Threats to Native Aquatic Insect Biodiversity in Hawai'i and the Pacific
THREATS TO NATIVE AQUATIC INSECT BIODIVERSITY IN HAWAI'I AND THE PACIFIC,
AND CHALLENGES IN THEIR CONSERVATION
A DISSERTATION SUBMITTED TO THE GRADUATE DIVISION OF THE UNIVERSITY OF HAWAI 'I IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
IN
ENTOMOLOGY
AUGUST 2005
By Ronald A. Englund
Dissertation Committee:
Mark Wright, Chairperson Dan Rubinoff Neal Evenhuis Dan Polhemus Andrew Taylor TABLE OF CONTENTS
ACKNOWLEDGEMENTS , ii
ABSTRACT iii
LIST OF TABLES vi
LIST OF FIGURES viii
CHAPTER 1. THE IMPACTS OF INTRODUCED POECILIID FISH AND ODONATA ON THE
ENDEMIC MEGALAGRION (ODONATA) DAMSELFLIES OF 0'AHU ISLAND, HAWAI'I 1
CHAPTER 2: EVALUATING THE EFFECTS OF INTRODUCED RAINBOW TROUT (Oncorhynchus mykiss) ON NATIVE STREAM INSECTS ON KAUA'I ISLAND, HAWAI'I 40
CHAPTER 3. LONG-TERM MONITORING OF ONE OF THE MOST RESTRICTED INSECT
POPULATIONS IN THE UNITED STATES, Megalagrion xanthomelas Selys-Longchamps, 1876, AT
TRIPLERARMY MEDICAL CENTER, O'AHU, HAWAI'I... 76
CHAPTER 4. THE LOSS OF NATIVE BIODIVERSITY AND CONTINUING NONINDIGENOUS
SPECIES INTRODUCTIONS IN FRESHWATER, ESTUARINE, AND WETLAND COMMUNITIES
OF PEARL HARBOR, O'AHU, HAWAIIAN ISLANDS 91
CHAPTER 5. FLOW RESTORATION AND PERSISTENCE OF INTRODUCED SPECIES IN
WAlKELE STREAM, 0'AHU 125
CHAPTER 6: INVASIVE SPECIES THREATS TO NATIVE AQUATIC INSECT AND ARTHROPOD
BIODIVERSITY IN HAWAI'I, THE PACIFIC AND OTHER RELEVANT AREAS WITH DISCUSSION
OF CONSERVATION MEASURES 143 ACKNOWLEDGEMENTS
I would like to thank the many people that have made this dissertation possible. I especially would like to
extend my thanks and warmest gratitude to my advisor Mark Wright, whose sense of humor and keen
intellect made this process as enjoyable as it can be. My committee, consisting of Dan Rubinoff, Neal
Evenhuis, Dan Polhemus and Andrew Taylor provided valuable insights and advice throughout. I sincerely
acknowledge and appreciate the efforts of my entire committee throughout my time at the University of
Hawai'i at Manoa. I am also deeply appreciative of Neal Evenhuis and Allen Allison for their
encouragement, and allowing me the flexibility to pursue a Doctorate while being employed at the Bishop
Museum. I have greatly enjoyed the scientific and cultural 'ohana at the Bishop Museum that always provided an ideal research and working atmosphere. Rob Cowie and Frank Howarth of the Bishop Museum
also provided valuable reviews and advice for many of these chapters. Bishop Museum librarians Patti
Belcher and B.J. Short were always helpful in tracking down the many obscure references. Several key
organizations provided the support that allowed my research to take place, and I thank the following
organizations that funded this research: Bishop Museum, Hawaii Division of Aquatic Resources, Nature
Conservancy, Smithsonian Institution, and the Delegation ala Recherche Polynesie fran<;aise.
People too numerous to mention assisted me in various aspects of the fieldwork required for this wide
ranging dissertation, and I greatly appreciate help from David Preston, Betsy Gagne, Dan Polhemus, Jean
Yves Meyer, BenoIT Fontaine, Olivier Gargominy, Tina Lau, Stephanie Loo, Brian Naeole, Alison
Sherwood, and Steve Jordan. Special thanks goes to the Hawaii Division of Aquatic Resources crew
including Bob Nishimoto, Glenn Higashi, Darrell Kuamo'o, John Kahiapo, Skippy Hau, Bill Puleloa, and
Mike Yamamoto. Bill Devick and Bob Nishimoto were also instrumental in encouraging and funding much
of this research as well. La Vonne Furtado provided consistent moral support during the crucial final stages
of this journey. I am also privileged to have incredibly supportive parents and a wonderful family, without
whom I would have never attempted this work. I would like to dedicate this dissertation to my wonderful
parents, Stanley and Marjorie Englund.
11 ABSTRACT
Although the decline in numbers and diversity and threat to native insects in the Hawaiian Islands is widely recognized by field scientists there has been little progress in either documenting the real decline of native species, or in demonstrating specific causes of the overall decline of these species. Additionally, few conservation actions to either restore populations or mitigate actual threats to native arthropods have been mentioned in the literature. The following chapters examine several assessments of relevant aquatic systems and the native aquatic insects dwelling within, where there has either been a perceived or real decline of these native Hawaiian aquatic arthropods because of threats from invasive or introduced species.
The large adaptive radiation of the endemic native damselflies (Coenagrionidae: Megalagrion) in Hawai'i has received considerable attention and study since at least the 1880s. Endemic Megalagrion are in many ways reflective of a great loss because they are largely now found in remote upper headwater areas of streams, yet they also represent the hope of preserving highly diverse freshwater ecosystems found throughout the
Hawaiian archipelago. The first two chapters of this dissertation examine the impacts of two differing taxa of introduced fish on Hawaiian Megalagrion, Poeciliidae (livebearers or mosquitofish family) and
Salmonidae (trout). The effects of each fish species on native aquatic insects depended mainly on the invasive status ofeach group; for example, Chapter 1 (Englund 1999) examines the impacts of introduced poeciliids on native damselflies. Damselflies were completely eliminated on the island of 0'ahu wherever species in the highly invasive mosquitofish family were found, and only remnant populations were found in high elevations lacking introduced fish. Chapter 2 (Englund and Polhemus 2001) examines the impacts of the non-invasive rainbow trout (Oncorhynchus clarkI) on Megalagrion damselflies. Damselflies and all other native aquatic insects were not found to be harmed by trout in the uppermost elevations of Kaua'i streams where trout reproduce naturally, and even had more robust populations than in some nearby non-trout containing streams. The lack of impacts on native damselflies by a large, generalist predator such as rainbow trout pointed out a seeming paradox. Whereas the small but ubiquitous mosquitofish appears to have completely devastated native aquatic fauna wherever it has been introduced outside of its natural range,
iii trout, because of their restricted range and smaller population sizes have had minimal, if any impacts on native invertebrates in Hawai'i.
Because introduced fish species have caused either the extinction or severe range contractions of Megalagrion damselflies in Hawai'i, long-term monitoring of the remnant populations has become necessary to preserve these remaining populations. Chapter 3 (Englund 2001) provides a case study in both the monitoring and preservation of a remnant O'ahu damselfly population now found in only 95 m of fishless stream at the
TripIer Army Medical Center. Chapter 3 also provides several harrowing examples of how this species was nearly been eliminated in the past 10 years through accidents and mismanagement. Not only are the endemic
Megalagrion now missing from all lowland areas of O'ahu (with the exception of the TripIer population), lowland aquatic insect diversity throughout O'ahu is at a remnant status, and biodiversity surveys for native aquatic insects in the Pearl Harbor watersheds in Chapter 4 (Englund 2002) indicated a near absence of native aquatic insects in these freshwater habitats. Lower Pearl Harbor watersheds were documented to have lost many native aquatic insect taxa such as all native Heteroptera, damselflies, Coleoptera, and many
Diptera species, while introduced insect species were abundant.
A variety of conservation measures have been suggested to either restore or maintain the current levels of freshwater biodiversity in Hawai'i. In Chapter 5 (Englund and Filbert 1999), the case of significantly increasing and restoring stream flow in a formerly diverted stream was examined to determine whether this factor alone would lead to a restoration of native aquatic species. It was found that merely increasing stream flow by itself was not enough to rid the stream of any alien aquatic species, in fact, several new nonindigenous aquatic species became established after stream flows were increased. The results ofChapter
5 confirm that an integrated, balanced and possibly drastic approach will be required to maintain and preserve
Hawai'i's native aquatic insect fauna. A wide-variety of conservation measures in the Hawaiian archipelago will be needed to maintain current biodiversity levels, and also hopefully restore native freshwater biodiversity in selected areas.
iv To put the Hawai'i problem into perspective, a brief review of the impacts of invasive species on native insects in other tropical areas is provided in Chapter 6. This review chapter also provides a synthesis of the problem facing Hawaiian freshwater insects and other terrestrial arthropods in Hawai'i and elsewhere due to invasive species, and how the Hawaiian case study of invasive species impacts has many parallels to other vulnerable biotas. Finally, drawing on a mixed record ofpast mistakes and successes in Hawai'i and elsewhere, some potential practical conservation measures intended to preserve and restore endemic island aquatic insects are provided in Chapter 6.
v LIST OF TABLES
Table 1.1. Extinctions of Megalagrion species in surveyed 0'ahu aquatic habitats since 1936 8
Table 1.2. Remnant native Megalagrion species found in O'ahu streams, tributaries in parentheses, and
relative abundance in stream areas containing native damselflies, (rare (R) = < 3 individuals collected or
observed); (moderately common (C) = <: 3 individuals collected or observed) 9
Table 1.3. Biogeographic status and Hawaiian Island distribution of aquatic Megalagrion and introduced
Odonata species found on O'ahu 11
Table 2.1. Aquatic insect species and native or introduced status collected at each Kaua'i stream 50
Table 2.2. Total number of aquatic species collected during benthic, drift, and aerial (general) collections
during this study 51
Table 2.3 Presence or absence in surveyed K6ke'e State Park streams ofnaturally reproducing and stOcked
rainbow trout and native Megalagrion damselflies 57
Table 2.4. Geographic origin and terrestrial or aquatic status of prey items found in 80 K6ke'e trout
stomachs, 1997-1999 60
Table 2.5. Summary numbers and percent frequency of native prey items of special concern collected in
rainbow trout stomachs during this study, compared to the number of taxa collected per stream 61
Table 4.1. Summary of the native or nonindigenous status and total number (percent) of aquatic species
found in Pearl Harbor estuarine habitats 98
Table 4.2. Geographic source (year of introduction) and known (or probably known) mode of introduction of
nonindigenous species ofaquatic macrofauna found in Pearl Harbor streams and estuaries 107
Table 5.1. The range and mean water velocities (± standard error) recorded in transects downstream of
Waikele Springs 131
Table 5.2. Introduced and native species found in Waikele Stream, Oahu in 1993 and 1997-1998 from 250
m above Waikele Springs downstream to concrete weir. Oahu introduction dates from Beardsley
(1980), Devick (1991a), Cowie (1995), Polhemus & Asquith (1996), Randall (1996), Cowie (1998).
...... 132
vi Table 6.1. Extinction status of native insect taxa in the Hawaiian Islands that have recently had their
conservation status examined, to lowest taxonomic resolution 149
Table 6.2. Successfully eradicated invasive animal species in the Hawaiian Islands 161
vii LIST OF FIGURES
Figure 1.1. Limnological divisions on O'ahu defined for the purposes of this study 5
Figure 1.2. Status of stream and wetland dwelling damselflies on the island of 0'ahu. E = Extinct on 0'ahu
...... 8
Figure 1.3. Elevational distribution for Hawaiian Megalagrion damselflies and introduced poeciliid fish for
streams entering Kane'ohe Bay, O'ahu 12
Figure 1.4. Elevational distribution for Hawaiian Megalagrion damselflies and introduced poeciliid fish for
streams entering northern windward O'ahu 13
Figure 1.5. Elevational distribution for the Hawaiian Megalagrion damselflies and introduced poeciliid fish
for streams entering Pearl Harbor and leeward O'ahu 14
Figure 1.6. Elevational distribution for introduced Ischnura ramburii and Ischnura posita damselflies and
introduced poeciliid fish for selected Kane'ohe Bay and Pearl Harbor streams, O'ahu 18
Figure 2.1. Study area of sampled K6ke'e State Park Streams, Kaua'i Island, Hawai'i 45
Figure 2.2 Summary of aquatic species collected in K6ke'e State Park streams from all sampling methods
combined (general sampling, Malaise traps, drift, benthic samples) 52
Figure 2.3. Summary of all insect species collected from rainbow trout stomachs (n = 80) and their
terrestrial or aquatic, and native or introduced status in K6ke'e State Park Streams 52
Figure 2.4. Mean density by stream for the two most important constituents of benthic (Surber) samples,
the caddisfly C. pettiti and midge C. bicinctus 53
Figure 2.5. Summary graph of number of aquatic species in Kaua'i Streams and the presence or absence of
naturally reproducing trout in each stream; Lumaha'i and Hanalei have never been stocked with trout.
...... 53
Figure 2.6. The six numerically most abundant aquatic insect taxa captured in drift samples taken in K6ke'e
state park streams 58
Figure 3.1. Map of O'ahu, Hawai'i showing locations of current and historic records for Megalagrion
xanthomelas (from Evenhuis et a1., 1995) 78
viii Figure 3.2. TAMC Mitigation ponds prior to drainage, February 2000 79
Figure 3.3. Megalagrion xanthomelas captures at TAMC stream, May 1997-June 2000 82
Figure 3.4. Megalagrion xanthomelas captures at TAMC mitigation ponds from May 1997-February 2000.
...... 83
Figure 3.5. Oviposition scars on water lilies at the TAMC mitigation ponds July 1997-February 2000.. 84
Figure 4.1. Map of Pearl Harbor with sampling locations 95
Figure 4.2. Number of species by stream and native or nonindigenous status for combined aquatic fauna
found in estuarine regions of Pearl Harbor 98
Figure 4.3. Native or nonindigenous status of fish species and total numbers found at different salinity
levels in Pearl Harbor estuaries 101
Figure 4.4. Native or nonindigenous status of aquatic insects at varying elevations on O'ahu: upper Halawa
data from Polhemus (1994), upper Waikele data from Englund (1993) 112
Figure 5.1. Waikele Stream study area 128
ix CHAPTER 1. THE IMPACTS OF INTRODUCED POECILIID FISH AND ODONATA ON THE
ENDEMIC MEGALAGRION (ODONATA) DAMSELFLIES OF 0'AHU ISLAND, HAWAI'I
1 ABSTRACT
Since the beginning of this century there have been substantial declines in the distribution and abundance of native Megalagrion damselflies on the Hawaiian Island of O'ahu. Native damselflies have also vanished from most low elevation areas on other Hawaiian Islands, although historically, lotic and wetland dwelling damselfly species were once common throughout the archipelago. It is hypothesized that poeciliid fish introduced for biological control have caused the decline of four stream-breeding damselfly species on O'ahu, and the extinction or near-extinction of two other species statewide. This study documents the presence of remnant Megalagrion populations in O'ahu streams, wetlands, and estuaries, and records the elevational distributions of introduced fish in each waterbody surveyed. The distributions of introduced Odonata are also recorded, because the eight species ofdamselflies and dragonflies introduced to O'ahu since 1936 present another potential threat to native Hawaiian damselflies. Native damselfly and introduced poeciliid fish distributions were mutually exclusive on O'ahu, and it is concluded that this is probably due to predation by these introduced fish. By contrast, even the rarest native Megalagrion damselflies were found in areas containing introduced damselflies and dragonflies.
INTRODUCTION
The Hawaiian Islands have a rich native damselfly fauna with 26 recognized species and subspecies in the endemic genus Megalagrion (Polhemus and Asquith 1996). Hawaiian damselflies have radiated into a wide range of aquatic and terrestrial habitats, including streams and wetlands, coastal anchialine ponds, and upland bogs. Certain species also breed in upland terrestrial habitats, such as in the leaf litter of the native fern
Dicranopteris linearis, and in waterpockets formed in the base of climbing plants such as Freycinetia arborea
(Williams 1937a; Polhemus and Asquith 1996).
Although native damselflies were formerly one of the most conspicuous elements of Hawaiian stream and wetland communities, many species appear to be increasingly rare or have disappeared altogether. Because of efforts by early collectors, historic damselfly distributions and abundances prior to introductions of alien
2 aquatic species are relatively well known, particularly on O'ahu. Extensive collections by R.C.L. Perkins from 1892 to the early 1900s documented the presence of native damselfly populations prior to many environmental changes (Perkins 1899; Perkins 1910; Polhemus 1993; Liebherr and Polhemus 1997). These early surveys were followed by almost a century of subsequent damselfly collecting. Liebherr and Polhemus
(1997), assessing Megalagrion relative abundances over time, found a substantial decline in stream species on O'abu since 1892, even though terrestrial species were as widespread and abundant as they were in the
1890s.
By 1935, native damselflies on O'abu, such as Megalagrion xanthomelas (Selys-Longchamps, 1876), were becoming uncommon, and fish introduced for mosquito control were suspected to be involved in their decline (Williams 1937a). Subsequent researchers also implicated introduced fish as a cause of decline
(Zimmerman, 1948; Polhemus 1993; 1997; Polhemus and Asquith 1996). However, these observations were anecdotal, and definitive distributional data for introduced fish and endemic damselflies have not been published.
Hawai'i has a long history of purposeful and accidental introduction of aquatic species (Funasaki et aI.,
1988; Devick 1991). Among these were three species of fish imported from southern Texas in 1905 for mosquito control (Van Dine 1907; 1908). At least two of these species, Gambusia affinis (Baird & Girard,
1853) (western mosquitofish) and Poecilia latipinna (Leseur, 1821) (sailfin molly), both in the family
Poeciliidae, eventually established naturally reproducing populations in Hawai'i. In 1922, more poeciliid species such as Poecilia reticulata Peters, 1859, Xiphophorus helleri (Heckel, 1848) and Xiphophorus maculatus (Gunther, 1866) were introduced for mosquito control (Brock 1960). Since these initial biological control introductions there have been additional intentional and accidental introductions of fresh and brackish water fish species. By 1991, at least 44 species of introduced freshwater fish had become established in
Hawaiian waters (Devick 1991). These introductions have resulted from deliberate release by government agencies, or by casual release of domestic aquarium fish. The poeciliid fish suspected of having the greatest
3 impact on Oahu's Megalagrion damselflies, however, were those introduced for biological control of mosquitoes prior to 1923.
Another potential concern for native Megalagrion damselflies has been the introduction of alien Odonata species. Three damselfly and five dragonfly species have been accidentally introduced to O'ahu since 1936
(Zimmerman 1948; Harwood 1974; Nishida 1997). It is possible that introduced Odonata could have negative impacts on native Megalagrion species through competition or predation, and may be an additional or alternate reason that native damselflies are absent from lowland areas. Thus, in conjunction with surveys of native damselflies and introduced fish, distributions of introduced Odonata were also recorded.
Three species of amphibians have become successfully established on O'ahu since the 1800s (Devick 1991).
Two frog (Rana rugosa Schlegel, 1838, Rana catesbeiana Shaw, 1802) and one toad (Bufo marinus
(Linneaus, 1758)) species were common from sea level to even the highest elevations of O'ahu.
Distributions of introduced amphibians in stream and wetland areas were also recorded.
STUDY AREA
The study area covers lentic, lotic, coastal wetland and estuarine habitats on the island of 0'ahu, with 0'ahu freshwater habitat types classified according to Polhemus and Asquith (1996). One of the high islands in the
Hawaiian archipelago, O'ahu has 57 recognized perennial streams (Hawaii Stream Assessment 1990), with the longest, Kaukonahua, being 50 km long, although most are much shorter. The headwaters or sources of these streams vary in elevation from 1219 m on Mt. Ka'ala (Makaha and Hale'au'au Streams) to large springs emerging only 1.0-2.5 m above sea level in the Pearl Harbor area (Kapakahi Stream, Waiawa
Springs, Waiau Springs).
