Lepidoptera larvae are restricted to browsing on sessile, sedentary or otherwise passive prey with the only partially comparable exception to this statement being the West African lycaenids, which Lamborn (1914) found to crawl toward, drop upon, and eat leafhoppers. The first observation I made of this predatory behavior was of a caterpillar perched on a leaf holding and chewing an 8 mm blow fly (Calliphoridae) in a native forest on the slopes of Hualslai Volcano, Island of Hawai'i. This green caterpillar was taken to the laboratory securely perched along the lobeliad leaf edge on which it was cryptically appressed. After two days, the caterpillar had not eaten any of the leaf, and living drosophilid flies were placed in the vial. When a fly first touched the caterpillar's caudal region, a ready position was assumed in which the forehalf of the body is elevated. The caterpillar then struck and captured the fly (Fig. 1) which was then entirely consumed except for wing fragments. After feeding on three additional flies, the larva pupated and the resulting moth was identified as Eupithecia orichloris Meyrick.

Subsequently, E. orichloris has been reared through its complete life cycle in the laboratory, and larvae were proven predaceous throughout all four instars. One-day-old larvae (2.3 mm length) regularly caught and fed upon living book-lice (Liposcelis dCvinatoris). Larger caterpillars fed on a diversity of appropriate-sized prey, including DrosophiZa, mosquitoes and ants. Those prey too large to be held securely, or causing annoyance, were quickly thrown aside. The larvae do not react to flies moving near their heads and either ignore or attempt to shake off flies that crawl onto their heads or the anterior two-thirds of their bodies. Only prey that touch the rear third of the body, including the anal processes (paraprocts), are struck. Feeding experiments with 27 larvae showed that an average of 43 Urosophila of 3 mm length were consumed during the final larval instar, which attained an average body length of 21 mm. Adult moths fed on small amounts of diluted honey and females deposited 60-120 eggs placed in loose clusters of three to five eggs on the underside of leaves. Very similar obligately predatory behavior has been confirmed for an endemic complex of five other Eupithecia. E. staurophragma has been found on the Islands of Hawai'i, Maui, and O'ahu. It differs morphologically from E. orichloris in its longer anal prolegs and shorter anal processes. Also, the caterpillars are brown and almost without exception are found perching on twigs rather than leaves. Similar in morphology to E. staurophragma are three undescribed species from Moloka'i, O'ahu, and Kaua'i. In E. craterias, the caterpillar is light or brownish green and most frequently perched on leaf petioles. It has been collected on three islands and is distinguished by elongate setae on its posterior dorsum. E. scoriodes has a dark-colored larva and is known only from the subalpine zone of Maui, above 2000 m or 6500 feet elevation. All of these and two additional undescribed species from the Island of O'ahu possess elongate, heavily spined thoracic legs and have been observed to capture and consume prey in a manner near to that of E. orichloris. Although not fully investigated, E. rhodopyra is most likely a sixth obligatorily predaceous species, since the legs of the first instar are specialized for seizing prey. Most of these carnivorous geometrid larvae are sparsely but widely distributed in native forests and shrublands above 300 m in elevation. Over 300 specimens have been collected, usually well camouflaged and stationed within 1 m of the ground in partially shaded situations. They are quite sedentary, remaining in their perch position for days or weeks to snap up approaching insects.

The only Hawaiian species of Eupithecia which is principally phytophagous is E. rnonticotens. Swezey (1954) reported that it preferred the epacrid, Styphetia, but also fed on plants belonging to five families. I have learned that an important food is the protein-rich pollen of the abundant flowers of Metrosideros. E. rnonticotens has no leg or other specializations facilitating entomophagy, but cannibalism and necrophagy have been observed in the laboratory.

The radiation of Hawaiian Eupithecia into at least six predatory species is an intriguing instance of adaptive shift. These geometrids were probably inclined to predation by diet that included protein-rich pollen, as now seen in E. inonticotens. The behavior of snapping backwards is common to many geometrid larvae and is used in defense. With the Hawaiian Eupithecia, this behavior was intensified and refined into a grasping strike. A significant factor in allowing this behavioral shift in the geometrids was the disharmonic nature of the biotic environment. Of the approximately 250 insect colonizers of Hawai'i, unusually few were predators, with carabids, reduviids, nabids, hemerobiids, and chrysopids predominating (Zimmerman, 1948). Only the emesine assassin bugs, nabid bugs, and spiders were able to establish and radiate in the Hawaiian Archipelago Analogous groups present on the continents, such as the ambush bugs, mantids, and mantispids were totally lacking prior to contact by western man. The absence of many predators and parasites that attack lepidopterous larvae would also have permitted Eupithecia to adopt an ambush predation that involves waiting for prey in relatively exposed sites. Without doubt, these Hawaiian inchworms have one of the most interesting or "striking" life histories of any moth. Being unique to the Hawaiian Archipelago, it is a prime example of the adaptive shifts and radiations so notable in the biology of islands. Because at least four species are readily found within the National Parks in Hawai'i, it is recommended that the "grappling inchworms"' behavior be interpreted at Park visitor centers by film and photographic methods. If such a phenomenon could go undetected in Hawai'i until the 19701s,it prompts one to ask what other rewards await the curious naturalist in the earth's most singular island group. 2. Poises

Figure 1. Diagram of ambush predation by Eupithecia. Figure 2. E. stauropkragrna poised ready to strike

Figure 3. E. craterias Figure 4. The short-legged completing a meal. B. montico2ens is predominately a fruit and flower feeder. Acknowledgements

The author gratefully acknowledges the assistance of numerous colleagues, in particular William P. Mull of Volcano, Hawaii, who provided the photographs and many detailed observations. The drawing was prepared by David Kemble, courtesy af the Bishop Museum Entomology Department. Moth identifications were confirmed by Dr. Klaus Sattler who is revising the Hawaiian Eupithecia at the British Museum (Natural History) in London. This work was supported in part by National Science Foundation Grant GB 29288.

-Literature Cited

Balduf, W.V. 1939. The bionomics of entomophagous insects, Part 2. Swift Co., St. Louis. Clausen, C.P. 1940. Entomophagous insects. McGraw-Hill, New York. 688 pp. Lamborn, W.A. 1914. On the relationship between certain W. African insects, especially ants, Lycaenidae and Homoptera. Ent. Soc. London, Trans. (1913):436-498. Swezey, O.H. 1959. Forest entomology in Hawaii. Spec. Publ. B.P. Bishop Museum 44. 255 pp. Zirnmerman, E.C. 1948. Insects of Hawaii, Vol. 1, Introduction. Univ. of Hawaii Press, Honolulu. THE MAJOR VEGETATION TYPES AND ECOLOGICAL ZONES IN HAWAII VOLCANOES NATIONAL PARK AND THEIR APPLICATION TO PARK MANAGEMENT AND RESEARCH D. Mueller-Dombois Department of Botany University of Hawaii at Manoa Honolulu, Hawaii 96822

In 1965 I prepared a vegetation map for the Park after Dr. Fosberg had drawn tentative boundaries on a set of 1954 air photos. Ground work for this took about four months and following that, I had two graduate students working in 1966 and 1967 on vegetation samples (i.e., 64 relevb) that were distributed throughout the major vegetation units. These were established for deriving more detailed information on the plant composition, tree growth and soil substrate of each unit. The map was first presented together with the Atlas for Bioecology Studies (Doty and Mueller-Dombois 1966) on 55 individual 70 cm x 70 cm air photos as transparent overlays at a scale of 1:12,000. In this form it was available only in two sets (one located at Hawaii Volcanoes National Park Headquarters, the other at the University of Hawaii Botany Department). In this bulky format, the map has been used by a few researchers as a tool for orientation, but it was a time-consuming job for them to obtain the needed information. For the same reason little use had been made of the mapped information by management. Because of a general request, the air photo map was later transcribed to the 1:24,000 topographic map sheets covering the Park. Then it was further reduced to a scale of 1:52,000 and made generally available in a CPSU technical report (Mueller-Dombois and Fosberg 1974). In this format, in which the map appeared on 25 topographic half-sheets of each 8% x 11 inches, the map became useful as a field reference guide in notebook form. Several people disassembled this report and glued the individual map sheets together into a single sheet for providing a general overview. Apart from the technical difficulty of reading the map symbols from a distance, the mosaic of vegetation units of the whole Park presents further problems of interpretation which are inherent in any two-dimensional map projection of single ecosystem components. For overcoming this problem or for bringing the mapped information "to life," I would like to briefly discuss three tools for interpreting such a map. Tools for Map Interpretation

(1) Environmental Data: Plant communities, which can be defined spatially as recurring similar plant assemblages, form in response to environmental forces acting in certain patterns upon the flora of an area. One of the outstanding environmental forces is the climate, another the substrate on which the communities occur. The latter is a very obvious control on community development in this Park where lava flows of differing dates and ash deposits determine the vegetation boundaries in many cases. Therefore, a geological map drawn to the same scale and used as a transparent overlay could serve as a useful interpretive tool for the vegetation map. This aspect has not yet been utilized. However, here I would like to draw particular attention to the use of climatic data. The Park Management has accumulated rainfall data over the years in various places and this has not yet been fully exploited for interpretive purposes. I had begun to do that for the Atlas of Bioecology, and the outcome was the climate diagram map for the Island of Hawaii (Figure 1). Generally, the most important climatic parameters for plant and community distribution are rainfall and temperature. With respect to rainfall, perennial plants are controlled not only by the normal (or mean) annual amounts of precipitation occurring at a given site, but especially also by the variation of rainfall throughout the year. Therefore, mean rainfall isohyets explain vegetation distribution only in a very general sense. Month-to-month variations are more critical, particularly in the tropics. Temperature may control certain plant distributions more or less directly in the form of heat, but temperatureis more important indirectly as it determines largely the effectiveness of rainfall for plant growth. For example, 1000 mm of rain falling in the mountain parkland on Mauna Loa provides more effective moisture for plant growth than 1000 mm falling in the coastal lowland of the Park, where the mean air temperature is about 10°C higher. These ecological aspects are incorporated in the climate diagrams, which were prepared after the method of Walter (1957): The months of the year are shown on the abscissa of each diagram (from January to January with July in the center). The left ordinate has a temperature scale in "C and the right ordinate a precipiitation scale in nun. Each diagram shows two curves, an almost horizontal or slightly one-modal, smooth curve, which gives the mean monthly temperatures of the station, and an uneven or zig-zag curve, which shows the month-to-month rainfall variation. An index of rainfall effi.ciency is built into the diagrams by making 20 nun rainfall correspond to 10°C in the scaling of the two ordinates This has the advantage that one can recognize with one glance whether a place has a seasonal or a continuously moist or humit climate. Wherever a rainfall curve undercuts a temperature curve a drought season is indicated. An example is Hale-ma'uma'u Crater at the north end of the Ka'; Desert. In the tropics such severe seasonality is similar for the vegetation to a winter season in the temperate zone, i.e. the perennial vegetation can be expected to be dormant at that time. A dry season is indicated where the rainfall curve drops below 100 mm or 4 inches per month, and a wet season is indicated where the monthly rainfall remains above 100 m. A typical rain forest climate shows month-to-month precipitation in excess of 100 mm. Such climates are indicated at Park Headquarters, Hilo, Ka-poho and other places, where rainfall in excess of 100 mm is shown as black fields on the diagrams. For maintaining the diagrams within practical sizes, the scale above 100 mm was reduced by 10%. This is indicated by blackening the area under the curve. Each diagram shows also the station name, its altitude (here by map location) and the mean annual rainfall. Mean annual temperature is shown only for stations with on-site records. For the others the temperatures were interpolated by using the adiabatic lapse rate and elevation of the nearest recording station. Although there are diagrams for only five stations in the Park and four nearby, this information is extremely useful for explaining the major vegetation types in the Park. Another useful tool is the following. (2) Profile Diagrams: Figures 2 and 3 are two diagrams out o? the five that I prepared originally for interpreting the vegetation map. Their locations are indicated on the climate diagram map. The profiles give first of all a third dimension, which is missing on the vegetation map. This third dimension, indicated by the topographic variation along each profile, removes some of the abstractness that is inherent in any flat map projection. In addition, the profile allows one to further abstract and indicate climatic and substrate relationskips as well as the major vegetation-structural features which cannot be shown on the vegetation map. By displaying the four components, climate, vegetation, soil substrate and topography together, one can now speak of having established ecosystems in the structural sense. Twelve such structural ecosystems are shown for the east-flank of Mauna Loa (Profile 1) in the Park, and eleven are shown for Profile 2, which extends from the KT-lau-ea Forest through the Ka'u Desert via Hilina Pali to the coast. By establishing samples in all of these units and by comparing them with one another, one can then establish ecosystem types, i.e. they can be identified systematically. For example, ecosystem nine on Profile 1 and ecosystem eight on Profile 2 are the same (the closed mixed forest of ~I~uka-pua-ulu),also unit ten on Profile 1 and unit seven on Profile 2 are the same ecosystems, the open 'ohi'a (Metrosideros) scrub-forest with native shrubs and lichens, which extends from the Tree Molds to ~F~uka-puaulu),and south to around the Volcano Observatory. On this basis of comparison, the Park's vegetation was classified into 31 major vegetation or structural ecosystem types (Mueller-Dombois and Fosberg 1974) . (3) Ecological Zones: Such a large number of major vegetation types is still not easy to comprehend. Their relationships, however, are more easily understood when they are grouped into ecological zones. In this way, the number of vegetation types is reduced to less than 10 for each zone, a number more manageable for Park policy considerations. When based on vegetation research, knowledge of animal home ranges and environmental data, such ecological zoning is not entirely an arbitrary divisioning. An important criterion for ecological zones are macroclimatic boundaries. In a terrain with pronounced topographic variations (as existing here in Hawaii Volcanoes National Park), topography is a major integrator of climate. Therefore, ecological zones can be defined in part by topographic boundaries for mapping. Where certain types of substrate prevail over large areas, the soil-substrate may help in defining ecological zones. In Hawaii Volcanoes National Park, which in its entirety occurs on recent volcanic substrates, the finer substrate boundaries occur mostly on too small a scale for ecological zoning or they extend, like some lava flows do, through a range of macroclimates. Therefore, soil substrate was of more Limited value in defining ecological zones for the Park. A fourth important component is the vegetation itself. For example, the rain forest in the Park cuts across a considerable range of altitude (from about 5000 to 1500 feet), where it reflects the boundary (approximately along the Chain-of-Craters Road) of the moist windward and drier leeward orographically controlled air movements.