For this study, we divided O'ahu into three major geographic units (Figure 1.1). The wetter windward areas
of O'ahu, which contain the majority of the island's perennial streams, were separated into Kane'ohe Bay
4 OAHU
r-··· ---1 Leeward & I'earl Harbor 10km Streams Northern Windward Shaded areas indicate relief Streams above 600 m. elevation
21°30' 21°30'
158°00'
Figure 1.1. Limnological divisions on O'ahu defined for the purposes ofthis study. Bold line indicates hydrological boundary between leeward and windward (northeastern) O'ahu, cross hatching indicates surveyed northern windward and Kane'ohe Bay streams. Streams containing Megalagrion populations are numbered on this map as follows:
Kahalu'u = 1 Kaluanui = 7 Dillingham = 13 Kaukonahua = 19 Waihe'e=2 Ma'akua=8 Hale'au'au (Mt. Ka'ala) = 14 Waiawa=20 Ka'alaea= 3 Kaipapa'u = 9 Makaha (Mt. Ka'ala) = 15 Halawa=21 Waiahole =4 Koloa = 10 Kaupuni = 16 Moanalua = 22 Waikane= 5 Wailele = 11 Helemano = 17 Kalihi = 23 Kahana = 6 Kahawainui = 12 Poamoho = 18
5 and northern windward drainages, while the drier leeward Ko'olau, Pearl Harbor, and Wai'anae Mountain watersheds were combined into a third major hydrological unit. Most streams on O'ahu are naturally interrupted, with perennial flow only in the mountain headwaters and in low-elevation areas near the ocean
(Polhemus et ill. 1992). This is a natural condition caused by percolation of water into the alluvium in the mid-reaches. Such streams exhibit surface flow throughout their length only during periods ofextended precipitation, and streamflow is characteristically flashy, with high flood peaks and low baseflows (Nichols et a1. 1997).
METHODS
Damselfly sampling commenced in 1992 and continued through 1998; whenever possible, streams were sampled more than once to confirm aquatic species composition. As O'ahu streams are short, it was often possible to completely survey a stream from the ocean to the headwaters, especially those originating at low elevations. Sampling sites were also dependent on local terrain and access to private property. Altitude at each sampling station was determined by using a combination of topographic maps and a hand-held altimeter. Fish and damselfly species composition was recorded at each sampling site. A summary of sampling sites, capture locations, stream type (permanent or intermittent), elevations surveyed, and elevations where introduced fish and damselflies were found is provided in Appendix 1.
Damselfly Sampling
Sampling focused solely on stream-dwelling damselflies on O'ahu, and did not include terrestrial damselfly species such as Megalagrion oahuense (Blackburn, 1884) or Megalagrion koelense (Blackburn, 1884). Adult
Odonata were collected mainly with aerial nets. Immature damselflies were collected by benthic sampling with aquatic dip nets. Most damselfly collections were of adults, as it was usually difficult to locate
immature individuals. Voucher specimens have been deposited in the Bishop Museum and Smithsonian
Institution collections. Aquatic habitat type, such as riffle, run, pool, wetland or estuary (mixed fresh and
6 saltwater), was recorded at each site to ascertain habitat preferences for native Megalagrion damselflies.
Immature damselflies were identified according to Polhemus and Asquith (1996).
Immature Damselfly Behavior
Inferences obtained from immature damselfly response to the threat of fish predation could lead to a behavioral explanation as to why native damselflies have become so rare. To observe behavioral interactions with introduced fish, ten immature O'ahu Megalagrion xanthome1as larvae were placed in a 40 liter aquarium containing ten Gambusia affinis and ten Poecilia mexicana Steindachner, 1863. Additionally, several species of introduced Odonata such as Ischnura ramburii (Selys-Longchamps, 1850) and Orthemis ferruginea (Fabricius, 1775) were also placed in this same aquarium. The response ofthe immature Odonata
(such as swimming away, keeping still, or hiding) to predation attempts by the introduced poeciliid fish was observed.
Fish Sampling
Fish species composition was assessed through seining, netting, snorkeling, and above-water observations.
Fish species composition and number of individuals captured were also recorded during seining. Underwater visual observations using mask and snorkel were made at sites with sufficient water depth and clarity. The stomach contents ofeight Gambusia affinis and four Poecilia mexicana were examined for immature damselflies.
RESULTS
Native Damselflies
There has clearly been a significant island-wide decline in the distribution of O'ahu Megalagrion species when compared to historical records. Resurveys during this study in the 1990s found that almost none of the original collecting sites of early collectors such as R.C.L. Perkins, FX. Williams, and others contained native damselfly populations (Table 1.1). Appendix 2 contains specific O'ahu site locality information for
7 Table 1.1. Extinctions of Megalagrion species in surveyed O'ahu aquatic habitats since 1936.
M. hawaiiense M. leptodemas M. n. nigrolineatum M. oceanicum M. pacificum M. xanthomelas
Collector (Years) Original Currentl Original Current' Original Current! Original Current! Original Current' Original Currentl
Perkins (1892-1912) 5 1 5 0 8 0 11 0 3 0 4 0
Williams (1925-1936) 1 0 3 0 1 0 2 0 3 0
Timberlake (1919-1923) 1 0 2 0
Others (1901-1926) 7 3 0 0
Total: 6 1 8 0 16 0 17 0 3 0 10 0
lCurrent number ofextant damselfly populations in exactly or approximately the same aquatic habitats where collected prior to 1936, see Appendix 2 for sites collected by R.C.L. Perkins and F'x. Williams.
60 i I
00 50
en E 40 til ~ en '030 Qi .0 E ::J 20 Z
10
E o Oahu Streams hawaiiense leptodemas nigrolineatum pacificum oceanicum xanthomelas
Figure 1.2. Status of stream and wetland dwelling damselflies on the island of O'ahu. E = Extinct on O'ahu Table 1.2. Remnant native Megalagrion species found in O'ahu streams, tributaries in parentheses, and relative abundance in stream areas containing native damselflies, (rare (R) = < 3 individuals collected or observed); (moderately common (C) = ~ 3 individuals collected or observed). Kane'ohe Bay Native Damselfly Species and Relative Abundance Kahalu'u M. hawaiiense (R), M. nigrohamatum nigrolineatum (C), M oceanicum (R) Waihe'e M. nigrohamatum nigrolineatum (C), M. oceanicum (R) Ka'alaea M. nigrohamatum nigrolineatum (C) Waiahole (Waianu tributary) M hawaiiense (R) Waiahole (Uwau tributary) M. nigrohamatum nigrolineatum (R) Waikane M. hawaiiense (C), M. nigrohamatum nigrolineatum (C)
Windward Kahana M. hawaiiense (C), M. nigrohamatum nigrolineatum (C), M. leptodemas Kaluanui M. nigrohamatum nigrolineatum (C), M. oceanicum (C) Ma'akua M. leptodemas (C), M oceanicum (R) Kaipapa'u M hawaiiense (C), M oceanicum (R) Koloa M. hawaiiense (R), M. nigrohamatum nigrolineatum (C), M. oceanicum (R) '" Wailele M. oceanicum (R) Kahawainui M. oceanicum (C)
LeewardlPearl Harbor Hale'au'au (Mt. Ka'ala) M. hawaiiense (C) Miikaha (Mt. Ka'ala) M. hawaiiense (C) Dillingham (unnamed) M hawaiiense (C) Kaupuni (Honua tributary) M hawaiiense (R) Helemano M. hawaiiense (R), M. nigrohamatum nigrolineatum (C) Poamoho M. nigrohamatum nigrolineatum (C) Kaukonahua M. nigrohamatum nigrolineatum (C) Waiawa M. nigrohamatum nigrolineatum (C), M.leptodemas (C) Halawa M hawaiiense (C), M. leptodemas (C), M nigrohamatum nigrolineatum (C), Moanalua (TAMC) M. xanthomelas (C)' Kalihi M. nigrohamatum nigrolineatum (C) 'Found in only 95 m of artificial stream habitat on O'ahu Megalagrion damselflies collected by Perkins and Williams. Most of the remaining native damselfly populations on O'ahu were instead found in remote upper elevation areas that were not surveyed by early collectors.
Fifty-five of fifty-seven O'ahu streams and most coastal wetland areas were surveyed in whole or part.
Highly fragmented damselfly populations remained in only twenty-two streams (Figure 1.2); these streams are listed in Table 1.2. Due to rugged terrain, the upper elevation sections of many streams were difficult to access, and some headwater reaches in the Ko'olau Mountains above 350 m have not yet been surveyed.
These areas are likely to harbor a few additional remnant populations.
Five of the six native stream or wetland dwelling damselflies still persisted on O'ahu (Table 1.3), with one species, Megalagrion pacificum (McLachlan, 1883), now apparently extinct on O'ahu, and another,
Megalagrion xanthomelas, near extinction. Three species (Megalagrion hawaiiense (McLachlan, 1883),
Megalagrion oceanicum McLachlan, 1883, and Megalagrion leptodemas (Perkins, 1899) were rare and found in twelve or fewer streams. Two of these species are O'ahu endemics (M oceanicum and M.leptodemas), and were found in six or fewer small streams. Populations of M. hawaiiense, currently still common on all the main Hawaiian Islands except O'ahu, were found in low numbers in 12 stream and wetland areas.
Megalagrion nigrohamatum nigrolineatum (Perkins, 1899) was the most common native stream damselfly collected, and was found in high elevation sections of 15 streams, mainly in windward areas.
Kane'ohe Bay Streams
Only three of the six remaining O'ahu Megalagrion species were observed in Kane'ohe Bay streams. Of these species, fragmented colonies of Megalagrion nigrohamatum nigrolineatum were found in upper elevation reaches of six catchments (Figure 1.3), all above barriers that precluded the upstream movement of poeciliids. Megalagrion nigrohamatum nigrolineatum, a side-channel and calm water dweller in its larval stages, was the most common species in these Kane'ohe Bay streams (Table 1.2). With the exception of
10 Table 1.3. Biogeographic status and Hawaiian Island distribution of aquatic Megalagrion and introduced Odonata species found on O'ahu.
Biogeographic Current Known Island Historical Island Species Status! Distribution Distribution Megalagrion hawaiiense Hawaiian Endemic All Major Hawaiian All Major Hawaiian Islands Islands Megalagrion leptodemas O'ahu Endemic O'ahu O'ahu Megalagrion nigrohamatum O'ahu Endemic O'ahu O'ahu nigrolineatum Megalagrion oceanicum O'ahu Endemic O'ahu O'ahu Megalagrion pacificum Hawaiian Endemic Maui, Moloka'i, All Major Hawaiian Hawai'i Islands Megalagrion xanthomelas Hawaiian Endemic O'ahu, Maui, All Major Hawaiian Moloka'i, Hawai'i Islands Enallagma civile Introduction (1936) All Major Hawaiian Islands Ischnura posita Introduction (1936) All Major Hawaiian Islands Ischnura ramburii Introduction (1973) All Major Hawaiian Islands Crocothemis servilia Introduction (1994) O'ahu Orthemis ferruginea Introduction (1976) All Major Hawaiian Islands Pantala hymenaea Introduction (1989) O'ahu Tramea abdominalis Introduction (1977) O'ahu Tramea lacerata Introduction (1874) All Major Hawaiian Islands Iprom Polhemus and Asquith (1996)
Waikane Stream, M. hawaiiense was uncommon throughout the Kane'ohe Bay watershed. For instance, only a single M. hawaiiense was found in Waiahole Stream at 244 m elevation, well above the range of introduced fish. Megalagrion oceanicum were found in restricted areas ofonly two Kane'ohe Bay streams.
Megalagrion damselflies were completely missing from the lower reaches of all Kane'ohe Bay streams, and
associated low elevation coastal wetland areas. In addition, all taro (Colocasia esculenta) wetland areas in the
Kane'ohe Bay drainage were surveyed. Although historically found in taro fields and lowland areas (Moore
and Gagne 1982), Megalagrion damselflies were never observed in any O'ahu taro field during this study,
despite some areas being intensively sampled, e.g., fifty sampling trips made between 1994 and 1998 to the
Waiahole Stream watershed and adjacent taro fields.
11 300 .,.------....., 275 Megalagrion = 250 225 Poeciliids = • E 200 -; 175 i5 150 Figure 1.3. Elevational distribution for Hawaiian Megalagrion damselflies and introduced poeciliid fish for streams entering Kane'ohe Bay, O'ahu. Northern Windward Streams Four Megalagrion species were observed in northern windward streams. Megalagrion oceanicum was found in five northern windward streams, while M. hawaiiense and Megalagrion nigrohamatum nigrolineatum were each found in three streams (Table 1.2). Megalagrion leptodemas was found in two catchments: Kahana Stream and a short section of Ma'akua Stream. Although native damselflies were generally found in the upper elevations of northern windward streams (Figure 1.4), an exception occurred at Kahana Stream, where damselflies were found in relatively low-elevation areas (30 m). The Kahana watershed has numerous, small high-gradient tributaries and seeps lining the steep valley walls, and this was the lowest elevation site at which any Megalagrion damselflies were found on O'ahu. 12 800 ,------, Megalagrion = 700 I PoecilUds :::: 600 ffi §: 500 c: ~ 400 ~ + iIi 300 200 100 o Kahana Kaluanui Maakua Kaipapau Koloa Wailele Kahawainui Figure 1.4. Elevational distribution for Hawaiian Megalagrion damselflies and introduced poeciliid fish for streams entering northern windward O'ahu. Megalagrion damselflies were completely missing from all stream estuary and coastal wetland areas surveyed in and near northern windward streams. By contrast, relatively large populations ofsome species were found in the upper reaches of most of the same catchments. Malaekahana Stream was the exception; although this stream maintained perennial flow from 43-244 m elevation, neither introduced fish nor Megalagrion damselflies were found there. Pearl HarborlLeeward Streams Native damselflies were completely absent from the lower sections of Pearl Harbor and leeward streams, and from the large set of emergent basal springs in the Pearl Harbor region. Four of the six stream-dwelling Megalagrion species were observed in other leeward O'ahu streams, mainly in upper elevation areas (Figure 1.5). One exception was the presence of M. xanthomelas at the relatively low elevation of 79 m in a tributary of Moanalua Stream at the TripIer Army Medical Center (TAMC). Megalagrion nigrohamatum 13 nigrolineatum was the most common native species in leeward streams, with other damselflies relatively more rare (Table 1.2). Three species ofstream-dwelling Megalagrion damselflies were found in North Halawa Stream, which contained the most diverse assemblage of native lotic damselflies ofany leeward O'ahu stream. Two damselfly species were found in Waiawa Stream (from 213 to 335 m), M. nigrohamatum nigrolineatum and M.leptodemas. Between them, North Halawa and Waiawa Stream 550 .,...------::::------, 500 cIJ Megalagrion =I ~ Q 400 Poeciliids = CJ 350 300 0 250 l. ... 200 150 100 50 0+------t.J+--+--44--f---+-+--+--f-i-+----+-+4--+--t::::t--+--1-'+--l--t-'+---j Helemano Waikele Kapakahi Waiawa Waimalu Kalauao Halawa Moanalua Figure 1.5. Elevational distribution for the Hawaiian Megalagrion damselflies and introduced poeciliid fish for streams entering Pearl Harbor and leeward O'abu. accounted for two of the four known remaining M. leptodemas populations. Megalagrion hawaiiense was found in six widely separated leeward and Pearl Harbor watersheds, but was only common in the Mt. Ka'ala (elevation 1150 m) bog area, and in a small stream (near Dillingham Airfield) draining Mt. Ka'ala. Damselfly Behavioral Observations To observe the behavior of immature Megalagrion in the presence of poeciliids, larvae were placed in an aquarium with Poecilia mexicana and Gambusia affinis. In these aquarium observations, all 10 Megalagrion xanthomelas larvae swam to the water's surface, and smaller individuals were eaten whole and larger individuals were picked apart by Gambusia affinis and Poecilia mexicana. This resulted in 100% mortality 14 of M. xanthomelas larvae within 10 minutes of them being placed in the aquarium. This contrasts with the diving-to-the-bottom and hiding-in-the-substrate behavior of the introduced Odonata Ischnura ramburii and Orthemis ferruginea larvae in the same aquarium. Neither of these introduced species exhibited mortality from poeciliids after remaining in the aquarium for one day. The behavior of Megalagrion paciflcum larvae (from Moloka'i) was similar to that exhibited by O'ahu M. xanthomelas, as these larvae always swam to the aquarium surface when disturbed. Swimming to the surface by Megalagrion larvae appeared to be a completely ineffective escape response behavior against the surface-oriented poeciliid predators, since the native damselfly larvae were immediately consumed. By contrast, the lack ofmovement and hiding behavior by immature Ischnura rambuni appeared to reduce or eliminate successful attacks by poeciliids in aquariums. Introduced Fishes The introduced poeciliids Poecilia mexicana, Xiphophorus helleri, Poecilia reticulata, and Gambusia affinis were common, and were found in every low-elevation and coastal aquatic habitat surveyed on O'ahu. Introduced poeciliids such as Poecilia mexicana, Xiphophorus hellen, Poecilia reticulata, and Gambusia affinis were abundant in all 0'ahu streams, with some species occurring to an elevation of nearly 400 m. There were also three relatively uncommon poeciliid species mainly restricted to still water habitats such as springs, marsh areas, and estuarine or brackish water areas. These are the Cuban limia (Limia cf. vittata), platy (Xiphophorus maculatus), and sailfin molly (Poecilia latipinna). The following ~ection summarizes the distributional range and impacts of the most common poeciliid species found on 0'ahu. Poecilia mexicana This species apparently resulted from hybridization, possibly in the wild in its native range prior to its Hawaiian introduction (R.R. Miller and W. Fink, personal communication). However, Poecilia mexicana (shortfin molly) now exhibits stable morphological characters readily distinguishing it from other poeciliids found in Hawai'i. This species was the largest of the introduced poeciliids, up to 115 mm in length. 15 Poecilia mexicana inhabited waters with a wide range of salinities, from 0 ppt to 40 ppt (ocean water in Hawai'i has a salinity of 36-37 ppt). Because of this salinity tolerance, these fish were found in every coastal estuary and low-elevation wetland. In general, Poecilia mexicana was found in estuaries, coastal wetlands, and in the lower «70 m elevation) reaches of streams. One notable exception was the collection of Poecilia mexicana at the source (elevation 400 m) of Kipapa Stream, a tributary of Waikele Stream. The Waikele Stream system has a gentle gradient with no barrier waterfalls below 400 m elevation. As in other O'ahu streams, Waikele Stream water flow disappears into the alluvium and dries up in its mid-reaches during periods of low precipitation. A water connection (combined with a low stream gradient) through the often dry mid-reaches (120 to 240 m elevation) of the Waikele Stream system was clearly maintained long enough to enable P. mexicana to colonize the uppermost headwaters of this stream. Xiphophorus helleri Xiphophorus helleri (green swordtail) was found only in freshwater and was always found above areas of tidal influence. Xiphophorus helleri was generally (with the exception ofWaikele Stream) found higher than P. mexicana, and was found upstream as far as the fIrst barrier falls or rapids. Xiphophorus helleri were abundant in pools and also in relatively high water velocity habitats, including runs and slower sections of riffles. The largest Xiphophorus helleri measured was 105 mm long, attaining a size nearly double that of other common poeciliids such as Poecilia reticulata and Gambusia affinis. The highest elevation at which this fish was found during this study was 204 m at Poliwai Gulch, a leeward drainage with a gradual stream profIle. Gambusia affinis Gambusia affinis (western mosquitofish) was recorded at a maximum length of 56 mm, and was always found in high densities in the lower stream reaches, usually in the presence of Poecilia mexicana. The 16 highest elevation at which Gambusia affinis was found was at 185 m in Poliwai Gulch, near the Waiahole Ditch. Gambusia afflnis also exhibited a wide salinity tolerance of 0 to 40 ppt, but was usually found in areas having salinities of 16 ppt or less. Ten Gambusia affinis stomachs from Pearl Harbor streams were examined and found to contain various aquatic fauna such as chironomids, shrimp, and ants. Native damselflies were not found in these stomachs. It was not surprising because of the complete lack of native damselflies in all areas where Gambusia afflnis was found. Poecilia reticulata Poecilia reticulata (guppy) and Gambusia afflnis are similar in size and ecological requirements, and these were also the smallest of the introduced poeciliids found on 0'ahu. In comparison with Gambusia affinis, Poecilia reticulata were restricted to water of s 4 ppt salinity. Poecilia reticulata were also generally found at the highest elevations of all the introduced fish. This species penetrated the headwater source areas of Hakipu'u, Kawa, and Kane'ohe Streams, none of which were observed to contain native Megalagrion in their uppermost reaches. Poecilia reticulata was found up to 400 m elevation, and was sympatric with Poecilia mexicana in the Waikele Stream system (a Kipapa tributary). This species was generally more common than Gambusia affinis in the upper reaches of streams, but absent from estuarine areas. Tilapia Introduced into Hawai'i in 1951, tilapia (Sarotherodon melanotheron Ruppell, 1852) is considered to be a harmful and dominant introduced fish in low-elevation Hawaiian streams, wetlands, and estuaries (Nelson and Eldredge, 1991). We found high densities of tilapia in both limnetic and estuarine habitats, in water ranging from 0 to 40 ppt salinity. In contrast to poeciliids, tilapia have little ability to colonize areas of high stream gradient, but were found in high numbers in still-waters of low elevation streams, estuaries, and wetlands. 