Ecological Zones as Major Ecosystems By integrating climate, topography, substrate, and vegetation, the six ecological zones (here superimposed on the Park's vegetation map) were derived. An important functional element can now be added that renders each zone to be considered as a major ecosytem in the Park, which deserves to be treated as having some individuality of its own. (1) Alpine Zone: This zone contains only two major vegetation types, a Rkacomitrium moss desert [map symbol r(R)1 and a Vaccinium-Styphelia lowscrub desert [symbol r(ns)l. This sparsely vegetated area, which extends from the summit of Mauna Loa to 8500 feet (2590 m) altitude has a dry, cool (500-750 mm rainfalllyr; 6"-9.5"Cmean air temperature) climate. The peculiarity of this area is that ground frost occurs nightly throughout the year. This night-frost climate limits development and activity of many biota. Even horse droppings on trails remain undecomposed for years, it seems. (2) Subalpine Zone: Three major vegetation types were mappea in this zone: an open to closed globose scrub (without trees) [symbol ns; the shrubs Vaccinium peleanum, StypkeZia douglasii and Dodonaea viscosa form clusters of dominants], a globose scrub with scattered Metrosideros trees [symbol ns(M), the treeline ecosystem], and an open Metrosideros scrub-forest with scattered Sophora trees [symbol oM(So-ns)]. This zone extends in the Park from 8,500 down to 6,700 feet (2590-2042 m) altitude. The most interesting vegetation type here is perhaps the treeline ecosystem, which experiences night frosts at least through the winter months of the year. Here also we find- - the upper limit and well-established habitat of the Nene (Hawaiian goose) and an environment favorable for growth of the silversword (Argyroxipkium sp.). The climate is summer-dry with about 1000 mm annual rainfall and cool with 9.5" to 12'C mean air temperatures. Daily temperature ranges are pronounced and may be 10" to 12'C. The substrate is mostly lava rockoutcrop of both pahoehoe and a': with little soil overlay. This results in the sparseness of grasses. The subalpine zone thus appears less of a fire hazard zone. Also, pigs are largely absent because of the lack of soil for grubbing. In contrast, goats have been seen to browse in this zone. (3) Montane Seasonal Zone: This zone is characterized by frequent clouds near ground. It has also a summer-dry climate but with more rainfall (from 1100-1600 mm/yr) and warmer mean air temperatures (of 12"-17°C). Five major vegetation types were recognized and mapped in this zone. The three more important ones are: the mountain parkland [symbol mx-ns(AcSoM)] the savannah [symbol mx-AcSaMl and the forested Kipukas [symbol AcSaM(ad)]. The other two types are Metrosideros communities on more recent lava flows and or ash deposits [symbols oM(ns-L) and scM(ns)l. The zone extends from the upper limit of the mountain parkland (from the Mauna Loa Rainshelter) at 6700 feet (2042 m) to 3800 feet (1160) around and south of Volcano Observatory. The two grassland vegetations (parkland and savannah) had periodic fires in the past (Mueller-Dombois and Lamoureux 1967). A few native grasses are present (Eragrostis grandis, Panicum tenuifolium, Deschampsia australis and Trioetum gZomeratuml. Acacia koa sprouts vegetatively from root suckers. Fire in these habitats, if not too frequent and extensive, does not seem to form a serious threat to the integrity of these communities. Activity of introduced ungulates is high in this zone. It is a favored environment for cattle grazing outside the Park, and an area where feral goats and pigs overlap in the Park. After the more effective goat control now, feral pigs are still a major cause of damage (Spatz and Mueller-Dombois 1975). (4) Montane Rain Forest Zone: Six major forest types are mappemr this zone. Four of them have Metrosideros co~iina('ohi'a) as the sole dominant tree, one is a tree fern (Cibotium spp.) forest and the other a mixed koa-'ohi'a rain forest. The four pure 'ohi'a forest types are differentiated by structure (open vs. closed) and by the differing dominance of undergrowing ferns (Dicranopteris VS. Cibotium) . The climate is humid throughout the year (i.e. 100 mm monthly rainfall or more). A short dry season in June may occur as a normal event, however, for example, around the Park Ileadquaruers area. Mean annual rainfall is generally above 2000 m, md in the Park may go up to 4000 mm ('Ola'a Tract). The major perturbations in this zone are the feral pigs and the 'chi's dieback phenomenon. Both of these merit attention by Park Management as they may threaten the integrity by precipitating the invasion of exotics and destruction of endangered endemic species. Fire has never been a problem here, neither have feral goats. The substrates in this zone vary from lava rockoutcrop to deep ash. Most of the destructive pig activity is restricted to the rainforesttypes with ash substrates. (5) Submontane Seasonal Zone: This includes a large area in thmfrom south of Halema'uma'u and south of the Chain-of-Craters Road, including 'Aina_houRanch and Nau1.u Forest downslope to the palis (Holei and Hilina Pali). Six major vegetation types were mapped in this zone. These are the broomsedge (Andropogon) savannah above Hilina Pali [symbol An(ns-M) and An(ns-i) 1, a large area covered with an open 'ohi'a-broomsedge (Metrosideros-AndropogonJ-native shrub forest [symbol oM(ns-An)], a native closed shrub vegetat-ion with scattered 1 ohi'a- trees [symbol ns(i-An-M)] (which occurs mostly east of the 'Aina-hou Ranch in an area now poorly accessible because of lava flows), and an open and a very open 'ohi'a-lama (~etrosideros-Diospyros) forest [symbol MD(ns-i-An-r) and MDlr), respectively] mostly along the major palis on ' a'a rubble. The sixth important vegetation type in this zone is the Ka'G Desert with extremely sparse xerophytic vegetation [symbol r-ash(poik)]. Apart from its substrate peculiarity (shifting ash dunes) and its sulfur steaming influence from Halema'uma'u Crater, the ~a'cDesert also is influenced by a more severe dry season (i.e. drought) in the summer and thus can be recognized as a separate subzone of the submontane seasonal zone. Also with regard to management considerations it may be a separate entity. Fire in the other vegetation types of the submontane seasonal zone is a real threat, both from the viewpoint of easily combustible material (particularly the presence of broomsedge) and the destructiveness to endemic species. Metrosideros (in contrast to koa) does not resprout well after fires and noneof the native shrubs in these vegetations (except perhaps Dodonaea viscosa) appear to be fire adapted. Fire apparently has never been an evolutionary stress factor in this zone as it was in the montane seasonal zone. Fire has only become a real problem lately with the invasion of broomsedge. Thus, fire prevention techniques will need to be perfected particularly for this major ecosystem of the Park. (6) Coastal Lowland Zone: This Park ecosystem is defined as the area below approximately 1000 feet (330 m) elevation extending along the coast from Kalapana to below Ku-ka-lau-'ula Pali. Seven major vegetation types were mapped. Two were grasslands, an anuual Eragrostis teneZZa (love grass) grassland 3nd a perennial Heteropogon contortus (pili) grassland. One was a lowshrub savannah, i.e. Heteropogon grassland with low shrubs [symbol H(ls-i)] and three were predominantly woody vegetations, a widely scattered old Metrosideros tree vegetation without appreciable undergrowth mostly on old 'a'; lava flows [symbol r(M)], a mixed (native and introduced species) lowland scrub [symbols ls(i) and lsi(E-H), including the native Canthium odoratum, Wikstroemia ?phiZZyraefoZia, Diospyros ferrea and Erythrina sandvicensis and the exotic Sohinus terebinthifolius, PZuchea odorata, Psidium guajava and Eugenia cumini] and an open mixed lowland forest of mostly exotic trees (Mangifera indica, Samanea saman, AZeurites moluccana, l'hespesia, Pandanus and Cocosl. The latter vegetation types occur near Kalapana and the change from woody vegetations to annual grassland from east to west along to the coast is related to a decreasing mean annual rainfall gradient from about 1700 rmn to 700 mm (south of the ~a'iiDesert). The seventh major vegetation type mapped includes the salt spray and other strand communities found in a narrow (about 50 m wide belt) directly along the coast. The coastal lowland zone has a warm-tropical climate (mean air temperature 23°C) with an increasingly severe summer-dry season from east tcr west and frequent, strong and desiccating winds. This is the Park ecosystem that has suffered the strongest alteration under the feral goats that were very abundant for probably more than a century here up to 1972 (Mueller-Dombois and Spatz 1975). With the fencing and goat eradication program nearing completion the vegetation is undergoing considerable changes and should be remapped soon.

Two exotic, taller-growing woody plants (Leucaena ZatisiZiqua and Ricinus cornrnunis), both with an enormous seed-producing potential, are now invading the chamaephyte stage in a few locations. If this process is not closely checked, this area will convert rapidly into an undesirable thicket with very low species diversity. At the present time, these two weed-tree species form only a few small colonies in the vast area of this coastal lowland. As step 2 in the ecosystem restoration program of this formerly goat-infested territory, these two exotic tree-pests should be monitored for elimination at very frequent (preferably monthly) intervals. As step 3, a few strategically located seed source centers should be established by planting (in aggregation) native dry-zone tree species (such as Canthiurn odoraturn, Dispyros ferrea, Erythrina sandwicensis). The seed source for these centers should be obtained from within an approximate three-mile limit of the formerly goat-infested territory.

Conclusions In a recent article in the Journal "Parks," R. M. Linn (1976) (formerly NPS Chief Scientist) declared vegetation mapping and park zoning the basic tool for masterplanning in the U. S. National Parks. Mapping can be done for different purposes. For masterplanning a map should in my opinion accomplish three major objectives. It should form a useful frame of reference for: (1) Research: One of its functions here is to form a basis for sample stratification and for determining the spatial extent to which results from biological field studies can be extrapolated. Another, and perhaps even more important function of avegetationmap is to encourage research with more specific questions in mind. For example, the Mauna Loa Transect analysis of our IBP research was generated through this map. (2) Management: Here it is particularly the more generalized level of ecological zones that appears useful when superimposed on a vegetation map. As I have tried to emphasize in the foregoing discussion, the different nature of each zone merits individual treatment and the establishment of different management objectives for each zone. On a functional level the zones can be recognized as the six major Park ecosystems, while the vegetation. types may be regarded as forming important modifications in each ecosystem. (3) Interpretation: My special concern in this paper was to convert the more abstract information presented in the vegetation map into a more concrete form for Park interpretation. This 1 did by emphasizing three tools for map interpretation: environmental data, profile diagrams, and ecological zoning. The latter was given particular emphasis as a tool for summarizing the mapped vegetation types and as a key to interpreting the nature and distribution of the major dynamic events occurring as periodic or erratic perturbations in the Park, i.e. fires, goats, pigs, frost, volcanic activity, 'ohi'a dieback and encroachment of exotic plant species. These perturbations should further be interpreted for each ecosystem as to whether they represent long-standing evolutionary stress factors or are recent and therefore artificial stresses which form a threat to the integrity of the system in question. Figure 1. Climate diagram map of the island of Hawaii with location of five ecosystem profiles in Hawaii Volcanoes National Park. WIAN AMVUAL RAINFAU ~~~~-~-~~1500mn----t2500mn MEAN ANNUAL TEMPERATURE

6OC 9.5T : 149: 7 169:

Figure 2. Profile 1: East flank of Mauna Loa in the Park. APWOX. MEAN ANNUAL AIR TEMPERATURES 23°C - . 14-C SUMMER- UMMER-DRY SUMMER-DROUGHT CLIMAIE SUMMER-DRY WITH HUMID S-WY DROUGHT CLIMATE WlTH STRONG WINDS FREWEM LDW CLOUDS TRANSITCA 6 ! ,8910 I I

Figure 3. Profile 2: South slope on Kilauea via Hilina Pali (segment 3) to the Pacific Ocean. Literature Cited Doty, M. S. and D. Mueller-Dombois. 1966. Atlas for Bioecology Studies in Hawaii Volcanoes National Park.

Hawaii Botan. Science Paper No. 2. ' 507 p. Republished as Hawaii Agric. Exp. Sta. Bull. 89 (1970). Linn, R. M. 1976. Introduction to Masterplanning. Parks (Intern. J. for Managers of National Parks, Historic Sites and Other Protected Areas) l(1) : 5-8. Mueller-Dombois, D. and C. H. Lamoureux. 1967. Soil-Vegetation Relationships in Hawaiian Kipukas. Pacific Science 21(2): 286-299. Mueller-Dombois, D. and F. R. Fosberg. 1974. Vegetation Map of Hawaii Volcanoes National Park (at 1:52,000). CPSU-Hawaii Tech. Report No. 4. 44 p. Mueller-Dombois, D. and G. Spatz. 1975. The Influence of Feral Goats on the Lowland Vegetation in Hawaii Volcanoes National Park. Phytocoenologia 3(1): 1-29. Spatz, G. and D. Mueller-Dombois. 1975. Succession Patterns after Pig Digging in the Grassland Communities on Mauna Loa, Hawaii. Phytocoenologia 3(2/3): 346-373. Walter, H. 1957. Wie kann man den Klimatypus anschaulich darstellen? Die Umschau in Wissenschaft und Technik, Heft 24: 751-753. LIVE COLOR PHOTOGRAPHY OF ENDEMIC HAWAIIAN INVERTEBRATES Willam P. Mull Research Associate Entomology Department Bishop Museum Honolulu, Hawaii 96818

One of the great rewards for a field biologist is the excitement of finding a new species, a rare species or an unsuspected behavior. If the event can be captured on film in living color, the excitement of that discovery can be shared again and again with colleagues and other interested people. Also, the pictures of the organism alive and fresh can record morphological and behavioral data that are not revealed later by the preserved specimen. For the past five years, I've been working with invertebrate zoologists to photograph their discoveries among Hawai'i's endemic insects, spiders, land snails and other land invertebrates. In the next few minutes, we'll be looking at color slides that show some of the product of that experience--which has been at least as exciting and rewarding for me as for my scientist friends." Imagine the joyful shock Steve Montgomery (University of Hawaii) felt three years ago in the Wai-'anae forest on O'ahu, when he looked under a leaf and saw this happy little face beaming at him. This tiny theridiid spider with the unbelievable happyface pattern on her back, carefully guarding her cluster of eggs, is apparently an undescribed species. Two years later, Wayne Gagn6 (Bishop Museum) found this male relative of hers--with "eyebrows" added--laughing back at him from under a leaf in a Big Island forest; several more male and female specimens like this have since been collected here. Then last month, Klaus Sattler (British Museum, Natural History) found this happy little female, with the coy red mask, under a leaf in wet forest in the Kohala Mountains of the Big Island; she appears to be yet another new "happyface" spider species. At the same time, on Maui, Steve Montgomery came up with this "smiling" lady to add still another apparently distinctive species to this remarkable group of petite Hawaiian "happyface" spiders--which have eluded field workers here all these years.

*Editor's Note: The many color slides presented with this paper have not been reproduced here. Persons interested in viewing the slides should contact Mr. Mull directly. Another chain-reaction spider event was set off five years ago when Frank Howarth (Bishop Museum) made his startling discovery of whole new ecosystems in the dark zones of Hawaiian lava tubes--one inhabitant of which was this species (Lycosa howarthi), the world's first-known small- eyed big-eyed hunting spider, first found here in Hawaii Volcanoes National Park. The vestigial eyes of this Big Island lycosid spider were soon overshadowed, though, by Frank s discovery a month later of the world's first-known no e ed big-eyed hunting spider (Adelocosa anops)--in a low&va tube onKauati, With no possible means of underground dispersal betweenKauafiand the Big Island, these two amazing spiders seem to represent a classic example of parallel evolution in an unexpected direction by closely related organisms with similar opportunities. As logical follow-up to these discoveries, a search was pursued for close surface relatives. This produced several apparently new lycosid species, including two from the Park; "Clarence," here was found by Wayne Gagn6 at che 8,000 ft. level on Mauna Loa, and I found "Big Mother," here, at the 5,000 ft. level just before those 300-plus piggyback youngsters emerged from the now-empty egg sac clutched under her--and before she got this contented-motherhood look on her face. When Frank Howarth found this reduced-eyed Caconemobius cricket in lava tubes in and near the Park, he started a search of the Park's recent lava flows for a surface-dwelling relative and discovered this closely-related normal-eyed species, previously unknown. In another lava-tube cricket event, Frank found this ThaumatogryZZus cricket with reduced eyes in Big Island lava tubes several years ago and followed-up just last month with discovery of this ThaumatogryZZus, with even- more-reduced eyes, in a Maui lava tube. So far, the only known ThaumatogryZZus cricket species that is surface dwelling lives on Kaua'i. One of the rarest, most cryptic, most elusive groups of endemic Hawaiian insects are the thread-legged bugs. This fancy last-instar nymph of an apparently new species of NesidioZestes fell out of a tree onto their collecting sheet while Steve Montgomery and Jim Jacobi (University of Hawaii) were light-collecting moths at night on West Maui in early 1974. Six months earlier, in a Big Island lava tube, Frank Howarth and Wayne Gagn6 had come up with another new, even more exciting NesidioZestes find--the world's first- known fully cave-adapted thread-legged bug, with greatly reduced eyes and no pigmentation whatsoever. A year of searching the Big Island for a surface relative finally produced this male last-instar NesidioZestes nymph (followed by a dozen more specimens in thenext few weeks), which I found under epiphytic AsteZia on a mossy tree trunk in wet forest. It may be N. seZium, known only from a single damaged specimen in the British Museum collected over 70 years ago--the only previously known surface thread-legged bug endemic to the Big Island. Coincidentally, a month later, Frank, Wayne, Steve and Robin Rice (University of Hawaii) found a series of long-uncollected N. insuzaris during a night-time collecting foray in upper Makaha on male is one of them. A fortuitous thread-legged bug bonus from my Big Island search for NesidioZestes was this male SaiceZZa specimen, which also was under AsteZia, which is the first Big Island record for the genus, and which appears to be the first truly wingless species in this group of otherwise micropterous forms. This female came along with the male and started laying eggs--producing over 50 before she died four months later. The eggs hatched, and I doted over those SaiceZZa kids for more than six months--finally rearing 18 through to adulthood. Another bonus from my thread-legged bug search on the Big Island was this flightless mirid bug, also from AsteZia, which doesn't really fit into any described genus. It is close, though, to the technically monotypic O'ahu genus Nesidiorchestes, of which several undescribed species exist in collections--including this new one discovered on Maui by Wayne Gagn6 in 1973.

Still another bonus was this male Nabis, also from Big Island AsteZia, which fell into my net along with this female. They are distinct from the several Nabis species described from this island. Another apparently new Nabis species from the Big Island is this flightless male, collected on Sphagnum in wet forest by Wayne Gagn6 last year. In the past year, Robin Rice has made a series of remarkable collections of rarely found endemic Rhyncogonus weevils. He has turned up most of the described Hawaiian species and, in the process, discovered several new ones, such as this male from koa onKaua'i; Along with the male, he got this female--who laid these eggs during the several days I worked with the pair to get photographs. I'm including nothing about the substantial recent discoveries amongHawai'ilsdrosophilid flies, or about the incredible endemic caterpillers that catch them--since both subjects will be addressed later during this conference--but I must show you this engaging little male tephritid fly with the striking wing patterns, proclaiming his territoryon the Bidens leaf. Steve Montgomery reared him from a Bidens stem gall he found last December on West Maui. It was an undescribed species with apparently only one previous specimen known, from a Maui collection many years ago. Its biology was unknown. Rarely does one have the opportunity to take live pictures of "extinct" species. But this beautiful endemic tree snail, AchatineZZa ZiZa, was found in the Ko'olau range onO1ahuinMarch 1974, along with several other living specimens. The species was regarded as extinct--as was this A. byronii and this A. casta, both found in July 1975, also in theKotolau range. Their rediscoverer was A1 Hart (University of Hawaii), who in the past three years has made a tremendous contribution to our understanding of the current status of most members of this marvelously-evolved but fast-declining endemic O'ahu genus, AchatBneZZa. The other day, I took time out from working on biological rarities to photograph this knobby-headed caterpillar with its head next to a similarly knobby new flower bud of its Pipturus host plant. Is there any significance to this curious structural similarity between these two closely associated but very different organisms? This picture reminds me that the potential for exciting discovery isn't confined to new or rare species. Even this commonplace Hawaiian Pipturus plant with its commonplace Ka-mehameha butterfly larva can hold revelations to quicken the pulse of the field biologis t--and his photographer. HOST SPECIFICITY IN HAWAIIAN DROSOPHILA: NUTRITIONAL OR BEHAVIORAL? Alan T. Ohta Department of Genetics University of Hawaii Honolulu, Hawaii 96822

For the past thirteen years the Hawaiian DrosophiZa project has been intensively studying the endemic DrosophiZa of these islands and in particular the "picture-winged" species group. Hawaiian DrosophiZa use the rotting bark, leaves, and fruit and sap exudates of the endemic flora as breeding sites (Heed, 1968). Different species of Hawaiian DrosophiZa utilize from one to many different families of host plants as breeding sites thus displaying a range of highly specialized to highly generalized species. In the "picture-winged" species group 62 species have been classified as monophagous and five as polyphagous species. D. crucigera typifies the generalist species having been reared from different plant families (Montgomery, 1975). Heed (1971) recognized that specialism and generalism becomes a much more complex classification when we consider, in addition to the number of plant families involved, the parts of the plant used as a breeding site. Some specialists may utilize an entire plant (leaves, bark, root, and fruits) and some generalists may breed in only one part of several plant families.

The grimshawi subgroup of "picture-winged" DrosophiZa is of particular interest with regard to breeding sites. D. grimshawi populations from different islands show an interpopulational differentiation of host specificity. D. grimshawi from Maui Nui (Maui, Moloka'i, and La-na'i) are generalists in breeding sites and have been reared from twelve different families of plants including two introduced families. This makes it the most generalist "picture-winged" DrosophiZa besides D. crucigera.

However, D. grimshawi from O'ahu and Kaua'i and D. puZZipes from Hawai'i, a morphologically nearly identical species (Hardy and Kaneshiro, 1972), are specialists in breeding sites and will ovipost only in the genus Wikstroemia. Skotts- berg (1972) describes some 26 species and several varieties of this genus while Rock (1974) classifies only eight endemic species. This plant, known to the Hawaiians as 'Skia, contains an acrid narcotic and has been used to stupefy fish. It also provided the Hawaiians with a very strong fiber due to the tough fiberous nature of its bark. Laboratory rearing of these flies indicate that this ecological separation is not due to differences in the nutritional requirements of the larvae but rather to the ovipositional behavior of the females. Stocks of D. grimshawi from each island have been formed from isofemale lines of wild caught females and one stock, O'ahu, from flies reared from rotting Wikstroemia brought in from the field. These stock larvae (both generalists and specialists) appear to do equally well in identical laboratory media and the stimulation of female oviposition seems to be the only difference between these populations. The generalists D. grimshawi will ovipost on lab food and/or in the CZermontia soaked kerchiefs which are placed in the vials to inhibit mold growth. The specialists D. grimshawi and D. pulZipes however require a small piece of rotted Wikstroemia to be added to the vial to elicit oviposition. The larvae do not appear to require the bark as this can be recovered essentially "in tact" from a vial of specialist larvae and the larvae can complete their development in the same medium as used by the generalist larvae. It seems that the female ovipositional behavior, in this group, dictates the number of substrates on which the larvae will feed. This makes it inappropriate to use the adjectives monophagous and polyphagous to describe the breeding habits of these species. Webster's New World Dictionary defines the root word phagous as "that eats." Monophagous then is defined as feeding on only one kind of food and polyphagous as the eating of or subsistence on many kinds of food. These definitions, although not incorrect, imply that the larvae will only feed on one or many types of plant hosts. This however is misleading for although they may only be reared from one specific host plant it is a result of the specificity of female ovipositional behavior and not of larvae requirements.

Other species of Hawaiian DrosophiZa however do show a host specificity due to larval nutritional requirements. This is best exemplified in the "leaf breeder group" which is composed of a very loose class of Hawaiian DrosophiZa composed of several species groups. These flies breed almost exclusively in the rotting leaves of the endemic flora and most, if not all, are specialists in that they have been reared from only one or two plant families (Heed, 1968). Flies, from this group, brought in from the field will oviposit when stimulated with extracts from leaves of their host plant. Successful rearing in the lab of the larvae however, has thus far failed as the lab media provided appear to be lacking some essential element(s) for larval development (Robertson, et. al., 1968). A few adults have been reared in the lab through intensive efforts but these proved to be incapable of producing progeney due to the nondevelopment of the FI female ovaries. It is therefore evident that it is not the ovipositional behavior of the female which dictates the breeding substrate but rather the nutritional requirements of the larvae. The determination of the mechanism of host specificity may bean important step toward understanding the evolution of different species groups of Hawaiian DrosophiZa. More- over it may provide us with a method of pest control using host shift speciation. Insect pests breeding on economically important host plants may be shifted onto another host plant which may be of little or no economic importance. However in order for this to be accomplished much more investigation into the mechanisms which control this type of evolution is needed. References

Hardy, D. E. and K. Y. Kaneshiro, 1972, New Pictured-Winged Drosophila From Hawaii, Part 111 (Drosophilidae, Diptera) University of Texas pub. 7213:155-161

Heed, W. B., 1968, Ecology of the Hawaiian Drosophilidae University of Texas pub. 6818:387-419

Heed, W. B., 1971, Host Specificity and Adaptive Radiation in Hawaiian Drosophilidae. Taxon 20:115-121

Montgomery, S. L., 1975, Comparative Breeding Site Ecology and the Adaptive Radiation of Picture-Winged DrosophiZa (Diptera: Drosophilidae) in Hawaii. Proceedings of the Hawaiian Entomological Society 22(1):65-103

Skottsberg, C., 1972, The Genus Wikstroemia Endl. in the Hawaiian Islands. ACTA Regiae Socientalis Scientiarum Et Litterarum Gothoburgensis Botanica I 166 pages

Robertson, F. W., M. Shook, G. Takei, and H. Gaines, 1968, Observations on the Biology and Nutrition of Drosophila disticha, Hardy, An Indigenous Hawaiian Species. University of Texas pub. 6816:279-299

Rock, J. F., 1974, The Indigenous Trees of the Hawaiian Islands, Charles E . Tut- Co . Inc . , pp.6319 THE EFFECTS OF FIRE UPON A HAWAIIAN MONTANE ECOSYSTEM Terrv T. Parman Hawaii ~olckoes~ational Park Hawaii 96718

In the early afternoon of August 13, 1975, a small fire was reported by Park Service personnel along the eastern boundary of Hawaii Volcanoes National Park at an elevation of 4900 feet. Pushed by strong easterly trade winds,the fire spread rapidly westward from the Ke-au-hou Ranch boundary, across the Mauna Loa Scenic Road and into the seasonally dry mountain-parkland and savannah grassland segments of Hawaii Volcanoes National Park. During the next four days this fire consumed some 1800 acres of grassland, native shrubland, and Acacia koa colonies, before being contained on the Ka-p5pala section of Parker Ranch, 4.8 miles southwest of its origin. Vegetation within the burned area can be roughly divided into three community types: 1) closed canopy Acacia koa colonies with a mixed grass understory, 2) open native shrubland with native and introduced grasses, 3) mixed Acacia koa, Sophora chrysophyZZa forest with mixed grass and shrub understory. Ten study transects were established in the burn area three weeks after the fire, in order to monitor recovery of vegetation. Three of the study plots are located inside colony stands of Acacia-koa. One study site is situated inside a mixed forest kipuka consisting of Sophora chrysophylZa, Acacia koa and a few small Metrosideros sp. The other six transects are located in native shrub and mixed grassland communities. All ten transects are situated at elevations between 4400 feet and 5100 feet. Percent ground cover by species for each transect was determined by the point frequency method. Point sampling was done using finely sharpened bronze rods mounted in a one-meter long wooden frame. Five points per meter, at 20 cm intervals, were sampled in each of the ten sixty-meter transects, resulting in 300 points per transect. Sampling along each transect was repeated every two months from September 1975 through early August 1976. I would like to briefly discuss the recovery patterns which have occurred during the first twelve months following the fire. Grass-Shrubland Communities Six sixty-meter line transects were established in six separate grass-shrub locations within the burn area. Within these six sites there are eight commonly occurringgrasses: three native species, Deschampsia austrazis, Eragrostis atropioides, Eragrostis variabizis; and five exotic species, Microlaena stipoides , PaspaZum dilataturn, Holcus Zanatus, SporoboZus afr"icanus,and Anthoranthum odoratum. The most common shrub species occurring within these transects are Dodonaea sandwicensis and StypheZia tameiameiae. The introduced shrub Jerusalem cherry, Sotanum pseudocapsicum, is also a prominent member of many of the communities affected by the fire. In all six of the shrub-grassland plots, ground cover was reduced to less than five percent immediately following the fire. The only vegetation left standing after the fire were the charred remains of SCypheZia and Dodonaea. Changes after six months. During the first six-month period -fillowing the fire, two of the predominant inhabitants of the grass-shrub communities were exbtic herbaceous species, Hypochaeris radieata and PZantago ZanceoZata; these plus six other herbaceous species accounted for 27.5 percent of the total ground cover after six months. The dominant component of these communities was the small fern Pteridium aquilinum var. decompositum which accounted for 7.5 percent of the total cover. Grass species, both native and exotic, did not begin to reappear until early January 1976, and by March 1976 only eight species had become established within the six grass-shrubland transects. After six months, three endemic grasses, Deschampsia and the two Eragrostis species, accounted for 4.2 percent of the total ground cover. The five species of exotic grasses accounted for 8.5 percent of the cover. Of the three endemic grass species Eragrostis appeared to respond most favorably during the first six months. By March Eragrostis accounted for 81 percent of the endemic grasses in the shrub-grassland plots. By far the most successful grass to reestablish itself within the first six months in the grass-shrub plots was Paspalurn diZatatum. By March 1976 it accounted for 5.6 percent of the total ground cover and 45 percent of the grass cover in the six grass- shrubland communities. Holeus Zanatus, one of the most common exotic grass species in the savannah-grassland and mountain parkland regions, showed only marginal success in reestablishing itself soon after the fire. Six months after the fire HoZcus occupied only 1.5 percent of the total cover in the grass-shrubland communities. One of the most discouraging results of the six-month vegetation tally was the lack of regeneration among native shrub species. During the March sampling very little evidence of regeneration of native shrubs could be detected. Dodonaea sp. seedlings were found scattered very sparsely on only three of the six grass-shrub transects. Not a single live specimen of Styphelia sp. was observed on any of the grass-shrubland transects during the March sampling. Six months after the fire 47 percent of the grass-shrubland transects were devoid of vegetative cover while litter occupied 16 percent of the ground in the grass-shrubland communities. Changes after one year. Cover percentage for all dominant species, both grasses and shrubs, showed no major change at the end of one year's sampling, although small increases in cover percentages were noted for each of the eight previously occurring grasses. Paspalum dilatatum exhibited the greatest increase in cover. PaspaZum accounted for 11.2 percent of the total cover, followed by HoZcus Zanatus (5.1 percent), at the end of one year's sampling. The two species of native Eragrostis increased in cover from 0.8 percent in March to 1.8 percent in August 1976. A slight competitive advantage was seen in the exotic grasses, which at the end of one year's sampling accounted for 67% of the grass cover in the shrub-grassland plots, while the endemic grasses showed a 4 percent decline accounting for 32 percent of all grass cover in August 1976 as compared to 36 percent in March. A slight increase in plant litter, some 11 percent, from 6 to 17 percent, partially explains a reduction in the amount of barren soil seen in the grass-shrubland plots at the end of one year, from 47 percent in March to 36 percent in August 1976. Conspicuously absent at the end of one year's sampling again are the native shrub species. Dodonaea sp., although visibly present on all six of the shrubland plots, accounted for less than 0.5 percent of the total plant cover. Newly emergent sprouts of StypheZia sp., the other dominant shrub in these communities, were observed on only one transect during the August 1976 sampling.

Understory Regeneration in Acacia koa Communities A somewhat different pattern of understory regeneration is seen within the Acacia koa colonies. The increased shading of the understory by the remains of the Acacia canopy accounts in part for the differences in species composition that can be observed within the koa plots, as compared to the grass-shrubland communities.

Three exotic grass species, Anthoxanthum odoratum, HoZcus Zanatus, and Paspalum diZatatum have established themselves as the dominant understory components within the koa communities. Six months after the fire, these three species accounted for 21 percent of the total ground cover. The single most abundant species after six months, however, was Pteridium aquilinum which occupied almost 11 percent of the total cover. The three native grasses occurring in these plots, Deschampsia austratis and the two Eragrostis species, did not appear to do as well in the partially shaded koa plots as they did in the full sunlight of the grass-shrubland communities. These three species accounted for only 0.5 percent of the total cover in the koa communities as compared to 36 percent in the grass-shrubland plots after six months. Barren ground and plant litter together occupied 53 percent of the surface area within the koa plots, with cover values of 25 and 28 percent respectively. Species composition in the koa understory revealed few changes after the second six-month sampling period, with only two new plant species occurring within the koa stands. One of the two newly emergent species, Eragrostis grandis, an endemic grass, was observed along two of the Aoacia koa transects during June, and by August 1976 Eragrostis accounted for slightly over one percent of the total cover inside the koa plots. Rubus penetrans accounted for 5.5 percent of the plant cover inside the koa plots during the August 1976 sampling. Growth of this noxious exotic will be monitored carefully in the coming months with hopes that its spread will be contained by natural successional trends. Two of the exotic grasses occurring within the koa communities, HoZcus tanatus and Paspalum ditata~tum,both showed moderate increases in cover percentages after one year's sampling, with Paspalum becoming the dominant understory component inside the koa colonies. Together these two species comprised 32 percent of the total ground cover in the koa communities, one year after the fire, with Paspalunr alone accounting for 19 percent of the ground cover.

Acacia koa Regeneration Regeneration of Acacia koa was first noted in early November 1975 on transect $14 at 4700 feet. Several small, newly emergent root sprouts were observed on the outer edge of the transect. Abundant regeneration of Acacia koa in the other two Acacia study plots was not observed until late December 1975. Koa regeneration within the study plots was noted to be occurring almost exclusively by root sprouting. Fewer than 20 Acacia koa seedlings were tallied in the plots during the first twelve months of recovery following the fire. Newly emergent koa root sprouts accounted for only 0.5 percent of the ground cover six months after the fire, but encouragingly, koa sprouts increased their cover value by some 5.5 percent in the second six-month period, occupying 6.2 percent of the sampled area at the end of one year. Preliminary results of this study indicate that Acacia koa responds quite favorably to disturbance by grass and brush fires. Fires of moderate intensity, such as this one, do not appear to severely damage the root systems of Acacia koa, although most of the trees suffered extensive damage to the trunk and crown portions. While cover values for koa do not indicate it, sizeable increases in area for each of the colony stands have been noted. Much of the understory was removed in the areas adjacent to colony stands of koa, which has allowed numerous new root sprouts to emerge outside the previous perimeter of the old colonies. With little initial competition from surrounding vegetation, these new root sprouts appear to be successful in extending the area of the colony stands. Summary After one year's sampling of the three communities affected by the August 1975 fire, it appears that a large of the floral components of the mountain parkland- majoritxsavanna grassland ecosystems are favorably adapted to disturbance by fire, Species composition within these two ecosystems does not appear to have been appreciably altered by the fire, although the native shrubs in these ecosystems are responding at a much slower rate than the therophytes or other phanerophytes in this region. This slow regeneration response by the native shrubs is not altogether unexpected as regeneration of these species occurs almost exclusively by seed propagation. The moderately high heat created by the fire combined with the high heat-conducting capabilities of the fine ash soil destroyed virtually all of the viable seed from the shrubs within the burned area. It is anticipated that an increased rate of regeneration of native shrubs will be documented in the following year as new viable seed from unburned surrounding sources invades the burned area. Appendix--Species List Checklist of species by families. Citation names follow Fosberg (1966). X = exotic species, E = endemic species, and I = indigenous species.

COMPOSITAE Cirsium vuZgare (Savi) Tenore Conyza canadensis (L.) Cronq. GnaphaZium purpureum L. Hypochaeris radicata L. Sonchus oZeraceus L. Senecio sy Zvaticus L. CYPERACEAE Carex macZoviana D'Urv. var. subfusca (W. Boott) Kuek. Carex wahuensis C.A. Mey. var. meyeri Franch. & Savat . DENNSTAEDTIACEM Pteridium aquizinum var. decompositum (Gaud.) Tryon EPACRIDACEAE StypheZia tameiameiae (Cham.) F. Muell.

Vaccinium reticulatum Sm. GENTIANACEAE Centarium umbel Zatum Gilib.

GERANIACEAE Geranium caroZinianum var. austraZe (Benth.) Fosb. GRAMINEAE Anthoxanthum odoratum L.

Andropogon virginicus L. Bromus rigidus Roth Bromus catharticus Vahl Cynodon dacty Zon (L.) Pers. DactyZis gZomerata L. Deschampsia australis Nees ex Steud. Eragrostis atropioides Hbd. Eragrostis grandis Hbd. var. grandis GRAMINEAE Eragrostis variabilis (Gaud.) Hbd. E Festuca megalura Nutt. X

Microlaena stipoides (Labill.) R. Br. X

Panicum tenuifolium H. & A. E

PaspaZum dilatatum Poir. X

Sporobolus africanus (Poir.) X Robyns & Tournay

IRIDACEAE Sisyrinchium acre Mann JUNCACEAE Juncus bufonius L. LEGUMINOSAE Acacia koa var. hawaiiensis Rock LYTHRACEAE Lythrum maritimum HBK. ONAGRACEAE EpiZobium oligodontum Haussk.

OXALIDACEAE Oxalis corniculata L. POLYGONACEAE Rumex acetoseZZa L.

PRIMULACEAE AnagaZlis arrensis L. SOLANACEAE SoZanum nigrum L.

SAPINDACEAE Dodonaea sandwicensis Sherff RUB IACEAE Coprosma ernodeoides Gray E Literature Cited*

Fosberg, F.R. 1966. Vascular plants. Pages 153-238 in M.S. Doty and D. Mueller-Dombois, Atlas for bioecology studies in Hawaii Volcanoes National Park. College of Trop. Agric., Hawaii Agric. Exp. Station, Misc. Pub. 1/89. Mueller-Dombois, D. 1966. The vegetation map and vegetation profiles. Pages 391-441 in M.S. Doty and D. Mueller-Dombois, Atlas for bioecology studies in Hawaii Volcanoes National Park. Coll. of Trop. Agric., Hawaii Agric. Exp. Station, Misc. Pub. #89. Mueller-Dombois, D., and H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley, N.Y. 754 pp. Rotar, P. P. 1968. Grasses of Hawaii. Univ. of Hawaii Press, Honolulu. 355 pp. St. John, H. 1973. List and summary of the flowering plants in the Hawaiian Islands. Pacific Tropical Botanical Garden, Memoir Number 1, Lawai, Kauai, Hawaii. 519 pp. Spatz, G., and D. Mueller-Dombois. 1975. Succession patterns after pig digging in grassland communities on Mauna Loa. Phytocoenologia 3:346-373. Vogl, R.J. 1974. Effects of fire on grasslands. Pages 139-194in T.T. Kozlowski and C.E. Ahlgren, eds. Fire and ecosystems. Academic Press Inc., N.Y. Whitney, L.D.,E.Y. Hosaka, and J.C. Ripperton. 1939. Grasses of the Hawaiian ranges. Hawaii Agric. Exp. Station Bull. No. 82, Univ. of Hawaii, Honolulu. 141 pp.

* Includes literature cited for previous paper, "Hilina Pali Fire of 1975," by same author. HILINA PAL1 FIFE OF 1975 Terry T. Parman Hawaii Volcanoes National Park Hawaii 96718

During May of 1975, a fire control exercise was conducted by the staff of Hawaii Volcanoes National Park is order to train and familiarize Park personnel with current fire suppression and control techniques.