17 Introduced Odonata The introduced damselflies Ischnura posita (Hagen, 1862), Ischnura ramburii, and Enallagma civile (Hagen, 1862) were common in low elevation areas containing high densities of introduced poeciliid fish and tilapia; these same areas were devoid of native Megalagrion damselflies (Figure 1.6). For example, extensive 250 225 Ischnura = 200 Poeciliids = 175 E"-" c: 150 .9 -a;ell 125 ill 100 75 50 25 0 Kawa Kaneohe Kahaluu Waihee Kaalaea Waiahole Moanalua Waikele Figure 1.6. Elevational distribution for introduced Ischnura ramburii and Ischnura posita damselflies and introduced poeciliid fish for selected Kane'ohe Bay and Pearl Harbor streams, O'ahu. surveys of Pearl Harbor and Kane'ohe Bay wetland habitats found only these introduced damselflies with corresponding high densities of Poecilia mexicana, Gambusia afflnis, and Poecilia reticulata. In some cases, the distribution of introduced damselfly species did overlap with that of Megalagrion damselflies. For example, Megalagrion nigrohamatum nigrolineatum, M. oceanicum, and Ischnura ramburii exhibited a near-identical elevational distribution in Waihe'e Stream. Four introduced species ofOdonata (Ischnura posita, Tramea abdominalis (Rambur, 1842), Orthemis ferruginea, and Crocothemis servilia (Drury, 1770)) also co-existed with adults and larvae of the last Megalagrion xanthomelas population on 18 0'ahu at the TAMC Stream, a tributary of Moanalua Stream. Introduced fish were absent at the TAMC stream. Relative Distributions of Damselflies and Introduced Fish Little overlap in the distribution ofadult Megalagrion damselflies and other introduced fish was found in streams entering Kane'ohe Bay (Figure 1.3). For example, poeciliids were found in Kahalu'u Stream up to an impassable 2 m high concrete barrier encountered at 80 m elevation; above this barrier Megalagrion nigrohamatum nigrolineatum was present. In Kahana Stream, a few adult Megalagrion nigrohamatum nigrolineatum were recorded down to the maximum limit ofintroduced Xiphophorus helleri at approximately 33 m elevation, although they may have strayed temporarily from higher elevations. A small amount of overlap from the 52-61 m elevation level was also found between these two species in Waihe'e Stream. Only adult Megalagrion damselflies were captured from lower elevation areas of Kahana and Waihe'e Streams, and immature Megalagrion were not found in this small area of overlapping distribution. Thus, distributional overlaps in both streams ofintroduced fish and native damselflies may be explained by downstream adult damselfly flight. In upper Kahalu'u Stream, Megalagrion oceanicum, Megalagrion nigrohamatum nigrolineatum. and Megalagrion hawaiiense were found above two forks of the stream, with one fork containing a 2 m high concrete barrier and the other a series of high gradient cascades 2-3 m in height. A slight overlap in the distribution of Poecilia reticulata and adult Megalagrion nigrohamatum nigrolineatum was observed for 70 100 meters in the Waihe'e Stream channel. This was most likely caused by downstream flight of adult Megalagrion nigrohamatum nigrolineatum, which were common in upper Waihe'e Stream. Poecilia mexicana and Megalagrion distributions were always mutually exclusive. Northern windward O'ahu streams also had a dissimilar elevational distribution of introduced fish and Megalagrion damselflies (Figure 1.4). As with most Kane'ohe Bay streams (with the exception of Kahana), 19 the surveyed northern windward O'ahu streams were naturally interrupted with water flow sinking into the alluvium in the lower to mid-reaches (Polhemus et a1. 1992). This gap in water flow explains the lack of both damselflies and introduced fish shown in Figure 1.4 in the lower to mid-reaches of most of these streams. Stream flow in the interrupted sections occurs only for short times during periods of intense precipitation. Despite this, native stream biota such as fish, crustaceans, and mollusks were still found in the upper reaches of surveyed northern windward streams. By contrast, the relatively long dry sections and high gradient of these streams apparently preclude poeciliid colonization of their upper reaches (areas above ca. 210 m elevation). Kaipapa'u, Koloa, Ma'akua, Wailele, and Kahawainui were all dry in the lower to mid-reaches, but the upper reaches of these streams contained robust native damselfly populations, with Megalagrion oceanicum occurring in all five streams. By comparison, the terminal reaches, estuaries, and adjacent wetland areas of all northern windward O'ahu streams were found to have high densities of introduced poeciliid fish, and lacked native damselflies. The upper reaches of Punalu'u Stream, the only continuously flowing northern windward drainage, were not assessed, since access was denied by the landowner. In lower Punalu'u Stream, however, high densities of tilapia and Poecilia mexicana, were found, and native damselflies were not observed. Pearl Harbor drainages and leeward streams were found to be the most affected by introduced fish, possibly due to low stream gradients and general lack of waterfall barriers, which allow such fish to access the entire lengths of many of these streams. Introduced fish were not observed in the surveyed upper reaches surveyed (490 m) of Helemano Stream, but Megalagrion nigrohamatum nigrolineatum was common there. The most dramatic example of non-overlapping distribution of alien fish and native damselflies on O'ahu was the presence of a remnant population of Megalagrion xanthomelas in a small (35-70/1 min flow) unnamed artificial tributary of Moanalua Stream at !he TripIer Army Medical Center (TAMC). Poeciliids have been unable to access this tributary because of a 7 m high culvert at the downstream end of this section of stream. Megalagrion xanthomelas was not found during surveys of adjacent streams or in downstream sections of Moanalua Stream, all of which had high densities of introduced fish. 20 Pearl Harbor streams are also among the shortest on O'ahu, and this is reflected in the elevational distribution of introduced fish for some of the streams (Figure 1.5). Many of these streams flow in their mid- and headwater reaches only during periods ofextended precipitation, and maintain a perennial flow only downstream ofemergent springs in an area of caprock around Pearl Harbor (Stearns and Vaksvik 1935). Kapakahi, Honouliuli, Kalauao, and the Waiawa Springs outlet, are short, spring fed streams that flow for distances of 50-350 m before entering Pearl Harbor. These limnocrenes contain high densities of introduced poeciliids, and three species of introduced damselflies are common in this area. Native damselflies were not found in any of these wetland areas surrounding Pearl Harbor. Discussion Introduced Fishes Predator-prey relationships between fish and Odonata have been extensively studied in temperate regions, and fish have been shown to affect the distribution of many species of Odonata (Nilsson 1981; Pierce 1988; Henrikson 1988; McPeek 1989; 1990a; 1990b). In Sweden, experiments found that dragonfly larvae inhabiting only fishless lakes and acid bog" ponds attempted to actively escape from fish predators, while species from lakes containing fish feigned death and remained still when attacked (Nilsson 1981; Henrikson 1988). This naIve larval escape response to the threat of predation was similar to that observed for Hawaiian Megalagrion larvae both in the field and in aquarium observations conducted during this study. McPeek (l990b) identified similar naIve behaviors in Enallagma damselflies from fishless lakes, which attempted to swim away from predators. While this behavior was successful against attacks by immature dragonflies, it failed to deter fish predation. Poeciliid predation was also correlated with the elimination of native Odonata species in Australian lakes; lakes containing Gambusia affinis had only 3-4 species of Odonata, while nearby lakes without introduced fish contained 11 species (Davis et a1. 1987). In the presence of mosquitofish, declines in Odonata 21 populations were also found in California rice fields (Farley and Younce 1977). Although most research on the negative effects of poeciliids on native species has involved Gambusia affinis and Xiphophorus helleri, Arthington and Lloyd (1989) believed the carnivorous Poecilia reticulata might have impacts similar to that of Gambusia affinis on native aquatic biota. Megalagrion xanthomelas and Megalagrion pacificum, two low-elevation stream and wetland damselflies, are the species that appear to have been most adversely affected by introduced poeciliids. These damselflies are obligate coastal and low-elevation species, and poeciliids have invaded all of these habitats on O'ahu, with the exception of the artificial stream at TAMC. Although Megalagrion xanthomelas is still found on other islands, the O'ahu population maintains a precarious existence. Only 100-160 adults can be observed at anyone time in the 95 m (long) artificial TAMC stream (Englund 1998). Similarly, an analysis of Megalagrion pacificum collection records by Liebherr and Polhemus (1997) indicated this species was likely extirpated on O'ahu by 1920. The current absence of native damselflies in lowland coastal areas of O'ahu is clearly not due to a lack of suitable aquatic habitats. For example, large amounts of formerly suitable native damselfly habitat are still found adjoining springs in the Pearl Harbor region, with watercress (Nasturtium microphyllum) and taro cultivated here in large quantities. These spring areas also feed an extensive fresh and brackish water coastal wetland system that is now completely devoid of native damselfly species. Native damselflies such as Megalagrion xanthomelas that formerly were found in lowland stream areas, coastal wetlands, and brackish estuarine waters on O'ahu in areas of up to 8-15 ppt salinity (Polhemus 1996; Nishida 1997) are now completely extirpated, while introduced Odonata are abundant. All taro fields in the Kane'ohe Bay watershed were surveyed and were found to contain high densities of not only Poecilia mexicana, but also Poecilia reticulata, and Xiphophorus helleri. Neither Megalagrion pacificum or Megalagrion xanthomelas were found in any O'ahu taro fields, although both ofthese species were recently found in Moloka'i taro fields lacking introduced fish (R.A. Englund and W. Puleloa, unpublished). Thus, lowland native damselfly species appear 22 to be adaptable to artificial wetlands created by taro cultivation, but are not found in taro fields containing introduced fish species. Another endemic O'ahu species, Megalagrion oceanicum, has also disappeared from most streams, occurring only in high-elevation stream areas lacking introduced poeciliids. However, it is known that M. oceanicum formerly occurred in low-elevation stream habitats and was originally collected by Perkins in 1892 near sea level (Perkins 1899). This species also appears to be at high risk of extinction, and was usually only observed in low numbers. Additionally, the larval preference ofM oceanicum for fast-water habitats has made it vulnerable to loss of habitat through water diversions and channelization. With the exception of the bogs and small rivulets of Mt. Ka'ala, the highest mountain on the island, Megalagrion hawaiiense was also uncommon on O'ahu, and few individuals were observed in the streams where this species was found. This contrasts to its status as one of the most abundant native damselflies on other Hawaiian Islands, in habitats lacking introduced fish. On O'ahu, poeciliids were found in high densities in some of the habitats preferred by Megalagrion hawaiiense larvae, including springs, cut-off stream channel areas, and seep areas. Similarly to Megalagrion oceanicum, Megalagrion hawaiiense was originally collected by Perkins on O'ahu in 1896 at "no great elevation above the sea" (Nishida, 1997), and is still found near sea level on other islands, e.g. Moloka'i (R.A. Englund and W. Puleloa, unpublished). The lowest elevation this species was found at during the present study was 137 m on Kahalu'u Stream, which was above the range of introduced poeciliids for that stream. Megalagrion leptodemas, another O'ahu endemic, is known from scattered colonies in the upper elevation sections offour widely separated windward and leeward streams. Due to its overall scarcity, Megalagrion leptodemas may be the stream-dwelling damselfly species at greatest risk ofextinction in the Hawaiian Islands. No populations of Megalagrion leptodemas exist at sites where this species was collected prior to 1936 by Perkins and Williams (see Table 1.1), and these stream areas are now heavily infested with 23 poeciliids. Immature M. leptodemas favor areas of standing pools and slow water (Polhemus and Asquith, 1996) that are also favored by introduced poeciliids. This study found that native damselflies adapted to higher elevations have fared better than species restricted to lower elevations. For example, Megalagrion nigrohamatum nigrolineatum was found in the upper elevation sections of fifteen O'ahu streams, and was usually common in these areas. This species may have remained more abundant than other Megalagrion species on O'abu because of its broader habitat preferences. Megalagrion nigrohamatum nigrolineatum was locally common in high-elevation habitats in both leeward and windward drainages, unlike Megalagrion xanthomelas or Megalagrion pacificum, species preferring low elevation habitats. This may be due to the preference of Megalagrion nigrohamatum nigrolineatum larvae for side pools and slow-water channels in high elevation areas that are usually wetter and more shaded. Because no Megalagrion damselflies were found in poeciliid stomachs, this study did not establish a direct causal relationship between poeciliid predation and the island-wide decline of native stream damselflies on O'abu. However, the effects of introduced poeciliids on invertebrates have been widely documented, and predation by Gambusia affinis and other poeciliids is believed to be the primary cause of native species extirpation or reduction in other locations (Courtenay and Meffe 1989). On Hawai'i Island, an endemic hypogeal shrimp species was found in high densities in 187 pools lacking introduced poeciliids, but co occurred with poeciliids in only five of thirty-three pools, and then in very low numbers (Maciolek 1984). On Kaua'i, Megalagrion damselflies are no longer found within the Hanalei River National Wildlife Refuge, another area with high densities of introduced tilapia and poeciliids (Polhemus and Asquith 1996). The apparent susceptibility of immature Megalagrion species to poeciliid predation may be due to a combination of naive behavior and evolution to avoid benthic feeding by native Hawaiian stream fish in the families Gobiidae and Eleotridae. Way and Burky (1991), for instance, found that Odonata (dragonflies and damselflies combined) comprised 13% of the diet of the native stream goby Lentipes concolor. Furthermore, 24 most larval Megalagrion are not found in the benthic habitats most preferred by native insect-eating gobies, but instead live in habitats such as shallow side-channels, isolated side-pools, the highest velocity areas of riffles, or seeps, rheocrenes, and waterfalls (Polhemus 1997). These habitat preferences may in part be due to evolutionary predation pressure from native stream fish. Currently, introduced poeciliids are found in high densities in many of these preferred larval Megalagrion habitats throughout the Hawaiian Islands. Isolated Hawaiian streams, such as those on the north shore of Moloka'i, have healthy populations of native gobies and lack introduced fish; these areas also contain the greatest number of native damselfly species (R.A. Englund and W. Puleloa, unpublished). At the TAMC stream, immature M xanthomelas were frequently observed on the top of the stream substrate and free swimming at surface of shallow, small pools (Englund 1998), habitats unlikely to have native, benthic gobies. Introduced Odonata In the past century, five species of dragonflies and three species of damselflies have been accidentally introduced into Hawai'i (Nishida, 1997). It has been hypothesized that negative interactions could occur between these newly arrived species and native damselfly species. We found overlapping distributions of Megalagrion xanthomelas, one introduced damselfly, and three introduced dragonfly species in the 95 m section of the TAMC stream. Megalagrion xanthomelas also co-exists with other introduced dragonflies and damselflies on Lana'i and Hawai'i Islands (Polhemus, 1996). On Liina'i, for instance, Polhemus (1996) could not correlate the presence of introduced or native Odonata with the absence Megalagrion xanthomelas. Similarly, in lowland coastal wetland and riverine habitats of Moloka'i, the introduced Orthemis ferruginea, Ischnura ramburii, and Ischnura posita were found sympatrlcally with Megalagrion pacificum and Megalagrion xanthomelas (Englund 1999). This study found little or no evidence that introduced Odonata impact native damselfly distributions. 