A small section (ca. 44,000 sq. ft.) of open canopy 7 Ohi'a- forest with a native shrub-mixed grass understory, located along the Hilina Pali Road,was selected as the site for the controlled burn. Prior to the burning of the area, a three-foot wide fire break was cleared surrounding the one acre burn site to prevent accidental spread of the fire. Vegetation sampling was conducted before the burn in order to determine species composition and cover percentages. Sampling was conducted on two sixty-meter line transects using the point frequency method. Point sampling was done using finely sharpened bronze rods mounted in a one-meter long wooden frame. Five points per meter,at twenty centimeter intervals,were sampled for each of the transects so that 600 total points were recorded inside the study area. Point sampling along the same two transects was carried out at six month intervals after the fire in order to assess the regeneration and succession patterns. The vegetation along the first few miles of the Hilina Pali Road is typically very open canopy 'bhi'a forest. The understory is composed of four species of native shrubs, Vaceinium caZyeinum (lohelo), StypheZia tameiameiae (Piikiawe), Dodonaea sandwicensis ('A'ali'i) , and Dubautia ciZioZata (na'ena'e). Herbaceous ground cover in this area consists primarily of three introduced grasses: Andropogon virginicus, Andropogon glomeratus, and MeZinis minutiflora. Scattered clumps of BuZbostyZis capiZZaris and Carez wahuensis can also be found within this community. The understory vegetation, before the burn, was characterized primarily by Andropogon spp. and the four native shrub species. Andropogon spp. showed the highest cover value for any single species, accounting for 21 percent of the total understory cover. Taken as a group the native shrubs StypheZia, Dodonaea, Vaccinium, and Dubautia comprised 21.6 percent of the understory cover before the fire. StypheZia was the dominant shrub species accounting for 12 percent of the cover; 29 percent of the ground cover was litter. Barren ground, both rocks and soil, occupied 15 percent of the study area previous to the fire. Recovery of vegetation in this particular community is generally slow to occur due in part to the patchy seasonal distribution of rainfall. Seasonal rainfall is typically low, with the majority of it occurring during the late winter and early spring (February-March). Since the area was burned in early May, after most of the year's rain had fallen, regeneration six months after the fire was hardly noticeable. The November 1975 sampling showed vegetative cover to be occupying only 12 percent of the burned area. Andropogon spp. comprised the largest portion of new cover with 4 percent. BuZbostyZis covered 2.7 percent of the area and Melinis sp. accounted for a scant 0.4 percent. Not one newly emergent sprout of any of the native shrub species could be found in November. Likewise, of the 216 'ohi'a (Metrosideros coZZina) trees burned during the fire, not a single tree showed any sign of regeneration. The most striking feature about this area in November 1975 was the large portion of barren ground; 81 percent of the burned area was void of live vegetation six months after the fire. The August 1976 sampling period (15 months after the fire) showed several noteworthy changes. While plant cover and diversity did not increase dramatically, a major alteration within the structure of the community could be observed. The once-domina.nt native shrubs had begun to be displaced by the two introduced Andropogon species. Dubautia, StypheZia, and iraccinium, while present, accounted for less than one percent of the total cover. Andropogon spp., on the other hand, exhibited the largest increase in cover, jumping to 23.6 percent. The third introduced grass, MeZinis minutiflora, increased in cover by 2 percent, accounting for 2.2 percent of the cover in August 1976. Barren ground decreased in August 1976 to 47 percent,down from 81 percent in November 1975. Total live plant cover increased from 12 percent in November 1975 to 54 percent in August 1976. The most encouraging aspect of the August 1976 sampling was the large increase in 'Ohi'a regeneration; 102 out of 216 (47 percent) 'bhi'a trees showed some degree of basal regeneration. It will be interesting to note the success of these new root suckers, as it appears that they will be competing heavily with the increasing stands of Andropogon for both light and moisture. F'ifteen months after the fire one major trend appears to have emerged within the succession pattern of this open forest community. While it seems that all of the native species occurring in this region show some ability to with- stand and respond to perturbation by fire, several exotics, in particular the two Andropogon species, are very well adapted to and favored by disturbance by fire. Structural alteration of the community (i.e., displacement of native shrub species by exotic grasses) is very evident. The mean height of Andropogon in this area has almost doubled since the fire and the density of Andropogon also appears to have greatly increased. Preliminary results of this study indicate that Andropogon is an extremely aggressive invader of fire-disturbed mixed grass and shrub communities. It is therefore recormended that in order to best preserve and protect the remaining native plant species in communities such as thiswhereAndropogon coexists with native shrubs and trees, that every effort be made by Park Service personnel to quickly suppress and extinguish any fire which might occur there, be it accidentally or naturally caused.

Appendix--Species List

COMPOSITAE Hypochaeris radicata L.

Dubautia ci ZioZata (DC.) Keck CYPERACEAE Bulbosty Zis capiZZaris (L.) C.B. Clarke Carex wahuensis C.A. Mey. EFACRIDACEAE StypheZia tameiameiae (Cham.) F. Muell. ERICACEAE Vaccinium caZycinum Sm. GRAMINEAE Andropogon gZomeratus (~alt.)BSP. Andropogon virginicus L. Melinis minutifzora Beauv.

LILIACEAE ~ianezlasandwicensis H. & A. MYRICACEAE Myrica faya Ait. RUBIACEAE Coprosma ernodeoides Gray SAPINDACEAE Dodonaea sandwicensis Sherff

MYRTACEAE Metrosideros coZZina (J.R.& G. Forst.) Gray subsp. poZymorpha (Gaud.) Rock

Literature Cited See Literature Cited in next paper by same author, "The Effects of Fire upon a Hawaiian Montane Ecosystem." REESTABLISHMENT OF NATIVE FLORA AND FAUNA IN HAWAII VOLCANOES NATIONAL PARK D. Reeser Hawaii Volcanoes National Park Hawaii 96718 Hawaii Volcanoes National Park is the largest single land unit in Hawai'i whose management policy is primarily devoted to the perpetuation and maintenance of the primitive Hawaiian scene. Objectives guiding the management of the park's biological resources pertain to the preservation and restoration of native ecosystems. A resources management program has been developed to favor native ecosystems and minimize the effect of non-native invaders. Guidelines for this program are set forth in the park resources management plan. Relative to this plan, I would like to address three aspects of resources management: 1) what is the goal of resources management in Hawaii Volcanoes National Park, 2) what has been the progress since the park was established, and 3) what programs are underway or planned. What is the goal: We often speak of "preserving and restoring native ecosvstems" as being the ~rimarvDurDose or obiective. .. A But, exactly what does thig mean? Everyone here probably attaches a little different meaning to this concept, depending to some extent on his or her special interest. However, I think most would agree that Hawaii Volcanoes should be a natural preserve wsere ecosystems are maintained as nearly native as possible. Where the influence of non-native organisms is minor. Where native nlant communities flourish, supporting a broad array of native insect and birdlife. Where species native to the region are not unnaturally rare or threatened with extinction. Where natural processes associated with native biota are operating as free as possible from man's direct or indirect influence. This is the goal. Unfortunately, at this point in time the park does not meet these criteria, hence the need for a resource management program. Some persons might say that this goal can in reality never be achieved in light of the many adverse factors operating against native ecosystems. While this may be true, does this mean we should abandon the goal? Certainly not! For if we can come anywhere near this goal, the effort will have been worthwhile. Because in so doing at least a semblance of Hawai!ils unique native resources will have been saved and maintained. The importance of these resources will continue to grow as, almost inevitably, native biota declines elsewhere in Hawai'i. .. What has been the progress since the park was established: !Rie park waiZ%tablished in 1916. Through the years land additions have been made and today it is about 220,000 acres in size, ranging from over 13,000-feetelevation to sea level grading from rain forest to desert. The area which became Hawaii Volcanoes National Park was far from being a pristine Hawaiian scene in 1916. At least 75 years of destruction by cattle, goats, pigs, non-native birds and insects had gone before. Even Bird Park, considered to be a botanical treasure, was lacking understory vegetation as a result of extensive browsing by cattle. It is not unreasonable to assume that native plants, birds, and insects had already disappeared; some perhaps even became extinct. We can never know what the original native character of the area was. Under park management the area gradually began to change to the benefit of native flora and fauna. Cattle were removed from Bird Park; grazing permits were eventually discontinued on the Mauna Loa Strip. Some of the rarer native plants were propagated. HibiscadeZphus giffardianus was saved from extinction. And in the early 1970's an effective goat reduction began. These were the highlights of resources management during the last 60 years, but even with this L believe that native ecosystems throughout the park in general continued on a downward trend. A truly effective program was slow in coming. I sense now, however, that the tide has begun to change and indications are that progress towards the goal is going to accelerate. What programs are underway or planned: Continued goat control and completion of boundary fences is still the number one priority: The population has gone from an estimated 15,000 in 1970 to 300 today. As remaining segments of boundary fence are completed, this number should be further reduced. Feral pigs pose a major problem. Damage has certainly been alleviaeed in accessible areas but there are thousands of acres still receiving incredible abuse from these animals. Now that goat control and related fence construction becomes less of a burden, more attention is going to be focused on pigs. Presently we are experimenting with control methods to determine how we can best wage an effective and long lasting reduction program. Next, there are several exotic plants that we must keep from getting out of hand. These include: fire tree (Myrica faya Aft. ) , Zkoa or koa-haole (~eucaenaZeucocephaZa [Lam.I de Wit) , silky oak (GreviZZea robusta A.Cunn. in R.Br.), Christmas berry (Schinus terebinthifo Zius Raddi) , banana poka (~assifZoramoZZissima [HBR.] Bailey), and a few others deemed to be highly competitive with dominant native ecosystem species. Some work has been accomplished through tho years but beginning: this year this program will be substantially increased. There are over 400 exotic plants inhabitating park ecosystems Much research is needed to tell us which pose problems to the native ecosystems and what techniques are best for their control or elimination. The program I have discussed so far is aimed at extracting grossly interfering elements from the ecosystem. In many areas of the park, this is perhaps the only type of resource management action necessary. However, in some areas-- particularly the coastal regions from 'Aina-hou to Ka'G boundary, which have suffered extensive damage and alteration --merely removing the non-native elements is not enough. Therefore the park has developed a native plant restocking program. The purpose of this is not- to create a large botanical garden, but rather to do the following: Provide native plant competitors for exotic plants. This is a technique for the control of exotic plants which is often more effective than mechanical removal Without goats, natives can often compete successfully with exotics if they are given a helping hand. As a result of past goat and cattle devastation, many plants once common are rare. Some of these need attention where there is little likelihood that they will reestablish themselves on their own. Even with goats gone, some are yet ravaged by insects and rats and until programs can reverse their continuing decline the park must artificially supplement the natural population. Planting can be useful in restoration of habitat for native birds. It is presumed that the decline of certain keyplant species have beeninstrumental in the disappearance of native birds. We are concerned at this time primarily with the 'Alal; and NSnE but perhaps further research will pinpoint other birds which can be assisted in this way. The reintroduction of plar~ts into their former range will not be done on a wholesale basis. Certain restrictions are included in the resources management plan so that hybridization does not occur and genetic pools are not altered: 1. Only species known to have or strongly suspected of having occurred in the park will be used. 2. Propagating material sources must come from the nearest stock available.

3. Complete records must be kept and plantings must be periodically monitored to gauge success of the program 4. It would also be valuable if a biological inventory could be made of each planting area for comparison studies in future years. It would be important to know how plantings influenced native insects, birds, and how they compete with exotic plants. The concept of reintroduction of plants also holds true for some bird species. NCnC formerly occupied the lowland habitat, migrating there from the uplands for breeding purposes. It is unlikely that NGnS would reoccupy these areas on their own even with the best vegetation recovery plan. We now have eight breeding enclosures from 'Kina-hou to KG-ka-lau-'ula. There is much to be learned from this project, such as ways to control mongooses and feral cats and NkS! requirements for native food plants. We have high hopes that a substantial NSnE population can be reestablished to survive and perpetuate itself in the wild. Another bird formerlyoccurringin the park is the 'AlalZ (Hawaiian crow). Many concerned people feel this may well the next species to be added to Hawai'i's long list of extinct birds. We have recently prepared a captive-breeding- reintroduction program and are prepared to commit park resources to implement it. We are presently seeking U. S. Fish and Wildlife Service approval. It is an understatement to say that we don't have all the answers. An enormous amounf of management-oriented research is needed. It is encoura~ing- - however to note that many such studies are now underway, several of which are being reported on at this conference, These will provide important information on which to base future management decisions. MMANE FOREST DECLINE ON MAUNA KEA: A REALITY OR MYTH' Paul G. Scowcroft U.S. Forest Service Pacific Southwest Forest and Range Experiment Station Institute of Pacific Islands Forestry Honolulu, Hawaii 96813

Management policy and objectives for the msmane (Sophora chrysophy22a)-naio (Myoporum sandwicsnse) forest on Mauna Kea have been debated with increasing fervor in recent years. The issue in some segments of the populace and scientific community has taken on an emotional flavor. Simply stated. the central issue is, "should a huntable population of feral sheep (Ovis aries), i.e. 1000 animals, be maintained on the mountain?" One side answers no. These people believe that feral sheep in huntable numbers are detrimental to native vegetation and in turn to birds,and therefore should not be tolerated. Research to date supports their belief that sheep are a major detriment to the ecosystem. The other side answers yes, 1000 sheep should be maintained because sheep are an important source of meat for their families. They also regard sheep hunting on Mauna Kea as an important part of their cultural heritage. This group expresses the belief that the mzmane-naio forest on Mauna Kea is not declining, but is in fact regenerating itself quite well. They point to specific areas which they say were devoid of almost all ground cover 20 years ago, but which today support an abundance of plants including established mzmane reproduction. For the past 5 years, the U.S. Forest Service, in cooperation with the Hawaii Department of Land and Natural Resources, has been conducting research on Mauna Kea to gather quantitative data about changes in vegetation in the mzmane- naio forest ecosystem. This paper briefly discusses the preliminary results of a study to determine changes in tree canopy cover and number of trees between 1954 and 1975 using aerial photography. Changes were thought to be correlated with intensity of sheep browsing.

Study Area The study area, located on the western flank of the mountain,is about 12,000 acres in size and includes land between 5600 and 10,000 feet elevation (Fig. 1). About one-fourth scale -5 miles

Figure., 1. Located on the western flank of Mauna Kea. the 12,000 acre study area includes portions of the Mauna Kea and Ka-'ohe Game Management Areas and extends into cattle pasture in the northwest corner. Figure 2. Tree canopy cover classes, elevational strata, and land-use strata used in the analysis of cover change.

Tree canopy cover classes: 0 percent 1 - 25 percent 26 - 50 percent 51 - 75 percent 76 - 100 percent

Elevational strata:

> 9000 feet 8001 - 9000 feet 7001 - 8000 feet 6001 - 7000 feet < 6000 feet

Land-use strata: Cattle pasture Game Management Area (GMA) Mauna Kea Forest Reserve-Game Management Area Ka-bhe Game Management Area of the area is cattle pasture while the rest is within the Mauna Kea and Ka-'ohe Game Management Areas. The forest at lower elevations is predominately dense naio with some m5mane intermixed. As elevation increases, mgmane becomes the predominant species and the forest becomes more open. Pure mzmane stands occur near tree line and at lower elevations in the northern one-third of the study areas.

Methods To determine changes in tree canopy cover and tree numbers, the area was sampled using 927 one-half-acre circular plots (4 percent sample) systematically established on aerial photos taken -in 1954 and 1975. The scale of the 1975 color imagery (1:12000 nominal) was about 3.5 times larger than that for the 1954 B/W prints (1:40000 nominal). Thus, smaller tree crowns were visible and, therefore, measurable on the 1975 photos. Clear acetate overlays showing plot centers and perimeters were made for each scale encountered. Dot grids were used to facilitate cover estimates. Tree cover was defined as the proportion of ground in the plot hidden from overhead view by the foliage of live trees.

Preliminary Results Early in the photo interpretation, doubts were raised about the feasibility of tallying individual crowns, especially in the naio-mzmane areas where trees tended to form clumps. When our doubts were substantiatedby counts obtained in the field, we abandoned number of trees as a means of quantifying change. The tree cainopy cover data and conclusions reported in this paper are preliminary. Research is continuing using several cover estimating techniques in addition to the one originally used and reported here. In the future, much of the problem of accuracy may be solved by improving the quality of the photography. Color infra-red imagery may be the answer. Taken from aircraft flying relatively close to the ground, color infra-red pictures offer high resolution and clarity and outstanding contrast. Preliminary investigations indicate that we may be able to consistently differentiate between m5mane and naio on such photographs. Changes in canopy cover were analyzed using grouped and .ungrouped data stratified by elevation and land-use (Fig. 2). Plots were grouped into one of five cover classes. The two land-use strata were cattle pasture and Game Management Area (GMA) The latter included portions of the Mauna Kea and Ka-'ohe GMA's, both of which contain free-roaming herds of feral sheep.

I first determined the number of plots in the pasture area and GMA that had increased, decreased, or not changed by at least -one cover class. The GMA had a significantly (0.01 level) greater proportion of plots that changed by at least one cover class.' About 35 percent had changed compared to 21 percent of the cattle pasture plots. The significance was primarily due to increases in cover class rather than decreases. The GMA had a significantly (0.01 level) greater proportion of plots that increased by at least one cover class than did the pasture area. Differences in the proportion of plots that decreased in cover were not significant (0.05 level) . I stratified the pasture plots by type of change and elevation (Table I), to see if changes in cover class were affected by elevation. The hypothesis that cover class changes are the same for all elevational strata was tested' and rejected (0.05 level). The proportion of plots that increased and decreased for the 6001- to 7000-foot stratum were different from the others. The reasons for the differences have yet to be determined. I also stratified the plots in the GMA by type of change and elevation (Table 2). The changes associated with the zone above 9000 feet were significantly different (0.01 level) from those recorded below 9000 feet.' Specifically, a much smaller proportion of plots had increased by at least one cover class above 9000 feet as compared to the proportions for lower strata. Interestingly, the proportion of plots that had decreased in cover class in the zone above 9000 feet was not significantly different (0.05 level) from the proportions encountered in lower elevational strata.

So far I have been discussing changes in cover classes. I now want to discuss change from the standpoint of average percent change per plot. In the pasture area, the net change was -2.1 percent per plot while that in the GMA was -1.2 percent (Table 3). These were not significantly different (0.05 level). I segregated the pasture and GMA plots into two additional categories: those that increased in cover and those that decreased. The difference between the average change for the pasture and the GMA was less than 4 percent in both cases (Table 3). Statistically, however, the differences were significant (0.01 level); e.g., the magnitude of the average change for plots that lost cover was signific significantly greater in the GMA than in the pasture. 'G-test of independence. Table 1. Percent of cattle pasture plots in each elevational stratum that increased, decreased, or did not change by at least one cover class during the period 1954-75.

Number Cattle pasture Elevation of plots Increase Decrease No change Table 2. Percent of Game Management Area plots in each elevational stratum that increased, decreased, or did not change by at least one cover class during the period 1954-75.

Number Game Management Area Elevation of plots Increase Decrease No change

I I >go00 146 8001-9000 260 7001-8000 224 6001-7000 33 56000 -

All 663

Table 3. Average change in percent tree canopy cover for cattle pasture and Game Management Area --number of plots in parentheses. Difference Level of Type of plot Pasture GMA /WGMA-%p/ Significance2 Lost cover -9.2 (129) -12.8 (270) 3.6 0.01 Gained cover 7.6 (86) 1 10.8 (245) 3.2 0.01 All plots1 -2.1 (264) -1.2 (663) 0.9 NS

f~ncludesplots for which no change was recorded. 'Determined from the Wilcoxon two-sample test. One other point about differences between the pasture and GMA is that the proportion of plots that lost cover was significantly greater on sheep range (GNA) than on cattle pasture (0.05 level).' No significance was detected for proportion of plots that increased in cover. Because the feral sheep range, i.e., the GMA, is the principal focus of concern, I looked at data from it in detail. As with the cover class analysis, I stratified the GMA plots by elevation to see if this variable affected average cover changes (Table 4). There was a significant effect (0.01 level). When all GMA plots were used in the calculations, the average change per plot above 9000 feet was -5.5 percent, over 5 times greater than the average change below 9000 feet (0.01 percent). No significant difference due to elevation was detected when only plots that gained cover were used in the analysis. The same was true when only plots that lost cover were used in the calculations. I found in the data additional evidence suggesting that elevation affects cover changes. In the zone above 8000 feet, 123 plots increased in cover (Fig. 3). But 160 decreased. Below 8000 feet, 122 plots increased and 110 decreased. In other words, more plots lost cover than gained above 8000 feet while the reverse was true below 8000 feet. This was even more pronounced when the zones above and below 9000 feet were compared (Fig. 4). Above 9000 feet, 52 more plots lost cover than gained. Below, 27 more plots gained than lost. Changes in tree cover of 25 percent or more are probably ecologically significant. So I segregated the data into two categories: those that changed by at least 25 percent and those that did not. Out of 927 plots, only 76 (8 percent) changed 25 percent or more. Of these, 61 were in the GMA. The average change was -12.0 percent per plot. For the 15 pasture plots the average change was -18.1 percent. These values were not significantly different (0.05 level).2 I again looked more closely at the effect of elevation on change within the GMA using only plots that changed by at least 25 percent. Above 8000 feet, the area most heavily utilized by sheep,3 the average change was -17.4 percent per plot (Table 5). The average change below 8000 feet was 'G-test of independence. 2~ilcoxontwo-sample test used. 3~tateof Hawaii. 1975. Ecology of the feral sheep on Mauna Kea. Final report, Pittman Robertson Project No. W-15-5, Study No. XI, 1972-1975. Department of Land and Natural Resources, Division of Fish and Game. Table 4. Average change in percent tree canopy cover by elevational strata within the Game Management Area--number of plots in parentheses.

Elevational Type of plot stratum Gained cover Lost cover All >go00 11.6 (8) -14.9 (60) -5.5 (146)

Level of significance2 I NS 1 0.01 I 'Includesplots for which no change was recorded. 2~asedon the Kruskal-Wallis test.

Table 5. Average change in percent cover within the Game Management Area above and below 8000 feet elevation by type of plot using only plots which changed by at least 25 percent-- number of plots in parentheses.

Elevational Type of plot stratum Gained cover 1 Lost cover I A11

Level of significance1 NS NS NS

'Based on the Wilcoxon two-sample test. No change

~ainedcover

Lost cover

>80001 58000' Elevational strata

Figure 3--Number of plots that increased,

decreased, or did not change in cover

by elevational strata, >8000 feet and

58000 feet. C] NO change

1Gained cover

I2// Lost cover

>9000' 29000' Elevational strata

Figure &--Number of plots that increased,

decreased, or did not change in cover

by elevational strata, >go00 feet and

59000 feet. 3.2 percent. Although there was a difference of more than 20 percent between these mean values, they were not significantly different (0.05 level). The magnitudes ofthe average increase and decrease in cover above and below 8000 feet were not significantly different (0.05 level) either. But, the proportion of plots that had lost at least 25 percent of their tree cover was si nificantly greater (0.01 level) above 8000 feet than below. ? Conclusion Based on these results, it is apparent that changes in cover between 1954 and 1975 have been affected by land use and elevation. By way of summary, the important results are: 1. Change in cover has been more extensive in the GMA than in the pasture. 2. Within the GMA, the average loss of cover per plot was five times greater above 9000 feet than below. 3. Within the GMA, above 8000 feet, the number of plots that lost tree cover exceeded the number that gained, while below 8000 feet the number of plots that gained cover exceeded those that lost. 4. Within the GMA, the proportion of plots that lost at least 25 percent cover was greater above 8000 feet than below.

I believe it is important to relate these results to the question of decline of the mgmane-naio ecosystem. We can think of decline as the death of mature trees without replacement by younger ones of the same species. The data just presented suggest that more area above 8000 feet is losing mature trees or portions of their crowns than is gaining. If established msmane regeneration is also lacking in this area, then it is declining. Based on other studies I am conducting, I am inclined to say that established regeneration is lacking. Decline, in my opinion, is a reality. But it is not occurring over the whole ecosystem, only in areas frequently browsed by sheep. Further research is needed and already planned to get a more definitive grasp of the decline issue.

'G- test of independence. INTERPRETIVE TRAINING J. Shimoda City of Refuge National Historical Park Honaunau, Kona, Hawaii 96726 We have heard much this morning about research and natural research management. We'll be hearing more in the next couple of days, I'm sure. All of the "upper level" discussion is fine, but how do we reach the man on the city street and make him realize the importance of ecosystems, and research, and resource management? In our system of government, I believe it must be through interpretation, interpretation of the story. In interpretation the key words to me are: simplicity and motivation. It is only through awareness on the part of the public that we will be able to secure the commitment that will lead to action -- hopefully to preserve and manage wisely. Of my nineteen and a half years with the National Park Service, about twelve of them were spent in on-site interpretation, three and a half as an instructor for management and interpretation courses at our Stephen T. Mather Training Center in Harpers Ferry, West Virginia, and the last four yc:ars as a park manager involved in interpretive training for National Park Service employees in the State of Hawaii. In my years with the Park Service, too, I've been doing the two types of interpreting -- (1) interpreting, i.e. translating, for the American and Japanese delegates at the UJNR meetings, and (2) interpreting the park story and/or natural resource story. The Japanese have a word for translation, and they have a word for interpreting, just as they have a term for "hot", as in hot coffee, and "hot" as in chili pepper. Their word for interpretation is Kaisetsu, which trans- lates as explanation. But interpretation is not that simple. Today, the word to those of us who are involved in interpretation means much more than that. There is a book written by Freeman Tilden called Inter retin Our Herita e which is the best book in my opmlon+ on t e subject -7? or those of you who wish to dig into the subject of interpretation further. To me, interpretation means taking the facts that the scientistshave given us, and restating them in such a way that our park visitors will understand and be "turned on" toward protecting those resources, and become committed toward their wise management. Plutarch summed this up 2,000 years ago when he said, "The mind is a flame to be kindled; not a vessel to be filled."

What then, is the job of interpretive training? It is to fashion a Stradivarius out of a few pieces of wood. How- ever, let me hasten to add that the pieces must be of the right kind of wood. Let us assume then, that our selection process has provided us with a trainable interpreter. It does not matter whether he or she is easy to train or difficult to train, so long as the interpreter proves to be trainable in the end. Now that we have our guidelines in order, what then is the job that interpretive training has to do. First of all, an interpretive trainer must realize that the interpreter is the bridge between the scientist and the public. His job is to transmit accurately the data supplied by the scientist in laymen's terms. But how does the interpreter "turn people on?" It means teaching the interpreter to reach people's feelings; making a site or resource relevant to people, It means, -- don't just talk about it; whenever possible, take the people to the resource. In interpretation, the use of two or more of the five senses of seeing, hearing, touching, smelling and tasting by the visitor is very important. The more the better. Too many times he is confined only to hearing and maybe seeing. And to discuss the thing I call the sixth sense, interpreters must be taught to reach the people's feelings. Feeling is the one common denominator among all peoples of the world. It crosses language barriers. Who does not feel love, hate, warmth, jealousy, sorrow? People like to feel good and it disturbs them when someone or something threatens this "good feeling." Let me illustrate by reading something that was written by one of the trainees who at one time presented a straight interpretive talk without feelings. What I am going to read was his first attempt at reaching that sixth sense in the audience.

Once, when we were very young, the world stretched out so high and wide that we couldn't even see the other side. We were small and it was big. Our minds reached out through the world so zde. There was grass. There were tastes and things to touch. Wonder was ours. We wondered and learned. This sound bodes well and that one's Aunt Nell. There's a texture to mud and it squashes and swells. The snow is cold and grand- mother is old. Our world then was us, Just what we could touch and taste, see, hear, or smell. The world was high and wide and we were very small. Then we grew larger and the world expanded. Across the street and around the block. Down to the creek and over by the rock. We touched and ran and smelled and laughed. We wondered and learned of rabbits and mice and laughter and friends. The world was high and wide and full of wonderful things. We flexed our minds and flexed our muscles and grew wise and strong. We learned the how and we learned the why but the wonder slipped away. Responsibilities came and weighed us down and shoved the tadpoles out. Dignity grew, and who can squish in mud when there's dignity to do? One day we looked and we were big. Business and rush. Money and rush. Bosses and wives and husbands and rush. Hurry and pass and never look back. Too late to laugh. Too late to dance. Too late to jump and too late to wonder. There's no time now for a tadpole's twiggle or a little girl's giggle. So dance little child. Dance and sing. Laugh and steal my heart away. I can't go back. But your world's still wide and-full of grass and trees and ABC's. So dance little child, laugh and sing, down the hall ad across the creek to that wonderful secret place of frogs and water bugs and rabbit tracks and wonder. -- I wish -I had time to go.

by Lee Dalton-March 1972 Don't you think this person would act if he realized his sense of place was being threatened? Interpreters must also be taught to think in terms of wholeness or look at the whole story. They must see the total picture. Any interpreter worth his salt, I believe, must have the environmental message in mind. Otherwise he'd be tellin only a facet of the complete story. For example, the life cycle of the goats makes an interesting story but what about the envorinmental depredations? Many times interpreters merely entertain by telling partial stories with no real objectives in mind. A consistently good interpreter is a highly disciplined person. He understands objectives and outlining because he knows and believes what Seneca said nearly 2,000 years ago, "IE a man does not know what harbor he is sailing for, no wind is the right wind." And I might add a Shimoda corollary, "No amount of wind is the right wind." A good outlining prevents an interpreter's talk or program from meandering all over the countryside and aids in insuring a strong or memorable conclusion. I attended an interpretive talk once where the ranger spoke for 45 minutes and concluded his talk three times before he finally decided to quit for the night. My personal opinion is that a good interpreter can say anything worth saying in 20 minutes. Interpretive training must make an interpreter aware that without the visitor there is no need for an interpreter, just as without the student there is no need for a teacher. To carry this further, there is no need for an interpreter to fill his talk with factual garbage if there is no need for them in getting his message across to the audience. Too many times interpreters spout off facts just to show how smart they are to the confusion of the visitor or listener, and thereby fail to get their message across. Ideally, an interpreter must have a genuine liking for people. He will be at a definite advantage if he does. Interpretive training also aims to instill versatility in an interpreter as a communicator. There are a variety of methods or tools that are available to him. One is his voice, another are his hands and eyes, and still another is his appearance. The human voice is one of God's most wonderful gifts. If used properly it can make people happy or it can make people sad. We all know that if you combine the voice with the proper words, you can make people happy, you can get them excited, and you can instill fear in them. Andyet, when we get in front of an audience, many times we become formal, stiff and forget the voice's versatility. Speaking of the voice, volume is important, too. Remember to talk to the people in the back row. Next the hands. If one is reading a text as I am, the hands are pretty much tied down. However, hand gestures can be used even in reading situations and an interpreter must keep this in mind. In reading a text an interpreter has sacrificed much eye contact. He should make sure wherever possible that eye contact is direct, and not over the heads of the audience. An interpreter should also not favor one side of the audience in his eye contact. The third item, appearance, is equally as important for interpreters. In a park situation, a ranger's uniform seems appropriate, but this may not be true outside of the park. For example, it may not be appropriate when speaking to a militant group. At the same time, however, I would not dress in blue jeans and a buckskin jacket to show that I'm "in" with them, for the same reason that I would not attempt to speak in Pidgin Engllsh to a group of Hawaii youngsters because I know how ridiculous and forced it would sound. Sincerity in this case, I feel, is more important than attempting to put on a facade in our sorry attempt in trying td be accepted. Up to this point I've been discussing the lecture type of interpretive talk training, but there is also the lecture with aids -- visual and/or mechanical. Let me illustrate the use of a visual aid with this lemon (lemon demonstration). Interpreters must know the power of visual aids. They must also be aware of the use of pictures, flip charts, blackboards, slides, movies, demonstrations, overhead projectors, etc. An interpreter cannot belazy. If he is going to use aids such as a movie or slide projector, he must be professional about it. He cannot just "fly by the seat of his pants" and not check his slides or film, and equipment out in advance. There is nothing that ruins an interpretive program more than a slipshod operation where the interpreter is sloppy, where the movie film is threaded improperly or where the slides are poor, or are in upside down. Interpretive training looks at an interpreter from the point of view of, "Is he getting his message across?" rather than from the point of view of, "Is he using the King's English?"

And if he is trained to have the interest of his audience at heart and is trained in the proper techniques, he will be able to perform like Pagliacci, the clown, who turned in a good performance even though his son was dying at that very moment. Interpretive training can be broken down into attitudinal and skills training -- attitudinal in the sense of awareness of the importance of the resource, thereby instilling a desire to help protect it. Hopefully, the interpreter will not turn into a knight in shining armor who's sole mission will be to charge head-on into the dragon and to slay it. Hopefully he will use the strategy of awareness first. In the area of skills training the interpreter will.need to be trained in the art of looking at the total picture -- that he is stalking the goat, not merely for meat but also to make amends for the "evil," in quotation marks, that man has introduced to the native Hawaiian landscape. Following that he must be able to do a written outline whether he is going to give a talk, a talk with aids, preparing an interpretive publication, an interpretive sign system, building a guided or self-guiding trail, writing an audiovisual program, an audio script, a sound-slide program using a lapse-dissolve device, or making a movie. Beyond this the interpreter must be trained in the skills of comunications -- verbal and non-verbal. The voice, gestures, body control, choice of words and phrases, rate of speaking, diction. The voice I've already discussed. Gestures, including hand gestures, must conform with what the interpreter is saying. For example, he should not use a low hand gesture when he's talking about a high mountain. Perhaps every interpreter should be required to take hula lessons to master the art of gestures. Body control means that the interpreter must not sway from side to side or backward and forward as he is speaking. It also means that he must not shift from one foot to another. It also means that he should not play with his eyeglasses or fidget with his papers. Choice of words or phrases must also be made carefully when speaking before the public. They're very quick to take things out of context and the interpreter will lose his audience or be taken aback by an unexpected laughter, for a few minutes. Let me cite a couple of examples. There's a battlefield park, which will remain nameless, where there was a sign that said, "The general leaped up on the parapet and exposed himself." Another occurred as a passage in a handbook in a historical park which said, "General ~urgoyne's right and left flanks lay in the woods, but his front was open." Continuing on, rate of speaking arid diction are fairly obvious. They mean, for one thing, opening your mouth. Nervousness does tend to make an interpreter speak faster, but I don't know of a single case where a person can speak rapidly if he opens his mouth and pronounces his words carefully. And then, one word about diction. Good diction saves your vocal cords because you don't need to shout! In interpretive skills training, the use of a tape recorder and videotape is a must. No interpreter can deny what he did, for better or worse, because it's right on the tape or film. Besides, he can hear and see himself as others do. These are two very effective methods of feedback or evaluation for the interpreter. Their value cannot be overemphasized in interpretive training. But in the end, interpretive training can only point the way. It is up to the interpreter to work with his trainer to determine what techniques best suits his particular per- sonality. For example, they may find that he is not at all a lecture type speaker, but will be more effective in conducting team activities or problem-solving. The possibilities are many. Interpretive training is in a way like the wife in Japan. There is an old saying there that says, "The man is the head of the house, but it is the wife who is the neck which turns the head." FIFTEEN YEARS OF VEGETATION INVASION AND RECOVERY AFTER A VOLCANIC ERUPTION IN HAWAII Garrett A. Smathers National Park Science Center National Space Technology Laboratories Bay St. Louis, Mississippi 39520

A major volcanic eruption on the Island of Hawaii in November 1959 devastated an existing montane-rain-and- seasonal forest covering an area of about 500 ha. After the eruption, six new habitats were recognized by kinds of substrate and remains of former vegetation: Habitat 1. A lava lake consisting of massive pzhoehoe lava with many joint cracks in the KT-lau-ea Iki crater. Habitat 2. The summit and sides of the conder cone, Pu'u-pua'i. Habitat 3. A spatter area with dead tree snags. Habitat 4. A pumice area with dead tree snags Habitat 5. A pumice area with surviving trees. Habitat 6. A thin fallout area that covered a former partially cemented ash-crust habitat surface (Ka'ti Desert). A continuing study was made of plant invasion and recovery in permanent quadrats and transects from the time of the disturbance. The first report on the study was made 9 ears (1968) after the eruption (Smathers and Mueller-Dombois 19r4). Th e present report covers from the 9th year forward to year 15 (1974) after the eruption. In year 9 it was found that patterns of plant invasion and recovery observed depended on a number of specific factors or a combination of factors. In year 15 many of these same factors were still present, and individually or in combinations they were affecting the developing vegetation patterns. The recorded patterns of plant invasion and recovery are summarized as follows: 1. A continued directional progression of invasion on the floor of KT-lau-ea Iki (habitat 1) that correlated with cooling of the surface area. 2. Microhabitat development in major habitats affected connnunity development of synusia and aggregation. 3. Regression in earlier successional patterns were caused by feral pig activity. 4. Observations were made on relationships of exotic and native plants in the recovery and invasion process. 1. Factors related to directional invasion and recovery By year 9 an analysis of the directional invasion patterns revealed that: 1) the concentric inward advance of plant life on the KT-lau-ea crater floor was closely related to the substrate heat gradient that showed initial cooling at the crater floor margin; 2) the nearness of the seed source from the undisturbed forest was a major factor in the early invasion of habitat 3; 3) snag density accounted for accelerated invasion of habitats 3 and 4. The snags created favorable microhabitats by intercepting moisture which provided a moist soil at their bases for plant establishment; 4) ash-depth gradient correlated with both the survival and invasion of plants. The fast recovery in habitats 5 and 6 was attributed to the shallow ash deposit. In contrast plant life was still sparse on the deep open ash deposits of habitats 2 and 4 in year 15. Also the only additional directional invasion observed was on the ash deposits of habitat 4. Here the exotic weedy shrub Buddleja asiatica had completed its invasion from both the east and west side of the habitat. However, Buddleja could only get established at the snag bases, and in most instances the plants were dying or exhibiting low vigor. This species is being gradually replaced by the native low shrub Dubautia scabra. Concentric invasion of the KT-lau-ea Iki crater floor was continuing in year 15. Temperature measurements made by USGS through a drill hole in the lake surface in 1974 revealed that the interior of the lake was still molten rock at approximately 1075O. However, as the cooling process continued the crater floor had thickened from 30 m in year 9 to 40 m in year 15 (personal correspondence R. I Tiling. This thickening trend with a corresponding drop in surface temperature accounted for the advancement of plants into the center of the crater. Continued cooling of the lake interior will affect the invasion pattern of plant life for some time on the crater floor. As the interior cools, shifts in the crater can occur causing surface temperatures to rise in some localities. These shifts in temperature could explain the loss of some plants from several quadrats in the 1974 survey (year 15). Shifts of temperature in cooling volcanic materials is a common occurrence. It seems likely that it will be at least another decade before vascular plants will have arrived at the center of habitat 1.