25 Introduced Amphibians Liebherr and Polhemus (1997) indicated that introduced amphibians might be one of the reasons for the decline of Megalagrion, and at least three species of amphibians have become successfully established on O'ahu since 1896 (Devick 1991). This study found no evidence ofintroduced amphibians impacting native Megalagrion. The effects of the two frog (Rana rugosa, Rana catesbeiana) and one toad (Bufo marinus) species on Megalagrion damselflies are unknown; however, frogs and toads are common even in the most isolated and pristine aquatic habitats of the Hawaiian archipelago. For example, all six species of stream Megalagrion are abundant in areas of Moloka'i containing introduced amphibians, but lacking introduced fish (Englund and Puleloa unpublished). Toads were commonly observed in the 95 m reach of stream at TAMC, and during the present study both frogs and toads were found to co-occur with Megalagrion damselflies throughout O'ahu. Thus, although it is possible that these introduced amphibians have some detrimental impacts on both immature and adult damselflies, no clear adverse impacts were ascertained. Conclusions Introduced poeciliids present potential threats to the biodiversity of aquatic ecosystems throughout the Hawai'i. Megalagrion damselflies are now absent from virtually all lowland areas where early collections occurred prior to poeciliid introductions. Protective measures will be necessary to prevent the decline of remaining upland populations, as translocations offish by people to areas above waterfall barriers has already occurred in a number of Hawaiian streams. To preserve native aquatic invertebrate biodiversity in Hawai'i, research on the elimination ofthese introduced fishes should have high priority. Chemical or rotenone treatment in selected stream, spring or wetland areas could eliminate alien fish species and allow establishment of additional populations of now rare endemic damselflies, such as Megalagrion leptodemas, Megalagrion oceanicum, and Megalagrion xanthomelas. Chemical treatment to eliminate introduced fish has already been selectively used mainly for tilapia (and poeciliid) elimination by the U.S. Fish and Wildlife Service in Hawai'i, in areas such as Hanalei National Wildlife Refuge on Kaua'i (A. 26 Asquith, personal communication). Once poeciliids are removed from a reach of stream, it would also be feasible to construct fish barriers that could prevent poeciliids from recolonizing upstream areas, as most O'ahu streams are small in volume and size. Since native stream gobiids and crustaceans are accomplished waterfall climbers (Englund and Filbert, 1997), small fish barriers in restored streams would not hinder recolonization after chemical treatment by native species. Additionally, a high priority should be placed by management agencies on educating the public to the harmful effects ofreleasing aquarium fish into Hawaiian waters. This study has shown a negative correlation between the distribution of introduced fish and native damselfly species. While this does not prove cause and effect, it is highly suggestive that introduced fish are one of the key mechanisms in the reduction in Megalagrion damselfly populations. Further behavioral and experimental studies can build upon the findings of this study, perhaps ultimately allowing the restoration of endemic Megalagrion populations on O'ahu. 27 Appendix 1. Summary of O'ahu streams surveyed, elevations (above sea level) assessed, stream type according to Polhemus and Asquith (1996), and damselfly species present. Stream Elevations Stream Type Damselfly species present (elevation above sea level) Poeciliid species present (elevation above surveyed sea level) Kane'ohe Bay tributaries Kawa 0-74 m Permanent 1. ramburii (0-74 m), 1. posita (0-74 m) P. mexicana, P. reticulata, G. affinis, X. helled (all 0-30 m) Kane'ohe 0-80 m Permanent 1. rambudi (0-30 m) P. mexicana, P. reticulata, G. affinis, X. helled (all 0-61 m) He'eia 0-134 m Permanent 1. posita (134 m), M. n. nigrolineatum (134 m) P. mexicana (0-23 m), G. affinis (0-15 m), x. helleri (0-61 m) Kahalu'u 0-152 m Interrupted by 1. rambudi (0-12 m), M. n. nigrolineatum (85-152 m), P. mexicana (0-37 m), P. reticulata (37-80 channelization M. hawaiiense (137-152 m), M. oceanicum (137-152 m), x. helled (25-55 m) m) Waihe'e 0-243 m Permanent 1. posita (31-210 m), M. n. nigrolineatum (52-243 m), G. affinis (0-12 m), P. mexicana (0-12 m), M. oceanicum (243 m) P. reticulata (0-61 m), X. helleri (0-61 m) Ka'alaea 0-91 m Permanent 1. ramburii (0-37 m), 1. posita (28-37 m), M. n. P. mexicana (0-37 m), P. reticulata (0-37 nigrolineatum (67-91 m) m), x. helled (0-37 m) N 00 Waiahole 0-244 m Permanent 1. ramburii (0-25 m), 1. posita (0-152 m), M. P. mexicana (0-25 m), P. reticulata (0-97 hawaiiense (244 m), M. n. nigrolineatum (244 m) m), x. helled (0-97 m), G. affinis (0-10 m) Waikane 0-244 m Permanent M. n. nigrolineatum (150-244 m), M hawaiiense (200 P. mexicana (0-15 m), P. reticulata (0-67 244 m) m), X. helled (0-67 m), G. affinis (0-15 m) Hakipuu 0-91 m Permanent 1. ramburii (0 m) P. mexicana (0-4 m), P. reticulata (0-31 m), X. helleri (0-31 m) Windward Tributaries Kaelepulu 0-2 m Permanent 1. ramburii (0-2 m) P. mexicana (0-2 m), G. affinis (0-2 m) Maunawili 0-50 m, Permanent 1. rambudi (0-50 m) P. mexicana (0-50 m), G. affinis (0-50 m), 220-240 m x. helleri (0-50 m), P. reticulata (0-50 m) Makaua 335-520 m Intermittent M. oceanicum (335 m ) None Kahana 0-300 m Permanent 1. posita (0-30 m), 1. ramburii (0-20 m), M n. P. mexicana (0-20 m), P. reticulata (0-30 nigrolineatum (30-300m), M. hawaiiense (240-300 m), m), x. helleri (0-30 m) M. leptodemas (60-70 m) Appendix 1 (cont.). Summary of O'ahu streams surveyed, elevations (above sea level) assessed, stream type according to Polhemus and Asquith (1996), and damselfly species present. Stream Elevations Stream Type Damselfly species present (elevation above sea level) Poeciliid species present (elevation above surveyed sea level) Punaluu 0-1 m Permanent none P. mexicana (0-1 m), P. reticulata (0-1 m), X. helleri (0-1 m) Kaluanui 0-107 m, Intermittent M. n. nigrolineatum (671-762 m), M. oceanicum (671- P. reticulata (0-107 m), X. helleri (0-107 671-762 m 762 m) m) Ma'akua 0-245 m Intermittent M. oceanicum (220-305 m), M. leptodemas (220-305 P. mexicana (0-2 m) m) Kaipapa'u 0-305 m Intermittent M. hawaiiense (220 m), M. oceanicum (220-305) None Koloa 0-366 m Intermittent M. hawaiiense (244-274 m), M. n. nigrolineatum (274- P. mexicana (0-6 m), P. reticulata (0-6 m), 366 m), M. oceanicum (244-366 m) X. helleri (0-6 m) Wailele 0-396 m Intermittent 1. ramburii (0-3 m), M. oceanicum (388-396 m) P. mexicana (0-6 m), P. reticulata (0-6 m), X. helleri (0-6 m) Kahawainui 0-590 m Intermittent M. oceanicum (396-488 m) P. reticulata (0-122 m), X. helleri (0-122 m) tv \0 Malaekahana 0-244 m Intermittent 1. posita (61-244 m) none Leeward Waimea 0-76 m Permanent none P. reticulata (0-76 m), X. helleri (0-76 m) Anahulu 180-317 m Permanent none P. reticulata (180-317 m), X. helleri (180 317 m) 'Opae'ula 260 m Permanent none P. reticulata (260 m) Helemano 0-260 m, Permanent 1. posita (480-520 m), M. hawaiiense (480-520 m), M. P. reticulata (0-260 m), X. helleri (0-260 480-520 m n. nigrolineatum (480-520 m) m) Poamoho 480-570 m Permanent M. n. nigrolineatum (480-570 m)l P. reticulata (480-570 m) Kaukonahua 430-490 m Permanent M. n. nigrolineatum (430-490 m) none Dillingham 30-122 m Intermittent M. hawaiiense (30-122 m) none Makua 0-3 m Intermittent none none (dry stream) Hale'au'au 1130-1160 Intermittent M. hawaiiense (1130-1160 m) none m Makaha 0-2 m, Intermittent M. hawaiiense (1130-1160 m) G. affinis (0-2 m), P. mexicana (0-2 m) 1130-1160 m Appendix 1 (cont.). Summary ofO'ahu streams surveyed, elevations (above sea level) assessed, stream type according to Polhemus and Asquith (1996), and damselfly species present. Stream Elevations Stream Type Damselfly species present (elevation above sea level) Poeciliid species present (elevation above surveyed sea level) Kaupuni 0-2 m, Intermittent I. ramburii (0-2 m), M. hawaiiense (400-430 m) G. affinis (0-2 m), P. mexicana (0-2 m) 365-487 m Mailiili 0-1 m Intermittent none none (dry stream) Ulehawa 0-2 m Intermittent none G. affinis (0-2 m), P. mexicana (0-2 m) Nanakuli 0-2 m Intermittent none G. affinis (0-2 m), P. mexicana (0-2 m) Pearl Harbor and south shore tributaries Honouliuli 0-3 m Intermittent Enallagma civile (0-3 m) G. affinis (0-3 m), P. mexicana (0-3 m) Waikele 0-487 m Intermittent Enallagma civile (0-10 m), I. posita (0-10), I. ramburii G. affinis (0-40 m), P. mexicana (0-381 mY, (0-381 m), P. reticulata (1-381 m), X. helleri (1-381 m) Kapakahi 0-2 m Permanent I. posita (0-2 m), I. ramburii (0-2 m) G. affinis (0-2 m), Limia cf. vittata (0-2 m) P. mexicana (0-2 mY, P. reticulata (0-2 m), w X. maculatus (0-2 m) 0 Waiawa 0-2 m, 189- Intermittent I. posita (0-2 m), I. ramburii (0-2 m), M. n. G. affinis (0-2 m), P. mexicana (0-2 m), P. 336 m nigrolineatum (213-365 m), M. leptodemas (213-365 reticulata (0-213 m) m) Waimalu 0-3 m Intermittent I. ramburii (0-3 m) G. affinis (0-3 m), P. mexicana (0-3 mY, P. reticulata (1-122 m), x. helleri (1-122 m) Kalauao 0-3 m Intermittent I. posita (0-3 m), I. ramburii (0-3 m) G. affinis (0-3 m), P. mexicana (0-3 m) Aiea 0-11 m Intermittent I. ramburii (0 m) P. mexicana (0-11 m) Halawa 0-3 m, 240- Intermittent I. ramburii (0-3 m), M. hawaiiensis (240-305 m), M. n. G. affinis (0-3 m), P. mexicana (0-3 m), P. 305 m nigrolineatum (240-305 m), M. leptodemas (240-305 reticulata (1-183 m) m) Moanalua 0-365 m Intermittent I. posita ( 0-79 m), I. ramburii (0-79 m), M. G. affinis (0-122 m), P. mexicana (0-122 xanthomelas (79 m-in an unnamed tributary)2 m), P. reticulata (0-122 m), X. helleri (0- 122 m) Kalihi 0-2 m, 305- Interrupted I. ramburii (0-2 m), M. n. nigrolineatum (305-366 m) G. affinis (0-2 m), P. mexicana (0-2 m) 366 m Kapalama 0-1 m Intermittent none P. mexicana (0-1 m) Appendix 1 (cont.). Summary ofO'ahu streams surveyed, elevations (above sea level) assessed, stream type according to Polhemus and Asquith (1996), and damselfly species present. Stream Elevations Stream Type Damselfly species present (elevation above sea level) Poeciliid species present (elevation above surveyed sea level) Nuuanu 0-3 m, 230- Permanent 1. ramburii (0-240 m) P. mexicana (0-3 m), P. reticulata (0-240 378 m m) Makiki 0-300 m Interrupted none G. affinis (O-lO m), P. mexicana (O-lO m), P. reticulata (0-300 m) Manoa/palolo 0-360 m Permanent 1. ramburii (0-100 m) G. affinis (0-70 m), P. mexicana (O-lOO m), P. reticulata (0-360 m), X. helleri (0-300 m) Wai'alaenui 0-2 m Intermittent none P. mexicana (0-2 m) Wailupe 0-2 m Intermittent none none (dry stream) Pia 0-2 m, 61- Intermittent none none (dry stream) 400 m Kuliouou 0-1 m Intermittent none P. mexicana (0-1 m) Kuapa Pond 0-2 m Intermittent 1. ramburii (0-2 m) G. affinis (0-2 m), P. mexicana (0-2 m), tributaries Limia cf. vittata (0-2 m) Awawamalu 0-2 m Intermittent 1. ramburii (0-2 m) G. affinis (0-2 m), P. mexicana (0-2 m) ~ 1 M. n. nigrolineatumcaptured in small side tributaries lacking P. reticulata entering Poamoho Stream. 2See text for explanation of M. xanthomelas distribution. Appendix 2. Collection sites of O'ahu Megalagrion species by R.c.L. Perkins (1892-1912) and F.x. Williams (1925-1936). Colllector Collection Site Elevation Dates Species R.C.L. Perkins? O'ahu sea level none given M. hawaiiense R.C.L. Perkins Wai'anae Mountains none given February 1896 M. hawaiiense R.C.L. Perkins Honolulu Mountains 550 m March 1901 M. hawaiiense R.C.L. Perkins Honolulu 366 m 1910 M. hawaiiense R.C.L. Perkins Palolo Valley none given May 1912 M. hawaiiense F.X. Williams Herring Valley (Makiki) none given February 1930 M. hawaiiense R.C.L. Perkins Helemano Stream 610 m February 1893 M. leptodemas R.C.L. Perkins Kawailoa Gulch 550 m April 1901 M. leptodemas R.C.L. Perkins Kawailoa Gulch none given August 1901 M. leptodemas R.C.L. Perkins Kawailoa Gulch 550 m August 1901 M. leptodemas R.C.L. Perkins Palolo Valley none given 1912 M. leptodemas F.X. Williams Mt. Ka'ala, Hale'au'au Stream 610 m April 1931 M. leptodemas F.X. Williams Herring Valley (Makiki) none given June 1931 M. leptodemas F.X. Williams Herring Valley (Makiki) none given November 1933 M. leptodemas R.C.L. Perkins Wai'anae Mountains 610 m April 1892 M. n. nigrolineatum R.C.L. Perkins Honolulu none given November 1892 M. n. nigrolineatum w tv R.C.L. Perkins Wai'anae Mountains 610 m February 1896 M. n. nigrolineatum R.C.L. Perkins Honolulu none given 1901 M. n. nigrolineatum R.C.L. Perkins Waialua none given March 1901 M. n. nigrolineatum RC.L. Perkins Waialua none given April 1901 M. n. nigrolineatum RC.L. Perkins Palolo Valley none given May 1902 M. n. nigrolineatum R.C.L. Perkins Maunawili 244m 1912 M. n. nigrolineatum F.X. Williams Ha1e'au'au Stream (Wai'anae Mountains) 610 m January 1932 M. n. nigrolineatum R.C.L. Perkins? O'ahu sea level no date M. oceanicum R.C.L. Perkins Wai'anae Mountains 610 m April 1892 M. oceanicum R.C.L. Perkins Honolulu none given May 1892 M oceanicum R.C.L. Perkins Wai'anae Mountains 610 m February 1896 M. oceanicum R.C.L. Perkins Kawailoa Stream none given March 1901 M. oceanicum R.C.L. Perkins Kawailoa Stream none given April 1901 M.oceanicum R.C.L. Perkins Kawailoa Stream none given July 1901 M.oceanicum R.C.L. Perkins Kawailoa Stream none given August 1901 M. oceanicum RC.L. Perkins Wai'anae Mountains 610 m July 1901 M. oceanicum RC.L. Perkins Honolulu Mountains 550 m 1903 M.oceanicum R.C.L. Perkins Palolo Valley none given May 1912 M. oceanicum Appendix 2 (cont.). Collection sites of O'ahu Megalagrion species by R.C.L. Perkins (1892-1912) and F.X. Williams (1925-1936). Colliector Collection Site Elevation Dates Species F.X. Williams Hale'au'au Stream (Wai'anae Mountains) none given February 1930 M. oceanicum F.X. Williams Hale'au'au Stream (Wai'anae Mountains) none given May 1935 M. oceanicum R.C.L. Perkins Honolulu none given February 1892 M. pacificum R.C.L. Perkins Kawailoa Stream 457 m September 1900 M. pacificum R.C.L. Perkins Kawailoa Stream 457 m April 1901 M. pacificum R.C.L. Perkins Waialua, Ko'olau Range near sea level March 1892 M. xanthomelas R.C.L. Perkins Wai'anae Mountains 610 m April 1892 M. xanthomelas R.C.L. Perkins Honolulu none given September 1892 M. xanthomelas R.C.L. Perkins Honolulu none given November 1900 M. xanthomelas F.X. Williams Waipahu none given April 1925 M. xanthomelas F.X. Williams Wai'anae none given March 1935 M. xanthomelas F.X. Williams Wai'anae - lowland reservoir! none given July 1936 M. xanthomelas lJ.) lJ.) IFrom Williams (1937b) LITERATURE CITED Arthington, A. H. and Lloyd, L. N. 1989. Introduced Poeciliids in Australia and New Zealand. In Ecology and evolution oflivebearing fishes (Poeciliidae), (G.K. Meffe and F.F. Snelson, Jr., eds), pp. 333-348. Englewood Cliffs, New Jersey: Prentice Hall. Brock, V. E. 1960. The introduction of aquatic animals into Hawaiian waters. Intemationale Revue der gesamten und Hydrographie 45: 463-480. Courtenay, W. R. and Meffe, G. K. 1989. Small fishes in strange places: a review of introduced poeciliids. In Ecology and evolution oflivebearing fishes (Poeciliidae), (G.K. Meffe and F.F. Snelson, Jr., eds), pp. 319 331. Englewood Cliffs, New Jersey: Prentice Hall. Davis, 1. A., S. W. Rolls, and Balla, S. A. 1987. The role of Odonata and aquatic Coleoptera as indicators of environmental quality in wetlands. In: The role ofinvertebrates in conservation and biological survey (J.D. Majer ed), pp. 31-42. Report of Western Australia Department of Conservation and Land Management. Perth, Western Australia. Devick, W. S. 1991. Patterns of introductions of aquatic organisms to Hawaiian freshwater habitats. In: New directions in research, management, and conservation ofHawaiian freshwater stream ecosystems, Proceedings ofthe 1990 Symposium on Freshwater Stream Biology and Management, State ofHawaii, pp. 189-213. Hawaii Division of Aquatic Resources. Honolulu: Hawaii. Englund, R. A. and Filbert, R. B. 1997. Discovery of the native stream goby, Lentipes concolor, above Hawaii's highest waterfall, Hiilawe Falls. Bishop Museum Occasional Papers 49: 62-64. 34 Englund, R. A. 1998. Response of the Orangeblack Hawaiian Damselfly (Megalagrion xanthomelas), a Candidate Threatened Species, to increases in stream flow. Bishop Museum Occasional Papers 56: 19-24. Englund, R. A. 1999. New Records and Range Extensions of Native Odonata (Coenagrionidae) and Introduced Aquatic Species in the Hawaiian Islands. Bishop Museum Occasional Papers 59: 15-19. Farley, D. G. and L. C. Younce. 1977. Stocking density versus efficacy of Gambusia affinis in Fresno County rice fields. Proceedings and Papers ofthe Annual Conference ofthe California Mosquito Vector Control Association 45: 83-86. Funasaki, G. Y., Lai P. Y., Nakahara, L. M., Beardsley, 1. W., Ota, A. K. 1988. A review of biological control introductions in Hawaii: 1890 to 1985. Proceedings ofthe Hawaiian Entomological Society 28: 105-160. Harwood, P. D. 1974. Dragonflies (Odonata) observed in Hawaii in 1973 and 1974. Proceedings ofthe Hawaiian Entomological Society 22: 251-254. Hawaii Stream Assessment 1990. A preliminary appraisal of Hawaii's stream resources, Report R-84. Hawaii Cooperative Park Service Unit, Western Region Natural Resources and Research Division. Honolulu, Hawaii. 294 pages. Henrikson, B.I. 1988. The absence of antipredator behaviour in the larvae of Leucorrhinia dubia (Odonata) and the consequences for their distribution. Oikos 51: 179-183. 35 Liebherr, J. and Polhemus, D.A. 1997. Comparisons to the century before: the legacy of the R.CL Perkins and Fauna Hawaiiensis as the basis for a long-term ecological monitoring program. Pacific Science 51: 490- 504. Maciolek, J.A. 1984. Exotic fishes in Hawaii and other islands of Oceania. In Distribution, Biology, and Management ofExotic Fishes, (W.R. Courtenay, Jr. and J.R. Stauffer, Jr., eds), pp. 131-161. Baltimore and London: John Hopkins University Press. McPeek, M.A. 1989. Differential dispersal tendencies among Enallagma damselflies (Odonata) inhabiting different habitats. Oikos 56: 187-195. McPeek, M.A. 1990a. Determination of species composition in the Enallagma damselfly assemblages of permanent lakes. Ecology 71: 83-98. McPeek, M.A. 1990b. Behavioral differences between Enallagma species (Odonata) influencing differential vulnerability to predators. Ecology 71: 1714-1726. Moore, N. W. and Gagne, W. C. 1982. Megalagrion pacificum (McLachlan) - a preliminary study of the conservation requirements of an endangered species. Reports ofthe Odonata Specialist Group, Intemational Union for Conservation ofNature and Natural Resources 3: 1-5. Nelson, S. G. and Eldredge, L. G. 1991. Distribution and status of introduced cichlid fishes ofthe genera Oreochromis and Tilapia in the islands of the South Pacific and Micronesia. Asian Fisheries Science 4: 11- 22. 36 Nichols, W. D., Shade, P. J. and Hunt, C. D. 1997. Summary of the Oahu, Hawaii, regional aquifer-system analysis. U.S. Geological Survey Professional Paper 1412-A. 61 pp. Nilsson, B. 1. 1981. Susceptibility of some Odonata larvae to fish predation. Verb. Internat. Verein. Limnol. 21: 1612-1615. Nishida, G. M. 1997. Hawaiian Terrestrial Arthropod Checklist, 3rd Edition. Hawaii Biological Survey. Bishop Museum Technical Report 12. 263 pp. Perkins, R. C. L. 1899. Odonata. In Fauna Hawaiiensis, Vol. 2, Part 3 (D. Sharp ed.), pp.63-77. Cambridge: Cambridge University Press. Perkins, R. C. L. 1910. Supplement to Neuroptera. In Fauna Hawaiiensis, Vol. 2, Part 6 (D. Sharp ed.), pp. 693 696. Cambridge: Cambridge University Press. Pierce, C. L. 1988. Predator avoidance, microhabitat shift, and risk-sensitive foraging in larval dragonflies. Oecologia 77: 81-90. Polhemus, D.A. 1993. Damsels in distress: a review of the conservation status of Hawaiian Megalagrion damselflies (Odonata: Coenagrionidae). Aquatic Conservation 3: 343-349. Polhemus, D. A. 1996. The Orangeblack Hawaiian Damselfly, Megalagrion xanthomelas (Odonata: Coenagrionidae): clarifying the current range of a threatened species. Bishop Museum Occasional Papers 45: 30-53. 37 Polhemus, D. A. 1997. Phylogenetic analysis of the Hawaiian damselfly genus Megalagrion (Odonata: Coenagrionidae): implications for biogeography, ecology, and conservation biology. Pacific Science 51: 395-412. Polhemus, D. A. and Asquith, A. 1996. Hawaiian damselflies: a field identification guide. Bishop Museum Press, Honolulu, Hawaii. 122 pp. Polhemus, D. A., Maciolek, J., and Ford, 1. 1992. An ecosystem classification of inland waters for the tropical Pacific Islands. Micronesica 25: 155-173. Steams, H.T., and Vaksvik, K.N. 1935. Geology and ground-water resources of the island of Oahu, Hawaii. Territory of Hawaii, Department of Public Lands, Division of Hydrography, Bulletin 1,479 p. Van Dine, D. L. 1907. The introduction of top-minnows (natural enemies of mosquitoes) into the Hawaiian Islands. Hawaiian Agricultural Experiment Station Press Bulletin 20: 1-10. Van Dine, D. L. 1908. Mosquitoes. Annual Report of the Hawaii Agricultural Experiment Station for 1907, pp. 38 39. Way, C. M. and Burky, A. 1. 1991. A preliminary survey of the macroinvertebrates and a preliminary assessment of the diet of the endemic Hawaiian goby Lentipes conco1or (Gill). In: New directions in research, management, and conservation ofHawaiian freshwater stream ecosystems, Proceedings ofthe 1990 Symposium on Freshwater Stream Biology and Management, State ofHawaii, pp. 158-164. Hawaii Division of Aquatic Resources. Honolulu: Hawaii. 38 Williams, F. X. 1937a. Biological Studies in Hawaiian water-loving insects. Part 1. Coleoptera or beetles. Part 2. Odonata or dragonflies. Proceedings ofthe Hawaiian Entomological Society 9: 235-345. Williams, F. X. 1937b. Exhibitions and discussions of local material. Proceedings ofthe Hawaiian Entomological Society 9: 368. Zimmerman, E. C. 1948. Insects ofHawaii. Volume 2. Apterygota to Thysanoptera inclusive. Honolulu: University of Hawaii Press. 