2. Microhabitat development in major habitats Renewed analysis of the permanent transects and quadrats in year 15 revealed that community developments were taking place in some habitats through the development of plant aggregates. In habitat 2 an aggregate community of Vacciniurn retisuZatum was developing in a low depression on the lee side of the Pu'u-pua'i cinder cone. In year 9 plant community development had been observed to some degree in all habitats. On the crater floor (habitat I), several fern plants became established in the joint cracks. When one individual fern became established, it soon was joined by several other members. Also, the microhabitat conditions that favored this aggregation were not always present in other joint cracks since ferns did not get established there. Since other plants such as individual woody Metrosideros seedlings became established more in these cracks, and not associated with the ferns, it was reasoned that the controlling factor in individual and aggregate plant establishments was favorable moisture relations of the crack microhabitats. Thus, it seems reasonable to assume that the moisture conditions in the cinder cone depression were favorable for the establishment of the Vaccinium reticuzatum aggregate community. Further support is given by examining the upper profile of the Pu'u-pua'i cinder cone. The B layer, which is of a near impermeable flow sheet, acts as a hard pan that supports perched water for an indefinite period of time. It is thought that this B horizon extended widely over the cinder cone summit and sides; thus it could help bring better moisture relations to depressions such as the Vaccinium aggregate community.

3. Regression in year 9 successional patterns On local sites where feral pigs have scarified the developing soil, native plants have been destroyed and exoitcs have replaced them. The pig is not native to Hawaii. One form was introduced by the Hawaiians approximately 2,000 years ago. This animal has been "improved" by crossing it with the European domesticated pig. Since there are no native Hawaiian ungulates, the native plants do not have built-in mechanisms to cope with ungulate stress brought about by pigs, goats, and livestock. As a result large tracts of the native vegetation have been destroyed or seriously altered (Mueller-Dombois and Spatz, 1975). Park managers have tried unsuccessfully to eradicate pigs from the park since its inception (1916). As yet no definitive study has been completed on pig ecology. Such a study would provide park managers with better information for the control or eradication of this pest. The present study has shown that pigs engender the spread of exotic plants by their ground scarifying activities, but the most significant aspect of the study has been to prove that native plants can indeed competitively replace exotics (primarily Rubus penetrans, R. rosaefoZius), when substrate disturbing factors such as pig stress are removed. Mueller-Dombois and Spatz (1975) have also shown that when goat stress was removed, a new endemic legume vine (Cnnauolia kauensis) became established in two years. This was the first recording of this species in Hawai'i. Their study, like the present Devastation Area Study, revealed the high recovery potential of the native vegetation, once ungulate stress is removed.

4. Relationships-- of exotic and native plants in the. recovery and invasion process

The park manager of Hawaii Volcanoes National Park is charged to keep the park ecosystems in as near a pristine condition as possible. The invasion of exotics is looked upon as an unnatural phenomenon and a threat to the native fiora. Regardless of over a 50-year effort to eradicate exotic plants from the park, the majority of vascular species present are exotic (Smathers, 1968). Most of them are found on man-disturbed sites (road and trail sites, etc.) As was earlier noted, several exotic plant species such as Buddleja asintica, Rubus rosaefozius and Rubus penetrans .i.,cL- ..,-,- identified as early invaders in the recovery process in some habitats after the natural disturbance by ash fall. In the last several years an exotic low stature tree species, idyrica j'aga, has been observed to invade habitat 5, the pumice area with surviving trees. This plant, a native of the Azores and Canary Islands, was introduced in Hawai'i for reforestation. By 1944 it had succeeded so well that the Board of Agriculture and Forestry were pursuing a program to eradicate it (Neal, 1965). Although the state controls are still in effect, ~yricafaya continues to spread. By mid-1960 the plant was already established in the montane rain forest of the park and beginning to invade the seasonal dry forest. Over the past decade the park has extended and intensified its control of Myrica faya. Control is carried out by direct eradication (uprooting) and use of a herbicide (silvex). From 1967 to 1974, 62,776 Myrica faya trees were removed from the park (personal communication with former Park Superintendent Bryan Harry)--yet it continues to invade. A definitive study is needed to determine the ecological role of Edyrica falya in the Hawaiian islands. Such a study would provide the basis for the control of this species. General obse.rva;:i.ons tend to conclude that it rapidly invades most vegetation types in the park in great hordes as do many other exotic plants. It readily fills the open areas, especially in the montane rain and seasonal forests. On superficial view it appears to have replaced the native trees and shrubs, but on close inspection many native woody plants are still present. The means of occupying an open niche and the degree OE permanency of :hjrica faya must be evaluated. Study of the present populations of Myrica in habitat 5 and its possible invasion of the remaining habitat, if carefully monitored, can answer some important questions on the invasion potential and ecological adaptation of this species i.n this ecosys tern. This information could be undoubtedly USC~U%to understand the behavior of this exotic and its ?'.nteract:i.on with the nat;ve species ai~dother similar habi.tat:s. Literature Cited

Mueller-Dombois, D., and G. Spatz. 1975. The influence of feral goats on the lowland vegetation in Hawaii Volcanoes National Park. Phytocoenologia 3(1):1-29. Smathers, G.A. 1970. Proceedings of the Meeting of Research Scientists and Resource Managers, National Park Service, Albright Training Center, Office of Natural Science Studies Report. pp. 63-83. Smathers, G.A., and D. Mueller-Dombois. 1974. Invasion and recovery of vegetation after a volcanic eruption in Hawaii. National Park Service Sc. Monogr. Series 5. 129 pp. Neal. Marie C. 1965. In gardens of Hawaii. B.P. Bishop '~useumSpecial ~ublication50, Bishop Museum Press, Honolulu. 924 pp. THE CPSU UH C. W. Smith Department of Botany University of Hawaii at Manoa Honolulu, Hawaii 96822

On March 16, 1973, a Master Memorandum of understanding between the University of Hawaii and the National Park Service was signed which established the Cooperative National Park Resources Studies Unit at the University of Hawaii. The Unit is a jointly sponsored research unit whose function is to encourage and conduct biological research in Hawai'i's National Parks. The advantage for the National Park Service is the availability of qualified research personnel, well-versed in Hawaiian natural history, who can study the resources in Hawai'i's National Parks efficiently and effectively. The objective is to obtain information that will allow each Park to formulate and update its Resource Management Plan. For the University the advantages are the encouragement and support of biological research by faculty and students in the Parks. In this case students, normally encouraged to get as much supervised field experience as possible, are provided financial and logistic support which is generally the primary limitationon field work. The faculty also benefits by the continued support of their professional development. Thus the University benefits by the support and consequent upgrading of the graduate program in areas requiring extensive field work. Each National Park has a Resources Management Plan. This Plan is formulated from the Park Service experience within each Park. Within each Resource Management Plan, the various proposals identify where further information is needed. Each proposal requiring new action is then ranked in terms of its priority and an estimate made of the money and time needed to fulfill the proposal. The priority assigned to any proposal depends on a number of criteria, not least among which is the availability of funds and personnel. Each Park's resource management program is then evaluated at the State and Regional levels. The top priority items requiring research are then discussed with regard to their urgency, feasibility, and the availability of funds. A Resource Management Plan is not a static objective for a particular Park. It is updated and modified as necessary, that is, it changes with the requirements of the resource managers. However, there is a certain stability in the overall proposal, and programs once instituted are followed through to their expected conclusion. The CPSU UH Coordinating Committee is involved in this process to a limited extent in that we review management plans and proposals. At times, the CPSU UH Coordinating Committee (consisting of three entomologists, three botanists, one orn'thologist and the research biologist from Hawaii Volcanoes National Park) will identify resource management problems which it feels should be attended to; for example, the management of weeds. However, the role of the unit is to conduct research on subjects identified by the parks, and not to urge the parks to take care of specific problems. Individuals and groups should address their concerns directly to the park managers. Once a research program has reached the stage that it is of sufficient priority and funds are available, the CPSU UH is invited to find someone to investigate the problem. An initial estimate is made of the cost of the program, the availability of sufficient qualified people, and adequate facilities and equpiment. The Principal Investigator, who may be a faculty member or a graduate student, is asked to write a research proposal with a detailed budget. The research proposal is evaluated within the Unit and Graduate Division of the University before it is passed on to the National Park concerned, the State Director, and the Chief Scientist of the Western Region. At the Park, the proposal is checked to ensure that their resource management problem is being addressed and that it will provide the type of information that they need. At the State and Regional Offices the proposal is again assessed as meeting resource management needs but legal, budgetary and other criteria are also considered. I cannot overemphasize the fact that all research conducted by the Unit has as its primary objective the optimal management of some resource within the National Parks. At this point, if all requirements have been satisfied, a contract is drawn up between the Universi.ty and National Park Service which identifies the research proposal, and by implication the research program, which is to be sponsored. Once all the appropriate signatures have been obtained the CPSU UH administers the research program on behalf of the principal investigator; that is the Unit provides secretarial assistance, takes care of administration requirements (payroll, travel, supplies, etc.) and also ensures th.at the program is running smoothly and reports are received when needed. The Unit tries to relieve the researcher of all hassles so that the research program operates at its maximum efficiency. Most research programs are long-term and require several years for completion, yet the CPSU UH operates on an annually renewable contract. This apparent paradox is not a problem in reality. Assuming that the University and the National Park Service are satisfied with the overall program it is understood that once a research program has been contracted it will be continued through its conclusion on a funds available basis. No program would be prematurely terminated unless the researcher has failed to comply with the contract. I must stress the contractual nature of any research program conducted via the auspices of CPSU UH. There is a reporting schedule, every six months. These reports are an opportunity to briefly describe the highlights of progress to date and to identify problems experienced or anticipated. The report should normally be only two or three paragraphs long. However, as each segment of a research program is completed a technical report is required. This requirement is to fulfill the objectives of the research program; that is to give the Park resources managers the information they need as rapidly as possible. Later publication in an appropriate journal is also encouraged. Apart from the formal program, CPSU UH does have certain limited funds available to sponsor or assist researchers to work in Hawafi's National Parks. This assistance may be requested by one of the Parks which needs information, or broaden the coverage of an ongoing independent research program. The National Park Service is anxious to encourage responsible research within the parks as much as possible. Though the CPSU UH was organized in 1973, the program has taken some time to get into full operation. All programs have a certain latent period during which results are not forthcoming. Dr. Doty, Botany Dept., U.H. nursed the Unit through its formative years before handing it over to me last fall. I feel that now all major problems, misunder- standings, etc.,have been resolved and that information is becoming available rapidly and on schedule. The money available to support research is limited and there are many Parks competing for the available funds. I think it is true to say that Hawali's National Parks have done remarkably well in attracting the attention of the Regional Office to Support the research needed for their Resource Management Plans. 1 look forward to many years of fruitful cooperation between the University of Hawaii and the National Park Service during which we can develop a research program to provide the information for a resource management program that will allow our unique flora and fauna to flourish. C. W. Smith Botany Department University of Hawaii at Manoa Honolulu, Hawaii 96822

This report is a very severely summarized version of the work of Drs. Jack Beardsley, Frank Howarth, Sheila Conant, and Bill Hoe, Tissa and Ruki Herat, Robin Rice, Glenn Teves, Lani Stemrnermann, Paul Higashino, Bob Burkhart, Eddie Smith, and myself. Haleakala National Park is located on the island of Maui, the second largest island in the Hawaiian Archipelago. The National Park is subdivided into two districtsi the Crater District, the area under discussion; and the Ki-pahulu Valley District. The Crater District includes all areas of the National Park west of a line from Pshaku-pglaha, Kuiki, and the Koa Tree on the Eastern side of the Kau-p6 Gap. Compared with an area such as Hawaii Volcanoes National Park, Hale-a-ka-12 Crater has received little attention. This is particularly surprising considering the variety of ecosystems and its unique endemic flora and fauna. For example, there are several plants and animals unique to Hale-a-ka-la: Geranium arboreum, Geranium muZtifZorum, Artemisia mauiensis, SantaZum haZeakae, Stenogyne crenata, and depending on your taxonomy you may or may not consider the Hale-a-ka-15 Silversword as endemic to the Crater. Fourteen species of moss, eleven species of lichen, and in the insects, 33 of 143 endemic insects present in the Crater are unique to Hale-a-ka-1s. This listing is by no means complete since it does not include species from KI-pahulu or other areas of Hale-a-ka-15. The literature on Hale-a-ka-15 is sparse. Yocum (1967) in his report on feral goats in Hale-a-ka-15 Crater discussed the effects of the grazing pressure of these animals on the vegetation. Perkins (1896 and 1910) described the insects of Hale-a-ka-18. Zimmerman (1948) discusses the general flora and fauna of Hale-a-ka-18 in his introduction to the Insects of Hawaii. Beardsley has published several papers describing new species from Hale-a-ka-18 Crater. The birds are generally discussed in general reference to the Hawaiian avifauna. From the botanical viewpoint, Larson (1969), in an early Resource Management Plan for the Park, identified four major vegetation types: marginal rainforest, scrub, scrub-savannah, and barren areas. Vogl, Eenriclcson, and Forehand in 1970 and

"Contribution No. 01116 CPSU University of Hawaii. 1971 studied the Deschampsia grassland on the outer north face of the Crater. Finally, Kobayashi (1973) reported on his investigations on the ecology of the silversword. The above account essentially covers the major studies carried out in Hale-a-ka-l;?Crater. There are a number of other papers which include reports on Hale-a-ka-15.. The only exception to this generalization is the work on the taxonomy of the flowering plants which is relatively complete. The apparent disinterest in the Crater is without explanation. It is an accessible area but you cannot drive right up to any particular study area. This Resources Basic Inventory which should probably be more appropriately named a bioecological survey was initiated in June 1975 and will be concluded by April 1978. Its primary objective is to produce a list of the bryophytes and lichens, ferns, conifers and angiosperms, mammals, birds and insects that occur in Hale-a-ka-12 Crater. From these lists distribution maps (and general autecological statements) for indi-vidual species will be made. At the same time, general observations are made on the major vegetation units using a modified Braun-Blanquet method. These studies are preliminary to a vegetation map to be produced from field work initiated in June 1976 and scheduled for completion by October 1979. The principal method of study has been to work along seven transects; four in a north-south orientation: 1. Pu'u- nianiau to Red Hill; 2. 'Aina-hou to beyond H6lua Cabin; 3. Ka-palaoa through Pu'u-Maile to Waikau; 4. The Koa Tree through Pali-kti to the top of Kalapawili Ridge; the remaining three in an east-west orientation: 5. Along Kalapawili Ridge; 6. Pali-kc to Red Bill; 7. Pali-kc to Kala-haku. Each transect is samples approximately every thousand feet change in elevation or every mile, whichever is most appropriate. At each study site an area of about one acre is chosen which has a representative sample of the vegetation. This is a highly subjective choice and we fully realize that it can introduce a considerable bias into our evaluation. How- ever, this work will not produce the definitive statements on the vegetation of Hale-a-ka-18 Crater. It will provide some hypotheses that will be tested in the vegetation mapping program recently started. Should our study area be in a region where there is obviously more than one vegetation type present, each community is studied separately. During the course of our work, observations on management problems are also made. For example, estimates on pig and goat damage are recorded; the presence of noxious weeds in areas are also noted even though they may not be present in designated study area. Apart from the flowering plants and birds, species and organisms new to the Hawaiian Islands have been the rule rather than the exception, particularly in the mosses, lichens, and insects. Many of these species have simply been overlooked in previous studies. On the other hand, many of these groups have been collected by visitors with only a day or two to work in the area. However, the intensive systematic sampling technique used in this study has shown that many species are extremely localized in their distribution. For example, as far as we know to date: Andreaea rupestris is found in the Ka-palaoa Cabin Area and along the Kuiki Trail. Two species of Tortula are found only in the Pu'u Mamane--Ng-mana-o-ke-akua area. There are many other examples, all of which serve to emphasize the fact that the Hale-a-ka-18 Crater area has been studied in a very superficial manner at best. Because of the highly localized distribution of many species a reliable survey is dependent on a systematic approach that allows specialists to visit as many different areas as possible. A more serious problem in Hale-a-ka-1s is that the use of the Crater and surrounding areas as cattle ranch land during the past one-hundred years has seriously disrupted the native ecosystem. Until the early 1950's cattle grazed within the Crater. In fact, the National Park Service still does not have the grazing rights to certain areas though the rights have not been exercised for 25 years or more. Apart from cattle grazing, some areas such as the Kau-pE Gap were deliberately burned in an attempt to improve the pasture. The devastation caused by such activities is very difficult to assess now. However, the RBI team suspects that the eastern portion of the Crater, in particular, supported a very different vegetation compared to today's scrub. It is thought that previously there were extensive areas of open dryland forest of uncertain composition. Scattered remnants of Pittosporum, SantaZum, Sophora, Dodonaea, etc., suggest that these species were major components of the forest interspersed with Vaccinium and Styphelia. The damage done by cattle and the continuous suppression of regeneration by goats over the past one-hundred-and-fifty years has resulted in the establishment of a StypheZia-Vaccinium scrub with or without Sophora depending on the location. The damage done by feral pigs is relatively recent and generally confined to the peripheral areas of the Crater. Pigs come into the Park to forage from the surrounding forest The impact that they can have is amply demonstrated on Kuiki where the grassland which once occurred there is now reduced to lichen-covered rocks. The rooting by pigs has resulted in very severe erosion problems. In contrast to Larson (1969) we recognize six major vegetation types within the Crater: 1. Grassland. This vegetation type is normally confined to alluvial deposits, the ash aggregates at the center of cinder cones and to high elevations in the wetter regions. The controlling factors appear to be deep layers of porous soil and an adequate supply of water. In the pristine condition the dominant species is Deschampsia austraZis f. haZeakaZaensis but when disturbed by pigs, horses or humans ~oZcusZanatus and Dactylis gZomerata invade and frequently predominate in the area. 2. Rainforest. This community is found in the Crater at Pali-kc only. It is dominated by Metrosideros and some Ckeirodendron with various understory plants including SadZeria cyatheoides. In the shaded gullies, Metrosideros is absent and replaced by a mixture of PeZea, IZex, Myrsine, Psychotria, and Broussaisia. There are very few weeds in this vegetation type. 3. Koa Forest, but now a Koa Savannah. Though only represented by remnants, a major lcoa forest probably existed on the eastern flank of the Kau-p6 Gap. No regeneration of the community occurs because goats which occur here in very high numbers graze off all seedlings and suckers. The areas between the trees support grassland (predominately ~oZcus)and Eupatorium.

4. & 5. Scrub. This is the prevalent vegetation type inside the Crater and around the higher outer regions of Haleakala. However, within this type there are two different communities: one a climax community, the other a seral stage of the dryland forest climax. 4. The climax community. Thi.s community consists of Vaccinium, StypheZia, and Pteridium, with or without Sophora. It occurs in the drier regions of the Crater and. generally from 2,300111to 3,000m. It occurs at lower elevations in very dry climates, e.g.,West Kau-pB Gap and the Ko'olau Gap 5. The seral stage. This community occurs below 2,300m. From remnants scattered throughout the Crater, particularly the Pali-kO and eastern side of Kau-p6 regions, it is thought that the climax community would contain Sophora, Pittosporum, SantaZum, Myrsine, Euphorbia, and Dodonaea, with Styphe Zia, 0steome Zes and Vaccinium forming an understory. 6. Barren Areas. These are of two types. The high altitude (above 3,000m.) type contains scattered TetramoZopium humiZe, Dubautia (RaiZZardial menziesii, Agrostis sandwicensis, Trisetum glomeratum, and some Argyroxiphium macrocephaZum. In lower elevations (from 2,500m to 3,000m.) in low rainfall areas the substratum consists of very porous ash and plant growth is very limited. Here there are Argyroxiphium macrocephaZum, Dubautia, TetramoZopium, a species of SiZene, and Pteridium aquizinum var. decompositum scattered throughout the area. All of these areas except the rainforest are severely disturbed by goats. Pigs have a significant impact in the rainforest and scrubldryland forest areas. Even when the pressure from feral pigs and goats is removed, the dramatic revegetation as seen in the coastal lowlands of Hawaii Volcanoes National Park following goat population control will not be realized. Current exclosure studies in Haleakala Crater indicate that the rate of recovery will be extremely slow. This is not surprising. The work of Billings (1973) and many others have shown that arctic and alpine ecosystems have very slow rates of change. The sooner the feral herbivores are removed, the better. However, to conclude on a very dismal note, deer are seen in the Kau-pB Gap with increasing frequency in the lower and mid-elevations. Literature Cited

Billings, W. D. 1973. Arctic and alpine vegetations: similarities, differences, and susceptibility to disturbance. BioScience 23:697-704. Kobayashi, H. K. 1973. Ecology of the silversword Argyroxiphium sandwicense DC. (Compositae) Haleakala Crater, Maui. Ph.D. Thesis. Univ. of Hawaii. 91 pp. Larson, J. W. 1969. Haleakala National Park Natural Sciences Research Plan. U. S. Dept. of Interior, Natl. Park Service, Washington, D. C. Perkins, R. C. L. 1896. A collecting trip on Haleakala, Maui, Sandwich Islands. Ent. Monthly Magazine, 32:190-195. Perkins, R. C. L. 1899-1910. In D. Sharp, ed. Fauna Hawaiiensis. Cambridge Univ. Press. Yocum, C. F. 1967. Ecology of feral goats in Haleakala National Park, Maui, Hawaii. her. Mid. Naturalist 77:418-451. Zimmerman, E. C. 1948. Insects of Hawaii, Vol. I. Introduction. University of Hawaii Press, Honolulu. DISTRIBUTION AND VEGETATIVE ANATOMY OF HAWAIIAN SANDALWOOD Lani St emmermann Botany Department University of Hawaii at Manoa Honolulu, Hawaii 96822

Sandalwood was one of the earliest commodities of trade for the Kingdom of Hawaii with the European world. This trade was brought about by massive harvests of wood which led to the disappearance of the trees from the land. St. John (1947) has traced the beginning of the trade to as early as 1790, twelve years after the discovery of the islands by Cook. The height of the trade was between 1810 and 1825 but it was not until 1819, the middle of this period of heavy harvest, that Gaudichaud made the first collections of the genus in Hawaii for botanical study. Several other collections were made during the next decade, but there were no early records concerning the distribution of the plants prior to the trade. While the earliest explorers made careful and accurate descriptions of the avifauna, little description of flora and vegetation was made, probably due to the inability of the observers to distinguish between plant species. Evidence of this is included in the journal of Clerke, present on Cook's voyage, where a list of known plants was included, presumably provided by Nelson, the botanist. Mostly those plants that they had identified were plants that the voyagers had encountered in other ports, for at the end of the list Clerke indicated "that there were about twenty species of ferns, and fifty or sixty sorts of trees and shrubs that I knew nothing of." We are left, therefore, with no descriptions of former forests of sandalwood, but Rock (1916) indicated that there must have been great forests of the trees for there to have been as much as $400,000 income from the trade in one year. If the weight of the wood exported in a single year is calculated from this based on the price of $10.00 per picul (133.3 lbs.), twenty-five hundred tons of wood per year was exported at the height of the trade. Virtually the entire population was mobilized to supply the king's monopoly on the precious wood, and all districts were required to pay taxes of sandalwood. Degener (1930) has written that this massive employment of manpower was resented, and workers would destroy young plants so that future harvests would be impossible. Debts were incurred by the Kingdom because of the trade, and little trading was done after 1830. The later shipments were supposedly of inferior quality reportedly due to the absence of the once easily available choice wood, and the adulteration of the low quality sandalwood with substitutes. Such harvests drastically affected the distribution of the genus in the first half of the nineteenth century, but the effect of the trade on the present distribution of sandalwood can only be estimated. I feel there has been considerable recovery of the plants due in part to their ability to produce root suckers. This characteristic can be easily noted in areas where fires have burned, or roads have been cut. Even in areas that have not been disturbed, many young plants have been seen to originate as root suckers rather than by growing from seed. The felling of trees would likely be stimulus to sucker production. If the removal of young plants as suggested by Degener occurred over a long period of time, then this would certainly interfere with the recovery. Presently sandalwood is known throughout the main Hawaiian Islands, and is quite common in many areas. The genus may be broken down into two distinct groups. The first of these, the Freycinetianum group, consists of plants with medium to long flowers, which are usually red in color, especially in bud, as is illustrated by SantaZum haZeakatae, the sandalwood found in Haleakala National Park. The other group, the Ellipticum group, is characterized by plants with short flowers which are greenish to greenish brown in bud, illustrated by S. panicutatum, the species of sandalwood found in Hawaii Volcanoes National Park. Plants of the first group are trees found on the islands of Kaua'i, O'ahu, Moloka'i, and La-na'i in dry to mesic forests, and on Maui, in the alpine region as well. Plants in the second group may be shrubs or trees, and are found presently on all the major Hawaiian Islands with the possible exception of Ni'ihau and Ka-ho'olawe which have been poorly botanized recently. In the past a colony of Santatum eZZipticum var. Zaysanicum could be found on Laysan Island, in the Leeward Chain, but that taxon is now extinct. Each of these groups include species that are arborescent, and would have provided choice wood for the harvest. The following features of the wood were examined to see if any differences between the taxa could be observed: tangential and radial vessel diameter, vessel length, number of vessels per square millimeter, fiber length, ray height in cells and microns, rays per millimeter in tangential section, and number of crystal chains per millimeter in radial section. In transection, the wood of all species is diffuse porous, sometimes tending toward semi-ring porous, with the vessels usually distributed throughout the wood, but in some cases, rings are formed where vessels are either absent or very abundant. The tendency of the wood to show either of these patterns is not of taxonomic significance within SantaZum-- no taxon being consistent either in the presence or absence of the rings, or in the manner of their formation. These rings, as they are not found in a regular pattern in a sample, cannot be considered annual rings, but are more likely a response to some other environmental condition such as drought or fire. The wood has apotracheal diffuse parenchyma, and in longitudinal sections, chains of axial parenchyma cells, containing rhomboidal crystals, are seen. Also observed in longitudinal sections are tyloses, intrusive growths of parenchyma cells into old vessel members. Tangential sections of the wood illustrate the uniseriate and biseriate rays which may be filled with resin. When material is macerated the vessels are observed to be of various lengths and may have one, two, or no tails. Of all the characters examined, no characteristics of the wood were of taxonomic significance. In the accompanying graphs of tangential vessel diameter, crystal chains per millimeter, and rays per millimeter, it is seen that the ranges for all taxa, and groups, overlap considerably, and the means are also close. The other characters examined were equally as useless as taxonomic criteria. Characteristics of the leaves, being more a function of environmental conditions, were more distinctive. For instance, the thickness of the leaf varies considerably. In arid areas where plants of the taxon SantaZum eZZipticum var. ZittoraZe are exposed to sea spray, thick leaves develop where most of the cells are of the same type--mostly palisade parenchyma. In taxa that grow in areas not so extreme, a bifacial anatomy may be noted where there are two distinct tissue layers in the mesophyll. Another characteristic of leaf anatomy that appears to be correlated with environmental conditionsis the upper epidermis of SantaZum haZeakaZae. Plants of this taxon which grow in the crater region at high elevations and are subject to considerable solar radiation, have peculiar bottle shaped upper epidermis cells. However, plants of this taxon found in not so extreme environments have more conventional epidermal cells which indicate that the formation of the cells may be an environmental response. In conclusion, there appear to be no significant differences in the wood of the various taxa of SantaZum, and the differences in leaf anatomy probably reflect adaptation to local environments. Literature Cited

Clerke, C. 1967. In Beaglehole, J. C., ed. The Journals of Captain James Cook, V. 3. The Voyage of the Resolution and Discovery, 1776-1778. Cambridge University Press. Degener, 0. 1930. Ferns and Flowering Plants of Hawaii National Park. Star Bulletin, Ltd. Honolulu. Rock, J. F. 1916. The Sandalwoods of Hawaii. Botanical Bulletin #3, Territory of Hawaii Board of Agriculture and Forestry. Honolulu, Hawaii. 43 pp. St. John, H. 1947. The History and Present Distribution and Abundance of Sandalwood on Oahu, Hawaiian Islands. Pacific Science -1:15. THE INFLUENCE OF FOOD SUPPLEMENTATION UPON THE REPRODUCTIVE STMTEGY AND MOVEMENT PATTERNS IN THE HAWAI'I 'AMAKIHI (LOXOPS VIRENS)"

Charles van Riper I11 Department of Zoology University of Hawaii at Manoa Honolulu, Hawaii 96822

The objectives of this study were to investigate the influence of food availability on: 1) territory establishment, shape, and size; 2) nesting success; and 3) post-breeding dispersal in the Hawai'i 'Amakihi (Loxops virens). From 1970 through 1975 birds were color-banded and their territory boundaries recorded in the Pu'u-lz'au area on the southwestern slope of Mauna Kea, Hawai'i. Supplemental food sources, consisting of 20% sugar-water in feeders placed on yellow boards, were supplied in an area that heretofore had not been utilized for nesting. The feeders were placed at three different locations, each an 1, average" territorial distance from the other and from all contiguous territorial boundaries. Feeding bouts were recorded from 10 January through 10 June 1975 utilizing blinds, stopwatches, binoculars, and a tape recorder. Territory establishment and shape in the 'Amakihi were directly influenced by food availability; size was not. In the presence of a large food surplus, elongated overlapping territories were established around the periphery with one major territorial pair centrally located. The food surplus was utilized by territory holders, transient immatures, and introduced White-eyes (Zosterops japonica). Of the birds utilizing the food resources, only the major territorial 'Amakihi pair ranked above the White-eye, suggesting potential competition. Nectar appears to be an important factor during the breeding cycle of the 'Amakihi, with the greatest influence exerted during the latter stages of egg laying and early incubation due to an increase in desertion rate. It was found that post- breeding dispersal in the 'Amakihi can be suppressed when a continued food supply was made available.

* Abstract. PATTERNS IN THE DISTRIBUTION AND ABUNDANCE OF THE PALILA ON MAUNA KEA' C. van Riper 111' and J. M. Scott3

A recent cooperative survey involving the Hawaii Division of Fish and Game, Hawaii Division of Forestry, U.S. Forest Service, and the U.S. Fish and Wildlife Service resulted in the censusing of the entire known geographical range of the Palila (Psittirostra baitleui) This area was censused in January 1975 and again in September 1975. During these periods 597 and 586 kilometers of transects were walked and 2185 and 715 hectares respectively were censused for Palila. These figures comprise 13.6 and 5.3 percent of the total area surveyed. Palila were found only on the south and southeast flanks of Mauna Kea from L890 to 2,866 meters elevation. Estimated densities ranged from 1.83 to 125.73 birds per hundred hectares (median 17.28) during the non-breeding season, and from 2.53 to 88.96 in the breeding season. The Palila population was estimated to be 770 during the non-breeding season and 675 during the breeding season. These two figures are not significantly different from one another.

'Abstract 2~epartmentof Zoology, University of Hawaii at Manoa, Honolulu, Hawaii 96822. 3~ndangeredSpecies Program, Patuxent Wildlife Research Center, Hawaii Volcanoes National Park, Hawaii 96718. ROLE OF RESEARCH IN THE MANAGEMENT OF NATURAL RESOURCES IN AREAS ADMINISTERED BY NATIONAL PARK SERVICE IN HAWAII 0. L. Wallis Regional Chief Scientist Western Region National Park Service 450 Golden Gate Avenue, Box 36063 San Francisco, California 94102

Natural resources in areas administered by the National Park Service in Hawaii (Hawaii Volcanoes National Park,Haleakala National Park, City of Refuge National Historical Park, and Puukohola Heiau National Historic Site)are significant elements of endemic ecosystems--many are unique and of international importance and some are endangered. To fulfill a mandate to restore (where feasible) and to perpetuate these insular ecosystems and their constituent components requires effective management. Effective management is dependent upon sound and scientifically sufficient information of these resources. Contributions to knowledge about the natural resources of Hawaii by noted scientists over the decades have been significant. Yet we realize that so much more data is required to do the job. During this conference you will discuss your research projects that are adding so substantially to the fund of knowledge. Regional Director Howard Chapman is extremely interested in each of your research endeavors. He recognizes and thanks you for your role in securing the scientific data that provides the sound foundation for making informed resources management decisions and for the development and implementation of effective resources management actions.. He extends his appreciation to each of you for your participation in this-- the First Conference in Natural Sciences. He sincerely regrets that he is unable to personally meet with you during this research review. Briefly, I will outline and discuss some of the mechanics and steps involved in the identification, initiation and implementation of research in these areas in Hawaii: 1. Resources Identification and Evaluation. An initial identitication and evaluation of the natural resources is a prerequisite to effective management. Past and current research endeavors contribute to this step. 2. Resources Management Objectives. Based upon an identification of the park resources and their significance, the Park Superintendent establishes the basic Resources Management Objectives that guide his actions. Although based generally upon broad Service-wide policies and legislative requirements, his Resources Management Objectives must be Park- specific. 3. Resources Management Problems. With an under- standingy the park's resources and the establishment of management objectives, the park manager identifies and recognizes the problems which prevent him from achieving his objectives.