39 CHAPTER 2: EVALUATING THE EFFECTS OF INTRODUCED RAINBOW TROUT (Oncorhynchus mykiss) ON NATIVE STREAM INSECTS ON KAUA'I ISLAND, HAWAI'I 40 ABSTRACT Rainbow trout (Oncorhynchus mykiss) and other salmonids have been widely stocked into upland streams throughout the world to provide a basis for sport fisheries, but the effects of such introductions on indigenous and endemic aquatic insect assemblages are poorly documented. In this study, we examine the impact of rainbow trout on the indigenous and endemic entomofauna of upland streams in K6ke'e State Park, Kaua'i, Hawai'i, with particular emphasis on the potential threat trout pose to populations of endemic damselflies in the genus Megalagrion. Rainbow trout were introduced into the upland streams of Kaua'i beginning in the 1920s, with over 60 years of subsequent restocking. This study indicates, however, that streams in this area still maintain diverse populations of Megalagrion damselflies and other indigenous and endemic aquatic insects, both in catchments containing naturally reproducing trout populations and in catchments lacking rainbow trout. Our results indicate that the indigenous and endemic aquatic insect communities in the streams under study compare favorably in terms of density and taxonomic richness with other isolated and unimpacted streams elsewhere in Hawai'i, containing high densities and relative percentages of indigenous and endemic aquatic insect taxa. Our results demonstrate that the threats posed by conspicuous introduced species such as trout should not simply be assumed apriori on the basis of postulated negative interactions, because this may divert limited resources from programs aimed at control of other, potentially more destructive introduced taxa such as inconspicuous poeciliid fishes. INTRODUCTION Conservation biologists working with the faunas of isolated oceanic islands, such as Hawai'i, have long recognized that introduced species represent perhaps the most pervasive and persistent threat to the survival of insular biotas. At the same time, the effects of such introduced species are often simply presumed to occur, without rigorous scientific evaluation as to the actual impacts; this paper provides an example. This could lead to unnecessary expenditures of time and money in attempts to control introduced species that, though conspicuous, may be having relatively low impacts on endemic taxa, diverting resources that could be better utilized to deal with other, more pressing threats. In this paper we examine such a case, that of 41 introduced rainbow trout (Oncorhynchus mykiss) on the island of Kaua'i in the Hawaiian Islands, and their impact on Megalagrion damselflies and other endemic aquatic insects. Salmonids were first stocked on Kaua'i in 1894, with an unsuccessful attempt to establish brook trout (Salvelinus fontinalis) into the Waimea River (Needham and Welsh 1953). Rainbow trout (Oncorhynchus mykiss) were subsequently introduced starting in 1920, and in 1935, brown trout (Salmo trutta) were stocked in the Wainiha River and certain other Kaua'i streams, although the latter never became successfully established (Needham and Welsh 1953). Stocking ofrainbow trout into streams in the upland Koke'e State Park was perfonned on an annual basis after World War II (Needham and Welsh 1953), but was discontinued in 1992 because of concerns regarding the impacts ofrainbow trout on populations of endemic Megalagrion damselflies, certain species of which had by that time been proposed for listing under the U. S. Endangered Species Act. The introduction of trout to Kaua'i led in tum to the deliberate introduction of several alien mayfly species to Koke'e State Park streams in an attempt to supplement the naturally limited benthic invertebrate communities of these systems. Two mayfly species were introduced from Waddell Creek in the Santa Cruz Mountains of California to K6ke'e in 1961 (Needham and Usinger 1962; Usinger 1972): Nixe rosea (at that time called Heptagenia rubroventris) and Epeorus lagunitas (fonnerly Iron lagunitas). Although over 200,000 mayfly eggs were planted into various habitats in Koke'e and Kawaikoi Streams, this attempted introduction of herbivorous mayflies failed completely, with no adults of either species ever recovered from any Kaua'i stream, probably because these temperate zone species require an obligate cold diapause in order to complete their life cycles, a stimulus that was not provided under Hawaiian conditions (G. F. Edmunds, Jr., pers. comm.). In Hawai'i as a whole, the only mayfly species known to be successfully established is Caenis nigropunctata, a tropical mayfly indigenous to Indonesia, the Philippines, and southern China (Dudgeon 1999), which is found in lowland, disturbed and silted streams on O'ahu (Smith 2000), and does not require a seasonal diapause. In addition, by 1965 several species of alien caddisflies, including the large 42 Cheumatopsyche pettiti, had become accidentally established in Hawaiian streams (Denning and Beardsley 1967), providing additional food sources for introduced gamefish, including rainbow trout. The mode of introduction for these latter species into Hawai'i remains unknown, but probably involved the arrival of eggs in aquatic plants or plant substrates imported in association with the aquarium trade (l.W. Beardsley, pers.comm.). Whatever the means of establishment, by the early 1970s introduced Trichoptera had become abundant in the benthos of most Hawaiian streams, thereby inadvertently accomplishing what the earlier unsuccessful mayfly introductions had set out to do. In their natural condition, Hawaiian streams have poorly developed benthic insect communities in comparison to similar continental ecosystems. The largest endemic benthic insects are larvae of the secondarily freshwater chironomid genus Telmatogeton, and various Odonata species in the genera Anax, Nesogonia, and Megalagrion. Although recent studies have indicated Megalagrion damselflies with stream dwelling immatures are sensitive to the presence of introduced poeciliid fishes such as the Western mosquitofish (Gambusia affinis), green swordtails (Xiphophorus hellen) and shortfin mollies (Poecilia mexicana) (Polhemus and Asquith, 1996; Englund 1999), the impact of rainbow trout predation on indigenous and endemic damselfly populations has remained equivocal prior to the current study. In recent papers, different authors have alternately claimed that rainbow trout significantly impact endemic damselfly populations on Kaua'i (Kido et al. 1999), or conversely that trout in Hawai'i are an example of a failed invasion and therefore have little impact (Brown et al. 1999). The prevailing assumption among local conservationists is that because introduced trout are known to adversely impact indigenous and endemic species in other regions, they must also be having a significant impact on endemic aquatic species in Hawai'i. One of the primary aims of this study was therefore to objectively assess the relative threat posed by trout to Megalagrion damselfly populations and other endemic Hawaiian aquatic invertebrates in the context of the overalllimnological setting occupied by these organisms. 43 METHODS STUDY AREA Koke'e State Park is located in northwestern Kaua'i, the northernmost of the major high Hawaiian islands, and encompasses most of the Alaka'i Swamp. The park contains numerous streams draining from the flanks of Mt. Waialeale, lying near the highest summit on Kaua'i and putatively the wettest area on earth (Hazlett and Hyndman 1996). The Alaka'i Plateau surrounding Mt. Waialeale is a swampy tableland, heavily dissected by deep, erosion-formed canyons that cut into its relatively flat surface, and thickly covered by predominantly native Hawaiian upland rain forest. The five streams examined in this study, Wai'alae, Koai'e, Kauaikinanii, Kawaikoi, and Waiakoali, are situated relatively close to each other in the western section of the Alaka'i Plateau, lying at elevations ranging from 1035-1160 m (Figure 2.1). Because all five streams lie at relatively similar elevations, and exhibited low and stable streamflows throughout both phases of this study, it was possible to conduct uniform aquatic insect sampling protocols at each, resulting in comparable data for all watersheds. The stations sampled were as follows: Station 1: HAWAI'I, Kaua'i, Wai'alae Stream in vicinity of Wai'alae Cabin, 1095 m, water temp. 15°C, January 1999. 22°04.964'N, 159°35.170'W. Station 2: HAWAI'I, Kaua'i, Koai'e Stream in vicinity of USGS gaging station, near Mohihi-Wai'alae Trail crossing, 1125 m, water temp. 15°C, August 1997, January 1999. 22°06.778'N, 159°33.723'W. Station 3: HAWAI'I, Kaua'i, Waiakoali Stream above Camp 10 Road, 1035 m, water temp. 13°C, August 1997, August 1998, January 1999. 22°07.586'N, 159°37.225'W. Station 4: HAWAI'I, Kaua'i, Kawaikoi Stream above Camp 10 Road, 1035 m, water temp. 15°C, August 1997, August 1998, January 1999. 22°07.892'N, 159°37.280'W. 44 Station 5: HAWAI'I, Kaua'i, Kauaikinana Stream at Camp 10 Road, 1035 m, water temp. 14°C, August 1997, August 1998, January 1999. 22°07.974'N, 159°37.840'W. t N Figure 2.1. Study area of sampled K6ke'e State Park Streams, Kaua'i Island, Hawai'i 45 In addition, aquatic insect richness data from two other nearby trout-free Kaua'i streams, the Lumaha'i and Hanalei rivers, were incorporated into the current analysis based on studies by Polhemus (1995). General Methods This study was conducted from 1997 to 1999, during periods of low flow. Sampling areas were established in three streams containing rainbow trout with natural reproduction (Kauaikinana, Koai'e, Wai'alae), one stream where rainbow trout escaped from a nearby ditch stocked in 1998 (Waiakoali), and along a stream devoid ofrainbow trout (Kawaikoi). The presence or absence of naturally reproducing trout in all of the stream reaches under study was determined by visual observation at very low flows, angling, netting, and discussions with knowledgeable local anglers and wildlife biologists. To assess potentially available rainbow trout forage, the abundance and species richness of aquatic invertebrate populations were surveyed in each stream using a variety of methods. To evaluate actual rainbow trout diet, stomach contents were collected from rainbow trout in the four streams containing large trout populations, as noted above. Aquatic Insect Sampling In order to obtain comprehensive samples of the aquatic entomofaunas present in each study stream, collections of both immature and adult specimens were made using aerial sweep netting, dip nets and Surber (benthic) samplers; in addition, Malaise traps were placed over each stream for a 24-hour period. Visual observations of flying adult insects, particularly Odonata, were also conducted while traversing the stream reaches under study. Total numbers of aquatic insect taxa collected at each stream were determined by combining collection data from all the above methods. Sampling effort was focused on habitat suitable for Hawaiian aquatic insects, particularly splash zones around riffles and cascades, wet rock faces associated with springs and seeps, waterfalls, and rock overhangs. 46 The sampling of damselflies and dragonflies (Odonata) was emphasized, because six species of endemic Megalagrion damselflies are currently held as candidates for listing under the Endangered Species Act on the United States Federal Register. In addition, Hawaiian damselflies provide an indication of the relative 'health' of a stream system, and are typically absent in highly disturbed catchments (Polhemus and Asquith, 1996), or those containing high densities of introduced fishes (Englund 1999). Initial attempts were made to census native adult (i.e., flying) damselflies in order to obtain quantitative density data, but these were abandoned after it was determined that such damselfly species were too abundant to accurately census without the possibility of double-counting the same individuals. Instead, qualitative species richness data were utilized, similar to those gathered for the remaining groups of aquatic insects in this study. Benthic aquatic insect densities were assessed at each stream using randomly sited Surber samples set in riffles. All of the streams studied contained many moderate gradient riffles, with similarly sized substrate (6--25 cm), allowing comparable samples at each. Drifting invertebrates were collected with cone shaped Wildco® drift nets 0.9 m long with an aperture area 2 of 0.09 m and a mesh size of 450 jIm. The center of the net was placed in the middle of the stream and was set at approximately 60% of the water column depth (Filbert and Hawkins 1995). Each drift collection consisted of one lO-min sample collected at midday (1200 hrs) in each stream sampled. All drift collections were made within a one-week time period during the winter 1999 sampling. Time and budgetary constraints did not allow for comprehensive diel drift sampling that would have included nighttime or replicated sampling. Rather, these preliminary samples were collected to assess drift species composition and provide inferences on the propensity of endemic aquatic insects to drift. Such qualitative drift samples were collected in Kauaikinana, Kawaikoi, Waiakoali, and Koai'e streams, but not at Wai'alae Stream. 47 All insect specimens were stored in 75 percent ethanol and subsequently transported to the Smithsonian Institution and Bishop Museum for curation and identification. Voucher specimens are currently housed in the Bishop Museum and Smithsonian Institution collections. Rainbow Trout Diet Analysis Rainbow trout were captured by angling for this study, killed immediately after capture, and the stomachs injected with and placed in 70% ethanol for later laboratory analysis. Gut contents were removed from the portion of the gut between the anterior of the esophagus and pyloric sphincter (Kimball and Helm 1971). Contents of each rainbow trout stomach were identified to the lowest possible taxonomic level, and in many cases to genus and species. Species level identification was not always possible for partially digested insects, or in some orders such as Diptera where adult males are required for identification. The geographic origin (i.e., indigenous and endemic versus introduced) of individual arthropod taxa was evaluated using Nishida (1997). Total prey item numbers in each trout stomach were recorded, and if contents were broken into pieces, head capsules were then counted as an indication of individual prey item numbers. Terrestrial or aquatic status was determined for each identifiable prey item, and the percent composition oftrout diet represented by any given prey item class was calculated by adding the total number ofidentifiable prey items and dividing by the number in each prey item category. The proportion ofintroduced and endemic aquatic prey items in the trout stomachs was calculated for the aquatic species where such information regarding geographic origin could be ascertained. The processed stomach contents are stored in the collections of the Hawaii Biological Survey, Bishop Museum. Data Analysis Proportions Tests - Proportion tests were conducted to assess whether the presence of rainbow trout influenced the number ofendemic aquatic insect species in a particular stream (Ramsey and Schaefer 1997). The null hypothesis of no significant difference in the proportion of endemic aquatic insect species between 48 streams with and without trout was tested against the alternative hypothesis that the proportion of endemic species in streams without trout is higher than that in streams with trout. A two-sample test for equality of proportions with continuity correction tests was conducted using the S-Plus®2000 statistical program (MathSoft, Inc., 1999). Streams used in the proportion test included all assessed streams in Koke'e State Park, and two other relatively high elevation Kaua'i Streams lacking rainbow trout; the Lumaha'i River (430 m elevation) and Hanalei River (380 m elevation), using data from Polhemus (1995). Poisson Regression for Count Data -A Poisson regression model for count data was used to test if the main constituent of the benthic fauna collected in Surber samples, Cheumatopsyche pettiti, varied by stream. The null hypothesis of there being no significant difference in the number of C. pettiti per stream was tested against the alternative hypothesis that densities varied among streams. RESULTS Our data indicate that the streams sampled within Koke'e State Park maintain diverse populations of endemic aquatic insects. This is true both of streams containing rainbow trout, such as the upper reaches of Wai'alae, Koai'e, and Kauaikinana, and of streams where trout are lacking, such as the upper reaches of Kawaikoi, the latter being comparable to other trout-free streams such as the upper Hanalei and Lumaha'i rivers (Table 2.1). A large number ofaquatic and terrestrial taxa were identified during this study; these are summarized in Table 2.2. Terrestrial species were only included when found in trout stomachs or collected during invertebrate drift sampling (Table 2.2); even so, they comprise a large proportion (47.4%) of all the taxa consumed by rainbow trout. The aquatic entomofauna was composed largely ofendemic species, as shown in Figure 2.2. Except for Wai'alae Stream where drift samples were not collected, terrestrial species were included in Figure 2.2 in order to show the proportions ofendemic versus introduced species present in the stream drift. A summary of the aquatic and terrestrial taxa collected in rainbow trout stomachs is shown in Figure 2.3, and clearly indicates the importance of drifting terrestrial taxa to the diet of rainbow trout on 49 Table 2.1. Aquatic insect species and native or introduced status collected at each Kaua'i stream. I~ l': ~ '10 (l) .~ -; ·z l': ..>d 0 1il -; :;;l .(;; :"; ~ ..>d -; l': .(;; ~ ~ :"ed Aquatic Insects Species (I =Indigenous; E =Endemic; ~ 0 8 .(;; ::l ~ ~ ::l ::I: ~ ~ .....:I ~ ~ Int =Introduced) ~ Odonata Aeshnidae Anax strenuus (E) XXXXX X X Libellulidae Nesogonia blackbumi (E) X X X PantaIa flavescens (I) X Coenagrionidae Megalagrion eudytum (E) X Megalagrion heterogamias (E) XXXX XX X Megalagrion oresitrophum (E) XX XXXXX Megalagrion orobates (E)l X X Megalagrion vagabundum (E) XX XXXXX Heteroptera Saldidae SaIdula exulans (E) XXXX XX X SaIduia oahuensis (E) X Saldula procellaris (E) XX XXX X Veliidae Microvelia vagans (E) XXXX XX X Coleoptera Dytiscidae Rhantus pacifus (E) XX X Hydrophilidae Limnoxenus semicylindricus (E) X Lepidoptera Cosmopterigidae Hyposmocoma sp. (E) XX X Hyposmocoma sp. Dr montivolans XXXX XX X (E) Hyposmocoma sp. Dr saccophora (E) XXXX XX X Diptera Canacidae Procanace sp. (E) X Procanace bifurcata (E) X Procanace nigroviridis (E) X XX X Procanace quadrisetosa (E) X X X Procanace wirthi (E) X XX Ceratopogonidae Forcipomyia hardyi (E) XXXX XX Chironomidae Chironomus sp. (E) X XX Orthocladius sp. (E) XX Micropsectra sp. (E) X Telmatogeton sp. (E) X Telmatogeton hirtus (E) X Dolichopodidae Campsicnemus nigricollis (E) X XXXXX Campsicnemus n. sp. (E) XX Eurynogaster mediocris (E) XX Eurynogaster minor (E) X X Paraliancalus metallicus (E) X Sigmatineurum napali (E) X XX X Sigmatineurum n. sp. 1 (E) X Sigmatineurum n. sp. 2 (E) X Ephydridae Scatella cilipes (E) X X XX Scatella clavipes (E) X XX Scatella hawaiiensis (E) X XX X X Scatella kauaiensis (E) X XXXXXX Tipulidae Dicranomyia hawaiiensis (E) XX X XX Dicranomyia pcoba(E) XXXXXXX 50 Table 2.1 (cont.). Aquatic insect species and native or introduced status collected at each Kaua'i stream. l«l .~ l:: ° Il.l 'Q) «l 10 "«l <; «l l:: ~ Y 0 <; <; :.::l 'a ~ ~ l:: ~ 'a :; Aquatic Insects Species (I Indigenous; E Endemic; «l 0 8 = = 'a «l ~ ::l 'a :::c: g ~ .....:l ~ ~ Int =Introduced) ~ Dicranomyia perkinsi (E) X Dicranomyia stygipennis (E) XX XXX Odonata Coenagrionidae Ischnura posita (Int) X XXXX Ischnura ramburii (Int) X Enallgma civilie (Int) XX Heteroptera Mesoveliidae Mesovelia amoena (Int) X Mesovelia mulsanti (Int) X Coleoptera Hydrophilidae Tropisternus lateralis (Int) X Trichoptera Hydropsychidae Cheumatopsyche pettiti (lnt) XXX XXXX Hydroptilidae Hydroptila arctia (Int) XX XXX Oxyethira maya (Int) X X X Diptera Chironomidae Cricotopus bicinctus (Int) X XXXXXX Dolichopodidae Chrysotus longipalpus (Int) X X Syntormon flexible (Int) X Dolichopus exsul (Int) XXX X Ephydridae Ochthera circularis (Int) X Tipulidae Dicranomyia advena (Int) X X X 'Not found in upper Koke'e streams because it is only found only in lower to upper mid-reaches of Kaua'i streams (Polhemus and Asquith 1996). Table 2.2. Total number of aquatic species collected during benthic, drift, and aerial (general) collections during this study. Total Aquatic Native Total Native Method Spp.' Aquatic Terrestriae Terrestriae Total Species General!Aerial 48 33 n/a n/a 48 Drift 15 7 26 13 41 Benthic (Surber) 9 3 n/a nla 9 Trout Stomach 43 21 85 36 128 lIncludes fish, crustaceans, mollusks, and aquatic insects. urerrestrial species not sampled during benthic (Surber) or general species collections. 51 oNative Aquatic .Intro Aquatic ~ Native Terrestrial IIlntro Terrestrial VI 30 Q) 'g 2S a. (J) ~o ~o ... 15 ~ E 10 :::l Z 5 Kauaikinana KawaJk~ Koale Walakoali Waialae Figure 2.2 Summary of aquatic species collected in Koke'e State Park streams from all sampling methods combined (general sampling, Malaise traps, drift, benthic samples). oNative Aquatic • Introduced Aquatic 0 Native Terrestrial • Introduced Terrestrial 25 VI 2D Q) '0 Q) ~ 15 '0... Q) 10 .0 E :::l Z 5 Kauaikinana Koaie WaiakoaU Waialae Figure 2.3. Summary of all insect species collected from rainbow trout stomachs (n =80) and their terrestrial or aquatic, and native or introduced status in Koke'e State Park Streams. 52 oCheumatopsyche pettiti • Cricotopus bicinctus 18 Q) 16 a. E ell 14 V)... Q) 12 Q. >. +'" 10 'iii t: 8 oQ) t: 6 ell Q) ~ 4 2 o-IL-~~~----=~~--.-----=~~------':~~~-=~~ Kauaikinana Kawaikoi Koaie Waiakoali Waialae Figure 2.4. Mean density by stream for the two most important constituents of benthic (Surber) samples, the caddisfly C. pettiti and midge C. bicinctus. D Native Aquatic Insect Spp.•Introduced Aquatic Insects Spp. 3S VI Q) 'u Q) Q. 30 Vl +J U Q) 2S VI c:: .~ +J 20 ro :J «0 1S ~o ~ Q) .c 10 E :J Z S 0 Kauaikinana Koaie Waiakoali Waialae Kawaikoi Lumahai Hanalei Figure 2.5. Summary graph of number of aquatic species in Kaua'i Streams and the presence or absence of naturally reproducing trout in each stream; Lumaha'i and Hanalei have never been stocked with trout. 53 Kaua'i. Figure 2.3 also shows the lack of feeding selectivity by rainbow trout in regard to any particular aquatic or terrestrial taxon in their diet. Surber A total of 60 Surber samples were randomly collected from riffles in the five streams surveyed. Benthic sampling was completely ineffective in assessing aquatic insect diversity in upland Kaua'i streams. Only 5 aquatic insect species were collected via benthic sampling: Cheumatopsyche pettiti (Trichoptera: Hydropsychidae), Cricotopus bicinctus (Diptera: Chironomidae), Rhantus pacificus (Coleoptera: Dytiscidae), Megalagrion heterogamias (Odonata: Coenagrionidae) and an unidentified crane fly (Diptera: Tipulidae), the latter accounting for little more than 1% (by number) of all invertebrate taxa collected in the benthic samples. Of these five insect species, only two (Rhantus pacificus and Megalagrion heterogamias) are unequivocally known to be endemic, and comprised a relatively small percentage of the sample, while two demonstrably introduced taxa (Cheumatopsyche pettiti and Cricotopus bicinctus) comprised over 80% of the total collected benthic macrofauna. Mean benthic sample densities for Cheumatopsyche pettiti and Cricotopus bicinctus are shown in Figure 2.4. There is evidence that the extra-Poisson dispersion model fits this data (p-value = 0.5584 > 0.05). Statistical tests of Surber samples using this Poisson regression indicate that the counts of C. pettiti in Kauaikinana Stream, a stream with naturally reproducing trout, are higher than in upper Kawaikoi Stream, which lacks trout (p-value = 0.0004), with the mean number of C. pettiti counts being 3.5 times higher in Kauaikinana than in Kawaikoi (95% CI = [1.75, 6.86]). In summary, the Poisson regression statistical test indicates that the densities of the most important constituent of benthic aquatic insect fauna, C. pettiti, are independent of the presence or absence ofrainbow trout. Rather, densities of this species are hypothesized to instead be dependent on favorable water quality 54 characteristics, with more stable groundwater-fed streams such as Kauaikinana having higher densities of introduced caddisflies. There is also correlational evidence suggesting that introduced Trichoptera in Hawai'i exhibit lower densities in high acidity blackwater streams fed by surface runoff from upland swamps and bogs (Polhemus 1995) than in streams fed by pH-neutral deep groundwater. This is supported by results from Wai'alae Stream, where Surber samples were unintentionally taken upstream of several groundwater springs and tributaries that enter the stream just below Wai'alae Cabin. The stream in this reach is highly tannic, and the benthic samples had anomalously low numbers of C. pettiti; had the samples been taken downstream of the pH-neutral groundwater inputs, we suspect the numbers would have been more similar to those observed in other systems with less acidic waters. General Aquatic Insect Collections In contrast to benthic sampling, general collections, mainly conducted with sweep and dip nets in the splash-zones ofriffles and cascades, were an effective method ofqualitatively assessing aquatic insect species richness in the streams under study. Because the larval stages of most endemic aquatic insects in Hawai'i have not yet been described or adequately associated with the adults on which most classifications are based (Howarth and Polhemus 1991), collection of adult stages was emphasized. By this measure, all streams sampled were found to have large absolute numbers (and high relative percentages) ofendemic aquatic insect taxa. A low of 20 endemic aquatic insect species were collected at Wai'alae Stream, while 33 endemic taxa were found in Koai'e Stream. The endemic species assemblage found at Koai'e represents the greatest number ofendemic aquatic insect species so far recorded for any individual stream yet sampled in the Hawaiian archipelago (for comparative data see Polhemus 1995; Englund and Preston 1999), despite the presence of naturally reproducing rainbow trout in this system. A two-sample test for equality of proportions found the number of endemic aquatic insect species in streams with and without trout (p-value =0.6861) did not significantly differ. Streams without trout included Kawaikoi, Lumaha'i (430 m elevation), and Hanalei (380 m elevation); while streams containing trout 55 included Kauaikinana, Koai'e, Waiakoali, and Wai'alae (Figure 2.5). A summary of the presence of endemic damselfly taxa, and the presence or absence of naturally reproducing or stocked rainbow trout in the stream reaches sampled can be found Table 2.3. This table illustrates the full complement of Megalagrion damselfly species were found in streams both lacking and containing rainbow trout. The reasons for high endemic aquatic insect diversity in Koai'e Stream, despite the presence of trout, may be attributed to several factors. First, the reach of Koai'e Stream sampled exhibited greater habitat diversity than the other streams under study, having waterfalls, large rheocrene seeps (wetted rock faces), and a wide diversity of substrate size classes. Second, Koai'e Stream received significant groundwater spring inputs that may have provided more favorable water chemistry for aquatic insects than the acidic, tannin-rich surface waters typical ofWai'alae and Kawaikoi streams. Drift Samples Drift samples were collected in Kauaikinana, Kawaikoi, Waiakoali, and Koai'e Streams. While it was beyond the scope of this study to provide a comprehensive and quantitative analysis of diel stream drift, these preliminary findings are of interest and will possibly stimulate further research into endemic aquatic insect drift in Hawaiian streams. The terrestrial component of drift was relatively small (15%) when compared to the large numbers of aquatic insects present. Introduced aquatic insects dominated and numerically accounted for 60% of drift for all streams combined. Our drift samples were taken during low basal flows, concurrently with collection of rainbow trout stomach samples in the winter of 1999. It is likely that samples taken during a rain event would contain a higher proportion of terrestrial insects because of flooding of the riparian zones, and the action of rain in dislodging terrestrial insects from trees in the adjacent rain forest. By far the most numerically abundant aquatic insect species captured in the drift samples was the introduced midge Cricotopus bicinctus, which comprised 45% of the drift, although the endemic midge Forcipomyia 56 hardyi, at 14%, also represented an important, though lesser, component (Figure 2.6). Other taxa found in large numbers in the drift samples included endemic midges such as Chironomus sp. at 9%, and introduced caddisflies (Trichoptera) at 5% (Figure 2.6). Larvae of several undescribed semi-aquatic moth species in the genus Hyposmocoma (Cosmopterigidae), which are algal grazers on emergent boulders within the stream channels, were also found in the drift samples. Little is known regarding the biology of the Hawaiian Hyposmocoma species, and although there is anecdotal evidence that larvae enter the water to move between rocks (Merritt and Cummins 1996), it cannot currently be established whether they should be viewed as terrestrial taxa that are occasionally washed into the stream, or typical aquatic constituents of the stream drift. For this study we chose to categorize Hyposmocoma as aquatic because of their clear relationship to stream habitats. Though common as immatures in the drift samples, adult individuals of the endemic midge Forcipomyia hardyi were not collected by other sampling methods employed during this study (e.g., aerial collections, Table 2.3 Presence or absence in surveyed K6ke'e State Park streams of naturally reproducing and stocked rainbow trout and native Megalagrion damselflies. Rainbow Trout Naturally Reproducing or Stocked? Natural No Trout Natural Stocked Natural No Trout No Trout Kauaikinana Kawaikoi Koai'e Waiakoali Wai'alae Hanalei Lumaha'i (380 m) (430 m) M. eudytum +1 M. heterogamias + + + + + + + M. oresitrophum + + + + + + + M. vagabundum + + + + + + + 1Rheocrene habitat for this species only found at this stream. Malaise traps, trout stomachs, benthic sampling). This absence of adults in general collecting samples could be related to seasonal emergence patterns, or to adults leaving the stream corridor following emergence. This anomaly highlights the importance of using multiple sampling techniques at different 57 times of year to obtain comprehensive biodiversity information on endemic aquatic insects in Hawaiian streams. 250 200 50 O~----::==--..-----==:::...... ,.-----==:"""'--=='------r--==-----.--==--t' Cricolopus Forcipomyia Chironomus sp. Chironomidae Trichoplera Hyposmocoma bicinclus hardyi Figure 2.6. The six numerically most abundant aquatic insect taxa captured in drift samples taken in Koke'e state park streams A wide variety of terrestrial taxa were also captured in drift samples, with terrestrial mites (Acari) predominating, intermixed with numerous other groups including beetles (Coleoptera), flies (Diptera), bark lice (Psocoptera), planthoppers and aphids (Homoptera), and others. For those terrestrial drift taxa that could be identified to the species level, most proved to be endemic species. This reflects the nearly pristine character of the riparian vegetation found along many of Kauai's upland streams. Although ginger, blackberry and guava were encountered along the lower sections of streams such as Waiakoali and Kauaikinana, the surrounding forests are nevertheless generally dominated by an overstory of Metrosideros polymorpha (Myrtaceae) and an understory of other indigenous and endemic plants, which support a correspondingly native insect biota. Trout Diet A total of 80 rainbow trout stomachs were collected from the five streams under study. The potential impacts of rainbow trout on endemic aquatic insects could be unambiguously assessed, since the streams 58 examined were above the altitudinal ranges of endemic insectivorous fish. These streams were also above the normal elevational range of the one endemic upland freshwater crustacean species occurring in Hawai'i, Atyoida bisulcata, which is typically found below 900 m elevation. A complete breakdown of the aquatic versus terrestrial, and endemic versus introduced status of all identified prey items found in rainbow trout stomachs is presented in Table 2.4. As is evident in this table, rainbow trout consume a wide variety ofprey items, but endemic aquatic insects, the category of greatest interest to this study, accounted for only 7.5% of the diet. Introduced aquatic insects accounted for 28.9% ofthe diet and terrestrial insects accounted for 47.4% of trout diet, while organisms with an unknown geographic origin represented 16.2% (Table 2.4). Four species of Megalagrion were present along the streams involved in the present study: Megalagrion heterogamias - this was the most commonly encountered damselfly along streams in K6ke'e State Park during the course of this study, and is the largest and most conspicuous stream-breeding damselfly on Kaua'i (Polhemus and Asquith 1996). The immatures are confined to high velocity habitats in main stream channels, such as runs, riffles and cascades, and were one of the few aquatic insects taken by benthic sampling, while the adults were generally observed patrolling beats above these same areas. The remains of 5 M heterogamias immatures were among the 1,596 identifiable prey items recovered from the 80 rainbow trout stomachs analyzed. This result is interesting in light of the fact that rainbow trout are found primarily in slow pool habitats, while immature M. heterogamias inhabit fast riffles and cascades. The observed trout predation on these damselfly immatures is therefore hypothesized to occur when the latter become dislodged from their preferred habitats and are entrained in the water column, becoming a component of the stream drift. MegaJagrion vagabundum - this was the second most common damselfly observed in the study area; adults were commonly encountered around streamside seeps, in mossy riparian areas, and more 59 Table 2.4. Geographic origin and terrestrial or aquatic status of prey items found in 80 Koke'e trout stomachs, 1997-1999. Taxon Native Introduced Unknown Native Introduced Unknown Total Aquatic Aquatic Aquatic Terrestrial Terrestrial Terrestrial Amphipoda 23 23 Anura 2 2 Aranae 16 17 Arhynchobdellida 53 53 Basommatophora I I Blattodea 12 12 Coleoptera 9 2 18 12 16 58 Collembola I I Cypriniformes 3 3 Decapoda 6 3 9 Diptera 27 64 35 6 133 Gordioidea 6 6 Heteroptera 8 22 25 5 6 66 Homoptera 30 6 36 Hymenoptera 4 94 5 103 Isopoda 35 35 Isoptera I I Lepidoptera 37 11 46 94 Limacoidea 35 35 Lymnaeoidea 198 198 Neuroptera I Odonata 33 34 Oligochaeta 21 21 Oribatida 2 2 Orthoptera 5 5 II Phthiraptera I 1 Planorbioidea 2 2 Polydesmida 215 215 Psocoptera 2 2 Spirostreptida 115 115 Trichoptera 306 306 Totals: 120 462 258 227 435 94 1596 (%) (7.5%) (28.9%) (16.2%) (14.2%) (27.3%) (5.9%) (100%) infrequently along slow-water stream reaches. None were recovered from rainbow trout stomachs, probably because of their preference for riparian rather than lotic breeding sites (Polhemus and Asquith 1996). MegaJagrion oresitrophum - this slender species was also moderately abundant in the study area, being found mainly in slackwater side pools and along stream overflow channels. Adults were also relatively common along small, shaded tributaries away from the main stream corridors. Although M. 60 oresitrophum would appear to be the native damselfly species at greatest risk for consumption by rainbow trout, due to its preference for breeding in pools, no remains were found in any of the 80 trout stomachs examined. Megalagrion eudytum - although M. eudytum remains a locally common species throughout Kaua'i, occupying an elevational range from sea level to 1220 m, it was encountered during the present study only at Koai'e Stream. This localized occurrence is due to the hygropetric habits of the immatures, which are found only on mossy rheocrenes or wet vertical waterfall faces (Polhemus and Asquith 1996), sometimes at considerable distances from the main stream channel. The breeding habitat requirements appear to be vertical faces of 2 m height or greater; small seeps or in-stream cascades are apparently unsuitable. No immature remains were found in trout stomachs, probably due to the mutually exclusive habitats occupied by the two species. As noted above, 5 endemic damselfly larvae, all representing M heterogamias, were present among the total of 1,596 identified prey items recovered from the 80 trout stomachs during this study (Table 2.5), giving a frequency ofoccurrence of6%. Four of these five Megalagrion larvae came from fish collected in Wai'alae Stream, with the remaining individual coming from Koai'e Stream. The endemic dragonfly Anax strenuus was more commonly consumed by rainbow trout, with 28 individuals collected from 20% of the examined trout stomachs (Table 2.5). In contrast to the riffle-dwelling immatures of M heterogamias, the larvae of A. strenuus prefer slow-water habitats, where they are apparently more susceptible to consumption by trout. Table 2.5. Summary numbers and percent frequency of native prey items of special concern collected in rainbow trout stomachs during this study, compared to the number of taxa collected per stream. Stream Native Native Other Native Introduced Other Prey Damselflies Dragonflies Aquatic Insects Aquatic Insects Items' Total Kauaikinana 0 4 II 128 164 307 Kawaikoi 0 0 2 4 34 40 Koai'e 1 23 17 99 363 503 Waiakoali 0 0 6 31 184 221 Wai'alae 4 1 51 169 300 525 Totals (%) 5 28 87 431 1045 1596 (0.3%) (1.8%) (5.5%) (27.0%) (65.4%) (100%) lIncludes snails, terrestrial insects, other invertebrates and vertebrates. 61 Other endemic aquatic insects exclusive of Odonata constituted 7.5% (by number) of the prey items found in Kaua'i rainbow trout stomachs. This probably reflects the fact that drift-feeding trout are finding relatively few endemic aquatic insects in the drift, and instead must rely on introduced aquatic insects, terrestrial insects, snails, and any other opportunistic food source, such as introduced crayfish or frogs, that becomes available. Numerically, the three most important items in the rainbow trout diet were introduced caddisflies (Trichoptera) at 19%, followed by terrestrial millipedes (Polydesmida) at 13%, and aquatic snails of unknown geographic origin (Lymnaeoidea) at 12% (Table 2.4). Trout diet appeared to vary between streams for reasons likely related to food availability. For example, lymnaeid snails were consumed by trout only in Koai'e and Wai'alae Streams, but then sometimes in great numbers, with up to 41 snails found in an individual trout. DISCUSSION Native to the Pacific drainages of westem North America, rainbow trout have been introduced into the continents of Africa, Asia, Australia, Europe, and South America and formed naturally reproducing populations in all of these areas (MacCrimmon 1971). In the Pacific Island region (excluding Hawai'i) unsuccessful attempts have been made to establish trout on Fiji (Andrews 1985), Tahiti (Maciolek 1984), and New Caledonia (Gargominy et ai. 1996). Most research on the impacts ofintroduced trout has taken place in Australia and New Zealand and focused on the impacts in relation to indigenous fish (Fletcher 1986; Townsend and Crowl 1991, Crowl et aJ. 1992; Cadwallader 1996), although a limited number of studies or observations have been published on trout impacts to invertebrates as well. The available literature regarding the impacts of trout on the invertebrate taxa of most interest to this study, Odonata, consists of one study in Australia (Faragher 1980) and observations in South Africa (Samways 1995; 1996; 1999). In New South Wales, an Australian dragonfly, Hemicordulia tau, was found to play an important role in the diet of brown and rainbow trout in Lake Eucumbene (Faragher 1980). However, the long-term survival of the dragonfly population in Lake Eucumbene was not believed to be affected by trout predation because of seasonal changes in trout prey composition and density, as well as varying lake levels (Faragher, 62 1980). In South Africa the situation appears to be different. Samways (1995; 1996; 1999) observed that a rare and threatened damselfly (Ecchlorolestes peringueyl) living in clear upland streams in the southwestern Cape area may have had its range restricted by introduced rainbow trout. This Gondwana relict damselfly was found only in stream reaches above waterfalls that rainbow trout could not surmount, and the larvae appear to be behaviorally susceptible to trout predation as they crawl on the surfaces of rocks and plants. Impacts of Rainbow Trout on Native Damselflies and Other Aquatic Invertebrates Rainbow trout are opportunistic predators and usually select the largest and most readily