4. Current Management Actions. Actions being taken to solve these problems are identified and evaluated and questioned--are they adequate? 5. Needed Management Action. Management actions required to solve the problems and to achieve the objectives are identified and questioned --why are they not being undertaken? The answers may involve insufficient funds, equipment or personnel. However, many recognized management actions cannot be taken because of a lack of basic scientific knowledge about the resource. This is the step where research is required to provide the factual basis of management decisions and actions. Material from these steps is formulated and documented into an individual park-specific Resources Management Plan by the Superintendent and his staff. In this Plan he includes: a description of the park resources; the resources management objectives; the statement of current problems; and description of current and proposed actions required to solve these problems that obstruct the fulfillment of the resources management objectives. The Resources Management Plan is accompanied by an Environ- mental Assessment or Environmental Impact Statement that exposes the program to the general public and the scientific community and identifies the various alternative actions and the impacts of each. Public involvement assists the park manager in developing, evaluating, and implementing a program that is scientifically sound and socially acceptable. A Resources Management Program complements the Resources Management Plan. This program consists of a series of resources management and research action proposals and a projected programming sheet which provides for an orderly scheduling of these activites. Next, the projects enter the programming process for eventual funding when the Park Superintendent submits a form requesting an increase in his park base funding--we call it a Form 10-237--and places the project on his park's priority list where it competes with other park funding demands. These documents are sent to the Regional Office. One criterioneachproject must meet is known as "Briggle's Law." In general, it states that only those resources management or research projects that are responsive to a resources management need that has been recognized in an a roved Resources MGement Plan shall be financed with4f-h un s allotted to the National Park Service. Projects submitted by the Park Superintendents are reviewed and evaluated in the Western Regional Office and ranked on a Region-wide Priority List in competition with all needs of other areas within the Region--and, of course, the availability of funds. A tentative Regional Priority List is circulated to all field areas. Subsequently, after appropriate revisions based upon field review have been made, the List is approved by the Regional Director and turned over to the Regional Chief Scientist to facilitate. Research is undertaken in a National Park by one of several means, as follows: 1. Research Scientists. Seven research scientists are located in parks in the Western Region, and two serve on university campuses as unit Leaders at Cooperative National Park Resources Studies Units. Each works on specific research projects and provides the Superintendent and his staff with professional assistance and advice and research coordination. Research Scientist James K. Baker serves in Hawaii Volcanoes National Park. 2. Cooperative University Units. In the Western Region we enjoy cooperative arrangements with four universities: University of Hawaii, University of Nevada/Las Vegas, University of Arizona, and University of California/Berkeley. In Hawaii, we participate with the University of Hawaii, under a Master Memorandum of Understanding dated March 16, 1973, in operation of the Cooperative National Park Resources Studies Unit. This CPSI! operates on a University-wide scope through. the Department of Botany on the Manoa Campus. The program consists of two main elements; namely, (I) the Unit Support and Park Assistance Programs and (2) a series of specific research projects. Dr. Clifford Smith, a member of the University faculty, serves as the Unit Director.

3. ---.Contracts. Some research p:rojects are done unuer conr-:(:act wi.tlt educational institutions and other qualified organizations.

1, C~C,~JZ!.U...- .. . .. -. ivc. - ...... ,; . Coopcrai ive research projec. .:; 1 :u! ... i.12.. i:ci:!i cooperni-ive arrangements w-i.iA other Federal and state agencies such as the Geological Survey, the U. S. Fish and Wil-dlife Ser-vi.ce,the U. S. Forest Service, etc. 5. Indeerdent Res~iarch. A considerable amount of research is accoa~pl~~shedby researchers that are independently funded from other than National Park Service resources. Such research endeavors are encouraged so ?.oag as they do not alter or endangel: pafk .resources or cause an adverse irnpacir upon the el:.j'oyment of the park by other visitors. The Service may provide services and facilities, as available and appropriate, to implement these knvest%gations. Much scientific information, useful in the management and interpretation of park resources, is derived from i.rideycr:der;tl.y.~~fur,dedresearch endeavors.

hn independent researcher should contact the Superintendent of the park in which he proposes to undertake research for his advice and assi.st;mce. Before, funds a:re di.stributed for a research prdject on the Regional List, a Cull-term research proposal is requested. This proposal is reviewed by personnel in. the park and in the Regional Office pi:ior to approval by the Regional Chief Scientist. (1nstruct:i.ons aiitd procedures for preparation and submission of proposals may be secured from the Regional Chief Scientist.) The appro.ved research proposal serves as the basis for a contract (or the Research Work Plan for an individual NPS Research Scientist.) The contract provides for a specific product or products --usually progress reports and a detailed final research report. (A copy of our reporting requirements may be secured from the Regional Chief Scientist.) These reports supply the park manager with scientific conclusions and recommendations that provide the basis for sound resources management decisions and effective resources management action programs. Investigators are encouraged to publish the findings of their research, as appropriate. First consideration should be publication in professional journals. The Office of the Chief Scientist, Washington, can provide for page charges and reprint costs. In addition, the National Park Service has several series of scientific publications for which acceptable manuscripts are welcome. (A statement on the Service's format and manuscript requirements and a list of current Service scientific titles are also available.) Your continued participation in the scientific and resources management endeavors of the Service and your contributions to these efforts of mutual interest and concern are , greatly appreciated. Many thanks for sharing your ideas and findings with us. THE KALA-PANA EXTENSION: ITS VARIETY, VEGETATION AND VALUE Frederick R. Warshauer Botany Department University of Hawaii at Manoa Honolulu. Hawaii 96822

This paper provides a glimpse of the natural history of the eastern third of Hawaii Volcanoes National Park called the Kala-pana Extension. This area of approximately seventy- five square miles was added to Hawaii Volcanoes National Park following a 1938 Congressional authorization. It runs from the rocky coast up to 2500 to 3000 feet elevation, and almost all is east of the boundary between the Puna and Ka'ti districts, near the western edges of the recent Mauna- ulu lava flows. More interesting than the political limits is the main geographical one--that is, nearly the whole region lies along and down slope of the very active East Rift of Ki-lau-ea. There are two main geological features which are of particular importance to the vegetation distribution of the region: the many lava flows and the Hilina Fault System. The lava flows exhibit several textural surfaces of 'a's and pahoehoe lavas. In the upper slopes near the rift, most of the surface consists of geologically recent pHhoehoe flows with fairly vesicular rock. Some flows have smooth, shelly surfaces and others are very slabby and rough. A few of these pshoehoe flows turn into 'a's flows downslope, sometimes with a transition of very blocky pshoehoe. If an eruptive phase lasts long enough, as in the Mauna- ulu series, then large portions of the original flows may be covered by the denser, degassed pzhoehoe carried by extensive tube systems. Generally, these tube-fed flows extend much farther downslope, making up the bulk of the surface of the lower elevation of the Kala-pana Extension. Observations of vegetation growing on these different surfaces of a flow or flow series indicate that forest development occurs much more rapidly on the 'a's and rough textured pahoehoe than on the denser, tube-fed pghoehoe farther from the vent. An important _featureof lava flows is the common occurence of the kipuka, an island of older substrate and vegetation surrounded by younger lava flows. In such a volcanically active area as the East Rift of KT-lau-ea, much of the surface is a patchwork of different aged kFpukas. Walking through the area will provide one with examples of steps in the development of forests and the succession of species within them. The other important geological feature of the area is the Hilina Fault System, which has produced several fault scarps. The numerous earth movements along these scarps have extensively fractured the lava to give it a structure similar to that of 'a's. The scarp areas may also have increased moisture due to seepage. Where not covered over by more recent lavas, the scarp areas tend to act as kipukas, very often harboring the best developed forest of the immediate area. In addition to the lava substrates, there are two cinder cones in the Kala-pana Extension. These have on their slopes ash deposits, which give rise to a different kind of soil and have on at least one of them a few plants not seen elsewhere in the Extension. Superimposed on the geological setting of the region are two rainfall gradients. In addition to the usual orographic increase in precipitation with elevation, there is a northeast to southwest decrease representing a gradual shift from windward to leeward exposure. The range in precipitation over the Kala-pana Extension allows for eventual development of wet, mesic and dry forest zones, and examples of each can be found here. All or nearly all of the lava and ash surfaces appear to be young enough that their age is reflected by the degree of development of the vegetation upon them. The influence of the rainfall and elevation gradients on the different kinds and ages of substrate has resulted in a mosaic of various vegetation types. Disturbances from fires, agriculture and introduced plants and animals have added to the region's heterogeneity. Due to the frequency of newer lava flows at least partially covering the older ones, the exposures of the older ones usually have been reduced to fragments, and the-oldest tend to be the least common. It is in these older kipukas and flowsegments where the most. diverse plant communities are located. As most of the many rare and uncommon plants in this section of the Park are found only in the more mature communities, their distributions are naturally patchy. As mentioned, dry to wet moisture regimes allow for the eventual development of corresponding forest types. The most prevalent are the rain forest communities distributed in the wetter and higher parts of the region. Of these, the most diverse and pristine examples in the Park grow in a recent kipuka on and just downslope of an unnamed cinder cone along the East Rift. The dry forest communities are more scattered over the lower to middle reaches of the Kala-pana Extension. Some of the best lowland dry forest left in the Islands occurs on an old 'a'a flow in the Ka-moamoa ahupua'a here. There are several areas of mesic forest that have developed between some of the previous two types, mostly on the fault scarps above the Kala-pana Trail. Here grow some of the largest trees in the Park--Metrosideros well over a meter in diameter and 30 meters tall and, on another substrate, TetrapZasandra hawaiensis up to a meter and a half in diameter. There is also a sizable part of the extension in very open forest, scrub, grassland and barren lava. Those areas probably reflect past disturbances, early plant colonization, and limiting rainfall as well as very recent lava flows. The quality of these forest types varies greatly with the degree of successional maturity and especially with the amounts of disturbance to which they have been and are being exposed. In some of the lower and middle elevations there has traditionally been some agriculture, and later grazing, practiced. In an area of such volcanic activity, periodic fires have fringed the lava flows. Within the last fifteen years or so, two introduced species of Andropogon, a bunch grass, have invaded a large part of the park where the forest has been open enough. Since then, the size and severity of wildfires in this region have greatly increased. These grasses are just two of the many species of introduced plants and animals which threaten the integrity of the forests of the Kala-pana Extension. Other species include Myrica faya, two species of Psidium, Stachytarpheta jamaicensis, Leucaena ZeucocephaZa, Lantana camara, and Eugenia jumbos. Establishment of most of these and other problem exotics in the Extension is enhanced by disruptions of the native vegetation from fires and sustained feeding activites of feral mammals. In some of these disturbed areas one can see dense patches of one or more of the previously mentioned plants, and the prospect is for much more of the same. While they have recently been rendered scarce in the Park, the effects of feral goats can still be seen. In many of the drier scrub and forest areas of the Extension, their selective browsing has all but eliminated the native understory. Much more of a problem are the hordes of feral pigs, especially in the rain forests, where their ravages are the most evident. As with goats, their selective feeding has severely depleted certain species. More important is habitat destruction. The pigs disrupt and in large areas have eliminated the Cibotium understory by killing the tree ferns in the process of eating the quantities of starch inside them. Where the lava is not too rugged, they have stripped the ground of its naturally dense cover of vegetation by their digging activities, exposing it to invasion by weedy exotics. The spread of exotics in the genera of Psidium, Rubus, and PassifZora, at least, is no doubt accelerated by seed dispersal through their feces. The consequences of the tremendous numbers of pigs in the Park's forests is a clearing, plowing, and sowing not unlike a primitive agriculture, complete with considerable soil erosion. Even the epiphyte habitats have been altered by this exposure and resultant desiccation. The uncommon epiphytes are lost, and only the hardiest survive. It is no coincidence that many of the rare and uncommon plants in the Kala-pana Extension, as elsewhere in the Park, are restricted to places where the local geography allows for protection from feral animals. Pit craters, large earth cracks, a few of the recently isolated kFpukas, rough 'a'5 and slabby pzhoehoe substrates, and in a few instances proximity to human activities have acted as partial or complete protection from goats or pigs. Comparison of these fragments with their disturbed counterparts provides good examples of these alien animals' enormous influence. In spite of the considerable problems present in managing its biological resources, the Kala-pana Extension has tremendous value to its natural inhabitants and to the appreciative people who pass through it. The region is dominated by its active geology, and its interaction with life forces. There is no better place in Hawai'i to observe the sequences of forest development on both young and old lava flows. As sanctuaries for many native species it contains the best dry and rain forest habitats in the Park. In some of the few pig-free rain forest areas existsthe opportunity to observe relatively pristine forest. LIST OF PARTICIPANTS

Jay Aiu, Bishop Museum Barbara Allen * Suzy Allen, Mauna Loa Field Station, Russell A. Apple, Hawaii State Office, National Park Service Christina Ariola, State Dept. of Land and Natural Resources Ken Baker, Hawaii Volcanoes National Park N. Balakrishnan, Botany Dept., University of Hawaii at Manoa Cynthia Baldwin, University of Hawaii at Hilo Paul Banko, National Park Service Mr. and Mrs. W. E. Banko, U.S. Fish and Wildlife Service (retired) Bob Barbee, Hawaii Volcanoes National Park Bob Barrel, Hawaii State Office, National Park Service James Bertilacci, Jr., Los Angeles City Schools David L. Bishop, Moorpark College, California Mary Lew Breese ~auiL. Breese Bobby Camara, Director--Hawaii 2000 erai id Carr, Botany Dept., University of Hawaii at Manoa Hampton L. Carson, Genetics Dept., University of Hawaii at Manoa Meredith S. Carson, Genetics Dept., Univ. of Hawaii at Manoa Daniel P. Cheney, Neighbor Island Consultants, Hilo Paul L. Cleghorn, Bishop Museum Pat Conant, U.S. Fish and Wildlife Service Sheila Conant, General Science Dept., Univ. of Hawaii at Manoa Ranjit G. Cooray, Botany Dept. , University of Hawaii at Manoa Carolyn Corn, Botany Dept., University of Hawaii at Manoa Lisa K. Croft, Botany Dept. , Univ. of Hawaii at Manoa Clifton J. Davi-s, Bishop Museum Isa Degener, New York Botanical Garden Otto Degener, New York Botanical Garden Annette Mueller-Dombois Dieter Mueller-Dombois, Botany Dept., Univ. of Hawaii at Manoa David D. Dunatchik, Haleakala National Park Lee Goff, Entomology Depts., Bishop Museum and Univ. of Hawaii Sandra Guest, Honolulu Community College Don E. Hemmes, Biology Dept., University of Hawaii at Hilo Keith Hoofnagle, Hawaii Volcanoes National Park Terry Hunt, Bishop Museum Frances Jackson, University of Hawaii at Manoa James D. Jacobi, Botany Dept., Univ. of Hawaii at Manoa Kenneth Y. Kaneshiro, Entomology Dept., Univ. of Hawaii at Manoa Chris Katahira, Hawaii Volcanoes National Park Larry Katahira, Hawaii Volcanoes National Park Bruce M. Kilgore, Western Region Office, National Park Service John I. Kjargaard, Haleakala National Park E. Kridler, U.S. Fish and Wildlife Service Don Kyhos, Botany Dept., Univ. of California, Davis Edmuncl J. Ladd, City of Refuge National Historical Park Charles Lamoureux, Botany Dept., Univ. of Hawaii at Manoa 242 List of Participants (continued)

Elbert L. Little, Jr., U.S. Dept.Agriculture, Forest Service Mrs. E. L. Little, Jr., U.S. Dept. Agriculture (retired) H. F. Little, University of Hawaii at Hilo Carol A. Martin, National Park Service, Arizona Jean Martin, Jackson-Martin Consultants Patrick C. McCoy, Bishop Museum Judy McCoy, Bishop Museum Holly McEldowney, Bishop Museum Jeffrey Mermel, Hawaii Volcanoes National Park John M. Miller, Mauna Loa Observatory Steve Montgomery, Entomology Dept., Univ. of Hawaii at Manoa, and Natural Area Reserves Linda Mull, Hawaii Audubon Society Mae E. Mull, Hawaii Audubon Society William P. Mull, Bishop Museum Gail M. Murakami, Botany Dept., Univ. of Hawaii at Manoa Robert E. Nelson, U.S. Forest Service Alan T. Ohta, Genetics Dept., Univ. of Hawaii at Manoa Catherine K. Okano, Botany Dept., University of Hawaii at Manoa Greg Owen, Hawaii Volcanoes National Park Terry T. Parman, Cooperative Park Studies Unit, Univ. of Hawaii H. Douglas Pratt, Louisiana State Univ. Museum of Zoology C. J. Ralph, U.S. Forest Service Don Reeser, Hawaii Volcanoes National Park James Rutherford, Biology Dept., Univ. of Hawaii at Hilo Mike Scott, U.S. Fish and Wildlife Service Paul G. Scowcroft, U.S. Forest Service Robert Shallenberger,. Ahuimanu Productions Clara Shimoda Jerry Y. Shimoda, City of Refuge National Historical Park Clifford W. Smith, Botany Dept., Univ. of Hawaii at Manoa, and Cooperative National Park Resources Studies Unit, U.H. Wayne H. Souza, Hawaii State Dept. of Land and Natural Resources Lani Stemmermann, Botany Dept., Univ. of Hawaii at Manoa Maile Stemmermann, University of Hawaii Kimo Tabor, The Nature Conservancy P. Quentin Tomich, Hawaii State Health Department Monica Udvardy, Bishop Museum Charles van Riper 111, Zoology Dept., Univ. of Hawaii atManoa 0. L. Wallis, Western Region, National Park Service Rick Warshauer, Botany Dept., Univ. of Hawaii at Manoa Deborah Weiner, Cooperative Park Studies Unit, Univ. of Hawaii David Woodside, Hawaii State Dept. of Land andNatural Resources Layne Yoshida, Botany Dept., Univ. of Hawaii at Manoa Jamie Young, Bishop Museum SUBJECT INDEX

Acari, 89 'Ghi'a, 9, 75 adz quarry, 135 ornithology, 23, 29, 63, 'Alal5, 23 227. 229 'Amakihi, 63, 227 ~alila,229 anthropology, 3, 9, 55, 95, 135 petroglyphs, 55 archaeology, 3, 55, 95, 135 phenology, 127 behavior, 143. 167, 227 photography, 163 birds, 23, 29; 63,'227, 229 pigs, feral, 107, 217 botany, 19, 75, 107, 123, 127, Pu u-kohola, 47, 131 131, 149, 187, 207, 217, 223, rare organisms, 39, 123 3'37 research, 231 burial cave, Hawaiian, 3 resource management, 23, 35, cateroillars, 143 149, 183 cave, burial petroglyph, 55 sandalwood, 223 climate, 113 sheep, feral, 187 CPSU, 213 skinks, 11 Drosophi Za , South Point, 95 ecology, 11, volcanoes. 129 167, 171, entomology, 167 eruptions, volcanic, 129, 207 Eupithecia spp. , 143 evolution. 115. 143 exotic organisms, 11, 29, 107, 171, 179, 207, 237 fire, ecology of, 171, 179 food deoletion. 29 fumarole, 89 ' grassland, native alpine, 107 Hale-a-ka-lH, 107, 217 Hawaii ~olcanoesN.P., 11,19, 63, 89, 123,127,149,171,179,183, 207, 237 Hibiscade Zphus , 19 hydration-rind dating, 95 insects, 23, 39, 115, 143, 163, 167 interoretation. 199

koa, 9 mzmane, 187 marine fauna, 47 Mauna Kea, 135, 187, 223, 229 Mauna Loa, 129 Metrosideros, 9, 75 mites, 89 naio, 187 North Kona, 3