accessible prey (Tilzey 1977). Based on an analysis ofcontents from 80 individual trout stomachs, we conclude that rainbow trout in these systems are functioning primarily as opportunistic drift feeders, exploiting both aquatic insect drift, and the steady stream of terrestrial insects and other arthropods falling into streams from the adjacent forest. Although aquatic insect species accounted for 52.6% ofrainbow trout diet (by number), endemic aquatic insects amounted to only 7.5% of the diet. The low percentage of endemic aquatic insects directly reflects their low numbers in the stream drift, a phenomenon that is likely related to the natural absence in Hawai'i of the major orders of drift-prone stream insects such as Ephemeroptera (mayflies), Plecoptera (stoneflies), and Trichoptera (caddisflies) (Howarth and Polhemus 1991). Instead, most of the endemic aquatic insect fauna in the ~ve streams surveyed consisted of a diverse array of species dwelling in splash zones, all of which, though abundant, appear less prone to be entrained in the drift. While it is now impossible to recreate conditions prior to the introduction of rainbow trout into the inland waters of Kaua'i, our results also indicate that the endemic aquatic insect assemblages seen in the streams under study compare favorably with other isolated and unimpacted streams lacking introduced fish elsewhere in the Hawaiian archipelago (Polhemus 1995; Englund and Preston 1999). Unlike amphidromous animals such as endemic stream fish and shrimp, Megalagrion damselflies and other endemic Hawaiian aquatic insects complete the immature stage of their life cycles entirely within streams or other freshwater ecosystems. Until recently, such damselflies were common in virtually every such 63 ecosystem throughout the Hawaiian Islands, being found from slightly brackish basal spring wetlands near the ocean to the highest upland springs, seeps, rheocrenes, and streamlets, exploiting the full range of available lotic and lentic habitats (Polhemus and Asquith 1996). In recent decades, however, a combination of alien species introductions, stream channelization and diversion, and water quality degradation has caused a significant reduction or local extirpation of many endemic damselfly populations throughout Hawai'i (Polhemus and Asquith 1996; Englund 1999). Because Hawaiian streams differ from larger continental stream systems in that they are high gradient and do not generally form integrated networks, lateral and longitudinal movement of endemic aquatic insects occurs primarily during the aerial adult stage, rather than as upstream migration or downstream drift within the water column. As a result, the local extirpation of aquatic insect species such as damselflies within individual Hawaiian catchments can often be persistent, and has been conclusively documented from Bishop Museum collection records (Liebherr and Polhemus 1997; Englund 1999). Two primary ecological requirements for endemic damselflies are good water quality (clear, low turbidity water) and the absence of certain alien fish species such as tilapia (Sarotherodon melanotheron) and members of the family Poeciliidae. In particular, endemic damselflies have been shown to be sensitive to the presence of introduced poeciliid fishes such as the green swordtail (Xiphophorus hellen) or the shortfin molly (Poecilia mexicana) (Englund 1999). Extensive surveys on O'ahu have determined that the distributions of introduced Poeciliidae, including green swordtails, mollies and guppies (Poecilia reticulata), show a negative correlation with those of Megalagrion damselflies (Englund 1999), indicating an adverse interaction between the two groups. By contrast, a similar strongly allopatric distribution pattern was not evident in regard to introduced rainbow trout and endemic damselflies (or any other native aquatic insect taxon) in the K6ke'e State Park streams examined during this study. Our study found that Megalagrion account for only 0.3% of rainbow trout diet in upland Kaua'i. This compares to a 13% frequency of occurrence of Megalagrion recorded in the diet of the endemic Hawaiian 64 freshwater gobiid Lentipes concolor(n = 8) (Way and Burky 1991), the only published study to identify Megalagrion remains in the stomachs of Hawaiian native freshwater fish. This low dietary percentage, coupled with a high abundance of Megalagrion damselfly adults during the summer months along all the K6ke'e State Park streams sampled, and the presence of immatures in our benthic samples throughout the year, indicates that the low Megalagrion component in the trout diet is not simply indicative of the loss of this genus over time from the K6ke'e stream insect community. The fact that individual taxon frequency in trout diet is not merely a reflection ofrelative abundance of such taxa in the overall stream invertebrate community is further illustrated by the introduced midge Cricotopus bicinctus, which comprised a substantial portion of both the stream benthic fauna (23.3% by number in our benthic (Surber) samples) and also 46% of qualitative drift samples. Although it represented a dominant portion of stream drift and benthos, C. bicinctus comprised only 4% of trout diet, probably because this midge is small when compared to other available prey items. Likewise, the endemic midge Forcipomyia hardyi was the most important endemic component of stream drift, at 14%, but was absent from trout diet. Although Cricotopus bicinctus is nearly three times the size of the endemic F. hardyi (Hardy 1960), both taxa may either be too small for trout to notice, or their behaviors may exclude them from trout habitats such as deep pool areas. In any case, data from these taxa clearly further indicate that the proportional abundance of any given aquatic insect taxon does not necessarily correlate with its relative utilization by rainbow trout. Other Hawaiian aquatic insects, such as flies in the endemic genus Sigmatineurum are also sensitive to environmental disturbance (Evenhuis and Polhemus 1994; Englund and Preston 1999) and would presumably also be at risk for trout predation. However, this study found no indication of impacts from trout on such species. Habitat requirements for these endemic torrenticolous Diptera are quite narrow, and including seeps, and high velocity sections of riffles, cascades, and waterfalls. In contrast to other areas surveyed on Kaua'i (see Evenhuis and Polhemus 1994; Polhemus 1995), Sigmatineurum were common in 65 Koke'e streams with and without trout. The Kaua'i endemic Sigmatineurum napaJi was found in high numbers in two streams (Wai'alae and Koai'e) with naturally reproducing populations ofrainbow trout, and none were found in rainbow trout stomachs. These streams appear to have more robust populations of Sigmatineurum than other areas sampled in the Hawaiian Islands; during this study several hundred S. napali were captured at these sites, while only five individuals had ever been collected previously (Evenhuis and Polhemus 1994), despite a history ofextensive stream surveys on Kaua'i. The Future - Options and Concerns After examining 80 rainbow trout stomachs collected between 1997-1999 we have found no evidence that rainbow trout, from either stocked or naturally reproducing populations, significantly impact Megalagrion damselflies, Federally Endangered aquatic species such as Newcomb's Snail, or any other putatively sensitive endemic aquatic species not currently listed on the United States Federal Register. In fact, rainbow trout maintain a precarious existence in Kaua'i streams, apparently persisting only in those with the proper hydrological and geomorphological conditions. Streams at the identical elevations (1035 m) and less than a 1 km apart, such as Kauaikinana and Kawaikoi, vary widely in their ability to support trout. Contrary to the findings of Kido et a1., (1999), we found that stocked trout do not survive even in many of the upper elevation headwater reaches of the Koke'e State Park streams, much less in the warmer waters of the mid and terminal reaches below, and therefore represent a minimal threat to endemic biota in low-elevation areas. Furthermore, our findings that rainbow trout reproduce and hold-over only in high elevation sections of certain Koke'e area streams, feeding heavily on terrestrial drift, indicate that the impacts and risks, if any, to endemic aquatic invertebrate species as a result of the continued presence of trout in these catchments are relatively minor. These findings are at odds with the widespread but poorly supported perception that rainbow trout are highly destructive elements in Hawaiian stream ecosystems. This perception has arisen primarily from the reasoning that because the trout represent a conspicuous introduced species with known predatory habits, 66 they must automatically be having widespread deleterious effects. Our results clearly show that trout do exert some degree of predation pressure on endemic damselfly populations, but probably are no more harmful than endemic and indigenous gobiid species. Across the entire spectrum of threats facing Megalagrion damselflies, trout thus appear to represent a relatively minor component. Of far greater concern, in our opinion, are introduced poeciliid fishes, which are widespread in lower and middle stream reaches on all Hawaiian high islands, and have a demonstrable mutually exclusive distribution with endemic damselflies that would otherwise occupy such habitats (Englund 1999). By contrast, we have not been able to identify any indigenous or endemic Hawaiian aquatic insect species that appears to be eliminated from streams in which trout are present. Given this, the current concerns over the continued persistence of trout in the streams of upland Kaua'i, while understandable, are probably overstated. In terms of trout eradication, the only way to restore the upland streams in K6ke'e State Park to their original condition devoid of trout would be to use a chemical piscicide (fish poison) such as rotenone. Despite the rugged terrain that would make trout removal difficult, it would likely be possible, although prohibitively expensive. Not only would the extensive use of helicopters be required, but repeated treatments of the selected piscicide would be necessary. We see four major problems with using such methods for the complete or partial removal of trout from K6ke'e streams where naturally reproducing populations are maintained: 1) A piscicide would not necessarily eliminate other introduced aquatic fauna found in these watersheds. 2) Use ofpiscicides, particularly rotenone, would in all likelihood have non-target impacts on indigenous and endemic aquatic arthropods. Rotenone is toxic to insects as well as fish, and was originally used as an insecticide prior to the development of synthetic chemicals (Hynes 1970). Potential non-target impacts of the large-scale use ofrotenone in K6ke'e streams could include extirpation of locally common endemic aquatic insect species with ranges restricted to upper elevation areas, including Megalagrion 67 damselflies and endemic torrenticolous Diptera such as Sigmatineurum napali. The alternative piscicide antimycin would have lesser but still unpredictable impacts on such indigenous and endemic arthropods. 3) According to the results of our study, the removal of rainbow trout would not measurably enhance the indigenous and endemic aquatic fauna of these streams in terms ofoverall species composition. Kawaikoi Stream has not had rainbow trout stocked since 1992. Even so, the species composition of the endemic aquatic fauna in Hanalei, Lumaha'i, and Kawaikoi Streams in areas devoid oftrout is not markedly different from that of Koai'e, Kauaikinana, or Wai'alae Streams, ~l of which have naturally reproducing rainbow trout populations (Polhemus 1995). 4) Such a program would no doubt be politically unpopular; according to a recent review of the use of rotenone in North America from 1988-1997, public acceptance was one of the most important issues facing management agencies when using this method (McClay 2000). The chemical treatment of Koke'e State Park streams to remove rainbow trout would face nearly unanimous opposition from anglers on Kaua'i and throughout Hawai'i, and if the streams were restored to their formerly fishless state, it is almost certain these streams would be clandestinely restocked with rainbow trout by anglers, thus defeating any control or eradication efforts. Given these considerations, coupled with the minimal impact that trout appear to be having on such systems, the potential cure is clearly worse than the current problem. Rainbow trout, though undesirable, appear to have low impacts on endemic aquatic fauna, and their elimination would therefore be unlikely to result in substantive benefits to any endemic aquatic species. Unless extreme actions are regularly undertaken to remove rainbow trout, there will continue to be naturally reproducing populations in the upper reaches of certain catchments draining from the Alaka'i Plateau, at least in the near term. Considering the marginal existence of these limited naturally reproducing Kaua'i trout populations, however, it seems likely that they will decline over time in any case, falling victim to normal stochastic variations in temperature and discharge rate typical of Hawaiian stream environments, so that these streams will return naturally to a closer semblance of their original states. We believe that future efforts aimed at control of 68 introduced fishes in relation to Hawaiian damselfly conservation would be better concentrated on elimination of poeciliids, which constitute a much more pervasive and demonstrable threat to the long term survival of many rare Megalagrion species, particularly in the lowlands (Englund, 1999). Assessing the relative degrees of threat posed by the various constituents in suites of introduced taxa will continue to be a challenge for conservation managers, particularly on vulnerable tropical oceanic islands. The current study has illustrated that certain introduced taxa can take on the status of "flagship threats" due to their conspicuous or notorious nature, even though they may in fact represent lesser threats than other, less obvious introductions. In a world of limited resources for conservation initiatives, it is therefore prudent to obtain comparative data as to impacts, and to then channel resources first and foremost toward control of those species that represent the most clear and present danger. 69 REFERENCES Andrews, S. 1985. Aquatic species introduced to Fiji. Domodomo 3: 67-82. Brown, L. R., Brasher, A M., Harvey, B. c., and Matthews, M. 1999. Success and failure of nonindigenous aquatic species in stream systems: case studies from California and Hawaii. In: Nonindigenous freshwater organisms: vectors, biology, and impacts, pp. 415-430. R. Claudi and J.H. Leach (eds.). CRC Press. Cadwallader, P. L. 1996. Overview of the impacts of introduced salmonids on Australian native fauna. The Director of National Parks and Wildlife, Australian Nature Conservancy Agency. Canberra, ACT. 64 pp. Crowl, T. A, Townsend, A R., and McIntosh, A R. 1992. The impact of introduced brown and rainbow trout on native fish: the case of Australasia. Reviews in Fish Biology and Fisheries 2: 217-241. Denning, D. G. and Beardsley, J. W. 1967. The collection of Cheumatopsyche analis in Hawaii. Proceedings ofthe Entomological Society ofWashington 69: 56-57 Dudgeon, D. 1999. Tropical Asian streams. Hong Kong University Press. 830 pp. Englund, R. A. 1999. The impacts of introduced poeciliid fish and Odonata on endemic Megalagrion (Odonata) damselflies on Oahu Island, Hawaii. Journal ofInsect Conservation 3: 225-243. Englund, R. A and Preston, D. J. 1999. Biological Assessment of the Lower Hamakua Ditch on the Hawaiian Stream Fly (Sigmatineurum meaohl) and other aquatic insects. Contribution No. 1999 003 to the Hawaii Biological Survey, Bishop Museum. 31 pp. 70 Evenhuis, N. L. and Polhemus, D. A. 1994. Review of the endemic Hawaiian genus Sigmatineurum Parent (Diptera: Dolichopodidae). Bishop Museum Occasional Papers 37: 1-19. Faragher, R. A. 1980. Life cycle of Hemicordulia tau Selys (Odonata: Corduliidae) in Lake Eucumbene, N.S.W., with notes on predation on it by two trout species. Joumal ofthe Australian Entomological Society 19: 269-276. Filbert, R. and Hawkins, C. P. 1995. Variation in condition of rainbow trout in relation to food, temperature, and individual length in the Green River, Utah. Transactions ofthe American Fisheries Society 124: 824-835. Fletcher, A. R. 1986. Effects of introduced fish in Australia. In: P. De Deckker and W.D. Williams (eds.), Limnology in Australia, pp. 231-238. Dr. W. Junk Publishers, Dordrecht. Gargominy, 0., Bouchet, P., Pascal, M., Jaffre, T., and Tourneur J.-C. 1996. Consequences des introductions d'especes animales et vegetales sur la biodiversitie en Nouvelle-Caledonie. Revue D'ecologie (La Terre Et La Vie) 51: 375-402. Hardy, D. E. 1960. Insects of Hawaii, Volume 10, Diptera:Nematocera-Brachycera. University of Hawaii Press. 368 pp. Hazlett, R. W. and Hyndman, D. W. 1996. Roadside geology of Hawaii. Mountain Press Publishing Company, Missoula, Montana. 307 pp. 71 Howarth, F. G. and Polhemus, D. A. 1991. A review of the Hawaiian stream insect fauna. In: New directions in research, management, and conservation ofHawaiian freshwater stream ecosystems, Proceedings ofthe 1990 Symposium on Freshwater Stream Biology and Management, State of Hawaii, pp. 40-51. Hawaii Division of Aquatic Resources, Honolulu. Hynes, H. B. N. 1970. The ecology of running waters. Liverpool University Press, United Kingdom. 555 pp. Kido, M.H., Heacock, D.E. and Asquith, A. 1999. Alien rainbow trout (Oncorhynchus mykiss) (Salmoniformes: Salmonidae) diet in Hawaiian streams. Pacific Science 53: 242-251. Kimball, D. c., and Helm, W. T. 1971. A method of estimating fish stomach capacity. Transactions ofthe American Fisheries Society 100: 572-575. Liebherr, J. and Polhemus, D. A. 1997. Comparisons to the century before: the legacy of the R.C.L. Perkins and Fauna Hawaiiensis as the basis for a long-term ecological monitoring program. Pacific Science 51: 490-504. Maciolek, J. A. 1984. Exotic fishes in Hawaii and other islands of Oceania. In: W.R. Courtenay, Jr. and J.R. Stauffer, Jr. (eds.), Distribution, Biology, and Management of Exotic Fishes, pp. 131-161. John Hopkins University Press, Baltimore and London. MacCrimmon, H. R. 1971. World distribution of rainbow trout (Salmo gairdnen). Journal ofthe Fisheries Research Board ofCanada 28: 663-704. Mathsoft, Inc. 1999. S-Plus 2000. Seattle, WA. Software. 72 McClay, W. 2000. Rotenone use in North America 1988-1997. Fisheries 25: 15-21. Merritt, R.W. and Cummins, KW. (1996) An introduction to the aquatic insects of North America. Third Edition. Kendall Hunt Publishing Company, Dubuque Iowa. 862 pp. Needham, P. R, and Welsh, J. P. 1953. Rainbow trout (Salmo gairdneri Richardson) in the Hawaiian islands. Journal ofWildlife Management 17: 233-255. Needham, P. Rand Usinger, R. L. 1962. The introduction of mayflies into Hawaii. In: Job Completion Report, Investigations Projects, Proj. No. F-4-R-lO. Hawaii Division of Fish and Game, Bureau of Fisheries. Honolulu, Hawaii. 10 pp + appendices. Nishida, G.M. 1997. Hawaiian terrestrial arthropod checklist, 3rd Edition. Hawaii Biological Survey. Bishop Mus. Tech. Rep. 12.263 pp. (Searchable database on the web at http://www.bishopmuseum.orglbishop/ento/entodbhome.html). Polhemus, D. A. 1995. A survey of the aquatic insect faunas of selected Hawaiian streams. Report for the Commission on Water Resource Management, Department ofLand and Natural Resources, State of Hawaii. Department of Natural Sciences, Bishop Museum. Hawaii Biological Survey Contribution No. 1995-008. Polhemus, D.A. and Asquith, A. 1996. Hawaiian damselflies: a field identification guide. Bishop Museum Press, Honolulu. 122 pp. Ramsey, F. L. and Schafer, D. W. 1997. The Statistical Sleuth. Duxbury Press, Belmont, CA. 742 pp. 73 Samways, M. J. 1995. Conservation of the threatened, endemic dragonflies of South Africa. In: Proceedings of the International Symposium on the Conservation of Dragonflies and Their Habitats. P.S. Corbet, S.W. Dunkle and H. Ubukata (eds), pp. 8-15. Japanese Society for Preservation of Birds, Kushiro. 70 pp. Samways, M. J. 1996. Workability of the new 'IUCN Categories ofThreat' with Odonata in South Africa. Odonatologica 25: 347-354. Samways, M. J. 1999. Diversity and conservation status of South African dragonflies (Odonata). Odonatologica 28: 13-62. Smith, G. C. (2000) Rediscovery of the introduced mayfly Caenis nigropunctata (Klapalek, 1904) (Ephemeroptera: Caenidae) in Waimanalo Stream, O'ahu. Bishop Museum Occasional Papers 64: 30-31. Townsend, C. R. and Crowl, T. A. 1991. Fragmented population structure in a native New Zealand fish: an effect of introduced brown trout? Oikos 61: 347-354. Tilzey, R. D. 1. 1977. Key factors in the establishment and success of trout in Australia. Proceedings ofthe Ecological Society ofAustralia 10: 97-105. Usinger, R. L. 1972. Robert Lee Usinger: autobiography of an entomologist. California Academy of Sciences. San Francisco. 330 pp. 74 Way, C. M. and Burky, A. J. 1991. A preliminary survey of macroinvertebrates and a preliminary assessment of the diet ofthe endemic Hawaiian goby ('o'opu alamo '0), Lentipes concolor(Gill). In: New directions in research, management, and conservation of Hawaiian freshwater stream ecosystems, Proceedings of the 1990 Symposium on Freshwater Stream Biology and Management, State of Hawaii, pp. 158-164. Hawaii Division of Aquatic Resources, Honolulu. 75 CHAPTER 3. LONG-TERM MONITORING OF ONE OF THE MOST RESTRICTED INSECT POPULATIONS IN THE UNITED STATES, Megalagrion xanthomelas Selys-Longchamps, 1876, AT TRIPLER ARMY MEDICAL CENTER, 0'AHU, HAWAI'I 76 ABSTRACT Long-term monitoring of a remnant population of the Candidate Threatened Megalagrion xanthomelas (orangeblack Hawaiian damselfly) located at TripIer Army Medical Center (TAMC), O'ahu, Hawai'i began in May 1997 and continued to February 2000 for the mitigation ponds and June 2000 for the TAMC stream. This species has been reduced to little more than 100 m ofstream habitat on O'ahu at the TAMe. Threats to M. xanthomelas include alien fish species, stream dewatering, and habitat alteration. The TAMC stream now requires augmented water flow because construction of a facility up gradient of the TAMC stream disrupted the normal hydrology of the small stream. The O'ahu race of M xanthomelas will soon become extinct if the stream were allowed to become dry, as nearly happened in June 1997. The most cost effective way to ensure the survival of this species on 0'ahu would be to continue some mitigation water flows to the TAMC stream. The next step would be the establishment of another wild population of M. xanthomelas to a stream lacking alien fish species. It is highly recommended that a cooperative association of biologists from the Bishop Museum, University of Hawai'i, U.S. Fish and Wildlife Service, and U.S. Army environmental staff continue to monitor the population of M. xanthomelas, arguably the rarest insect population in the United States. INTRODUCTION Monitoring of the last remnant O'ahu population of Megalagrion xanthomelas Selys-Longchamps, 1876, a Candidate Threatened species on the U.S. Federal Register, began in May 1997 and continued through February 2000. This damselfly population is located in a small unnamed stream (henceforth called TAMC stream) in a U.S. military installation at the TripIer Army Medical Center (TAMC) near Honolulu, Hawai'i (Figure 3.1). Although formerly a common insect on O'ahu (Liebherr and Polhemus 1997; Englund 1999) this species has currently been reduced to little more than 100 m of stream habitat. Threats to the endemic M. xanthomelas include alien fish species, stream dewatering, and habitat alteration (Polhemus 1996). This species was considered extinct on 0'ahu until the discovery of this remnant population by N.L. Evenhuis of the Bishop Museum in 1994 (Evenhuis and Cowie 1994). Concerns were raised about the 77 Megalagrion xanthomelas 0= Historic records (1892-1989) Tripier Army Medical Center • = Recent records (1990-2000) Figure 3.1. Map of O'abu, Hawai'i showing locations of current and historic records for Megalagrion xanthomelas (from Evenhuis et a1., 1995). long-term viability of this remnant damselfly population because construction of a large U.S. Veterans Administration facility had the potential to disturb the watershed. Prior to the construction and large-scale upslope disturbance, four mitigation ponds constructed adjacent to the stream (Figure 3.2) and were completed in October 1995. These ponds were built to ensure the survival of M. xanthomelas in case a flood event or other catastrophe disrupted their habitat. Because ofbudgetary constraints the mitigation ponds were drained and dismantled in March 2000, however, stream augmentation flows currently provide suitable aquatic habitats for M. xanthomelas. 78 Figure 3.2. TAMC Mitigation ponds prior to drainage, February 2000. STUDY AREA The TAMC stream is located on leeward O'ahu at 79 m elevation, and flows through a forest ofintroduced plants. An in-depth description and map of the stream study area can be found in Evenhuis et al. (1995), Polhemus (1996), and Pange1inan (1997). The TAMC stream now requires augmented water flow because construction in 1995 ofa large Veterans Administration facility up gradient of the TAMC stream disrupted the normal hydrology of this small stream. The cement-lined mitigation ponds are approximately 200 m northwest of the TAMC stream, and measure 7.3 m long by 3 m wide and have an average water depth of 0.6 m. These ponds contained cobble substrate brought from the TAMC stream and aquatic plants such as algae, water lily (Nymphaea sp.), water lettuce (Pistia stratiotes), and a large aquatic sedge (Cyperus altemifolius). To ensure the survival of this damselfly species, quarterly monitoring of the four mitigation ponds began in May 1997 and continued until their being dismantled in February 2000. Quarterly monitoring of the TAMC stream began in May 1997 and continued until June 2000. 79 METHODS Stream Sampling The objectives of the damselfly mark-recapture sampling efforts were to 1) document recruitment of new individuals to the population between quarterly sampling efforts, 2) assess the relative abundance of damselflies between monitoring periods using a standardized methodology, and 3) provide a quick means of determining if the TAMC M. xanthomelas population was threatened by disturbance or stream dewatering. The entire length of wetted stream (95 m) was sampled starting at the upstream end of the man-made culvert where stream flow originates. Methods used were identical to previous research conducted on the M xanthomelas TAMC population (Pangelinan 1997). For this study, damselflies were marked beginning on 15 May 1997. Two observers conducted monitoring after the fIrst quarter (May 1997) in which methods were established. While slowly walking down the stream each observed damselfly was netted and its wings marked with a permanent black extra fIne felt tip marker. The number was recorded if a captured damselfly had been previously marked. After completing the slow downstream walk, which would take up to three hours, we returned slowly upstream and marked any previously unmarked damselflies. Collection and observation times were consistent during each quarterly monitoring event to standardize sample effort, and ranged between 3.5 to 4.0 hours. Individuals were not counted unless they were netted and the wings marked with a number. Pond Sampling Thirty-minute damselfly counts were conducted at each of the four concrete mitigation ponds. Damselfly sex and behavioral activity was recorded during each thirty-minute count. Individual adults were not counted unless they were captured and the wings marked with a number. Quantitative aquatic net samples were taken at the ponds starting during the November 1997 monitoring. Three aquatic net sweeps approximately 1.25 m in length were taken at the surface, middle, and bottom of each pond. The net contents were then placed in a 500-micron sieve and inspected for immature damselfly naiads. 80 The number of distinctive M. xanthomelas oviposition scars on water lily leaves in the mitigation ponds was counted each quarter to provide a measure of relative breeding attempts. After observing female M. xanthomelas ovipositing on water lilies, it was easy to distinguish Megalagrion oviposition scars from dragonflies and Ischnura spp. Megalagrion xanthomelas leaf scars were larger and more curved when compared to Ischnura spp. oviposition leaf scars. The presence of other aquatic insects was also recorded. Additionally, the presence of dragonflies, ants, water boatmen (Notonecta indica) and other potential predators or threats was noted at each pond. Pond levels were noted and when necessary the pond outlets were cleaned. RESULTS TAMC Damselfly Counts Stream Long-term monitoring of M. xanthomelas populations at the TAMC stream indicates a robust damselfly population after flow restoration started in June 1997 (Englund 1998). During the present study the number of adult damselflies marked each quarter varied from a low of 17 in May 1997 to a high of 162 in February 1998 (Figure 3.3). Observations of adults atthe TAMC stream were lower in the month of December 1999, and winter 1999 counts may not be directly comparable to other monitoring periods because of the cool, cloudy weather. Four separate monitoring attempts were made at the TAMC stream in November and December of 1999. Megalagrion xanthomelas adults are only active during sunny weather (Polhemus 1996), and a long period ofcloudy, cool, and very wet weather in the final quarter of 1999 reduced the numbers of captured adult damselflies at the TAMC stream. Additionally, the most severe weather and heavy flooding during the entire study period in November and December 1999 may have caused some adult mortality at the stream area. During this time heavy rains scoured the streambed to bedrock, and caused minor rockslides into some areas of the TAMC stream. Fortunately, numerous immature individuals were observed in the stream immediately after the heavy rains and flooding in December 1999. The restoration of stream flow 81 180 III 160 - ~ r- .a 140 - r- Q. - ~ 120 r- - r- r- - ~ 100 ~ 80 - E - nl 60 C nl 40 r- .... r- ~ 20 o n n May- June July Sept Nov Jan- Feb May Aug Nov Jan- May Aug Dec Feb- Jun- ~ 00 W 00 00 Figure 3.3. Megalagrion xanthomelas captures at TAMC stream, May 1997-June 2000 continues to provide optimal aquatic habitats at the TAMC stream, and also ensured this damselfly population could survive a severe environmental disturbance in the form of heavy flooding in November and December of 1999. Monitoring conducted in February and June 2000 confirmed December results resulted from poor weather conditions, and subsequent visits to the TAMC stream indicate M. xanthomelas remains abundant. Captures of M xanthome1as at the TAMC mitigation ponds were variable, but overall the ponds received some use during each quarter. During the present study the number of adult damselflies marked at the ponds each quarter varied from a low of 0 in May 1997 to a high of 23 in January 1998 (Figure 3.4). During this study, immature Megalagrion damselflies were not collected in dip net sampling at the surface, middle, and bottom layers of each pond. However, at least two cast skins and one teneral M. xanthomelas were found at 82 the ponds during the time-frame of this study by u.s. Fish & Wildlife Service personnel, indicating that there was at least limited reproductive success between 1997-2000. 1/1 25 l!! r-- :::J - C. 20 ltl () ~ 15 !E - G) - 1/1 10 - E r-- ltl - - - C 5 r-- - ...ltl o 0 n n n I- May- June July Sept Nov Jan- Feb May Aug Nov Jan- May Aug Dec Feb- W W ~ 00 Figure 3.4. Megalagrion xanthomelas captures at TAMC mitigation ponds from May 1997-February 2000. Five species of native and introduced dragonflies were abundant at the ponds both as adults and larva. Two native dragonfly species, AnaxJunius and Pantala flavescens were abundant, as were the introduced Tramea abdominalis, Orthemis ferruginea, and Crocothemis servilia. These immature dragonflies were captured in virtually every dip net sample. Cast skins of these dragonflies were also common and found on vegetation within the ponds. Two species of introduced damselflies were also observed and captured at the ponds, Ischnura posita and Ischnura ramburii, and cast skins of these damselflies were also found at the ponds. Introduced water boatman (Notonecta indica) were common and a potential predator, and introduced long- legged ants (Anoplolepes longipes) were observed by u.S. Fish & Wildlife Service personnel (A. Pangelinan, pers. comm.) to kill freshly emerged M. xanthomelas at the ponds. The most shaded and lowest pond received the greatest use by adult damselflies, with most damselflies being caught at this pond. 83 Damselfly oviposition scars observed on water lily leaves likely reflect a lesser percentage of attempts by females to lay eggs, as it has been observed that M. xanthomelas preferentially oviposit on algae both at the TAMC ponds and on Molokai (Englund, pers. obs.). However, these scars provided a useful relative index of breeding attempts at the ponds. Oviposition scars on lily leaves declined since the peak in January 1998 and were variable thereafter (Figure 3.5). The presence of some oviposition scars on water lily leaves indicates breeding occurred at the mitigation ponds throughout the study period. Thick algal mats were a continual problem in the ponds and necessitated weekly removal of algae. Pangelinan (1997) found very limited movement between the TAMC stream population and pond population; thus most of the individuals observed at the ponds were likely the result of pond reproduction. 20 l!! lU (J en 15 - l: 0 - E r-- //I 10 - 0 - - Q. - - '> 0 5 - - ~ J 0 n n n Jul-97 Sept Nov Jan- Feb May Aug Nov Jan- May Aug Dec Feb 98 99 00 Figure 3.5. Oviposition scars on water lilies at the TAMC mitigation ponds July 1997-February 2000. On 2 March 2000 the mitigation ponds were drained and as many damselfly larvae as possible were recovered and saved for later identification. Hundreds of Ischnura posita larvae along with high numbers of dragonfly larvae were examined, but no M. xanthomelas larvae were found. 84 DISCUSSION The experience of preserving the O'ahu population of M. xanthomelas at TAMC for the past 6 years has been one of the most notable conservation success stories in the Hawaiian Islands. A catastrophic flood in 1995 killed the entire stream population because of a massive sediment input into the stream during construction of the previously mentioned Veterans Administration facility. The mitigation ponds provided a reserve population for M. xanthome1as that would otherwise have gone extinct on O'ahu. The completion of the ponds and translocation of immatures and adults in October 1995 was fortuitous as the flood occurred in November 1995. This success has not come without a major expenditure of money and time by U.S. government agencies such as the U.S. Department of Veterans Affairs and the U.S. Fish & Wildlife Service. For example, the mitigation pond planning, engineering and construction cost totaled U.S. $200,000, while pond maintenance cost $6000 per year, starting in 1995. Other major expenses include a $200-300 per month water bill for maintaining stream flows and pond levels. While the mitigation ponds initially proved to be a major success and ensured the survival of M. xanthome1as on O'ahu, the long-term success in native damselfly recruitment was less dramatic, and hampered by sometimes unfavorable conditions. When the ponds were first filled in 1995 there were low numbers of predators such as dragonflies, and the water lacked thick algae growths. In 1995, reproductive success of M. xanthomelas was definitive and verified by numerous cast skins found around the ponds (Zoll 1995). However, over time, thick growths of filamentous green algae were the biggest problem at the ponds, interfering with emerging larvae and resulting in stagnant water conditions. Floating algae mats also allowed ants access to the interior of the ponds where they would prey on emerging M. xanthomelas. High densities ofnative and introduced dragonfly larvae at the ponds undoubtedly influenced M. xanthomelas numbers. These dragonflies were noticeably absent from the TAMC stream most likely because thick riparian vegetation and flowing water did not provide favorable dragonfly habitat. 85 With hindsight it appears that the ponds could have been designed to more closely mimic conditions found at the TAMC stream. The sunny and hot exposure of the ponds was the biggest problem because water temperatures at the ponds were much higher than in the TAMC stream. For example during a warm period in August 1999 pond temperatures were 27°C while stream temperatures were 22.5-23°C. A long-term solution for reducing water temperature and also algae growth at the ponds would have been to plant shade trees around the ponds. In contrast to the TAMC ponds, the stream area is heavily shaded but still provides some sun penetration at midday. In the future it is recommended that mitigation habitat more closely resemble a stream channel with riffles and pools and maintaining a moderate current velocity. This would also reduce predatory dragonfly numbers in the mitigation habitat. The TAMC stream nearly became dry in June 1997 (EnglundI998) after construction upslope of the stream two years earlier disrupted the hydrology of this small watershed. However, because of mitigative water flows paid for by the U.S. Department of Veterans Affairs, stream flow remained at 100% of stream length and aquatic habitats were optimal at the TAMC stream through May 2000. In March 1999, the U.S. Army and U.S. Fish & Wildlife service attempted to translocate M. xanthomelas to another O'ahu stream in attempts to establish a second O'ahu population. Fifty-five adults and 44 larvae were collected from the TAMC stream and transported to an unnamed stream near Dillingham Airfield. This attempt apparently failed because the translocation stream contained the introduced red swamp crayfish [Procambarus c1arkii (Girard, 1852)] (K. Johnson pers. comm.), a species absent from the TAMC stream. The taking of a large number of adult and immature damselflies from the TAMC stream had no observable impact on this restricted population (Figure 3.3). Although a more suitable translocation site needs to be found, the lack of impacts to the TAMC damselfly population is encouraging. It is imperative that current mitigation flows be provided at the TAMC stream for a minimum of at least two years after it has been shown that translocated populations of M. xanthomelas are naturally reproducing. If successful reproduction is observed in the form of unmarked damselflies at the next relocation site, then monitoring at this site 86 should be continued for at least two years to ensure the population is self-sustaining. An example that translocation of a rare tropical damselfly species can be successful is the translocation ofM xanthomelas from the TAMC stream to the mitigation ponds. Monitoring has also shown that removal of damselflies for translocations does not adversely effect populations at the TAMC stream. Without mitigation stream flows the TAMC stream will cease to flow and cause the extinction of damselflies from the TAMC stream, and O'ahu. Thus, it would be prudent to make additional translocations while the TAMC damselflies are abundant, perhaps to a second relocation site at the U.S. Naval Magazine at Lualualei, O'ahu. Possible candidate relocation streams were observed during Bishop Museum surveys of Lualualei (Evenhuis 1997a), although some of these streams may need to be treated with rotenone to remove alien poeciliid fish such as the western mosquitofish (Gambusia affinis), that are believed to be responsible for the decline of this species (Englund1999). The most obvious and cost-effective way to ensure the survival of M. xanthomelas would be to continue at least some water flow to the TAMC stream. The $200-300 per month water bill to provide mitigation flows is a small price to pay for one of the most formerly common insects on O'ahu. Construction adversely affected water flow at the TAMC stream and caused the stream to become dry before mitigation flows were implemented in June 1997. Now that the ponds have been dismantled there is only one O'ahu location where M. xanthomelas can be found. As stated earlier, the fate of the translocation efforts for this species are very uncertain, and have failed in at least one attempt to move them out of the TAMC watershed. Thus, to ensure the survival of this species it is necessary to maintain mitigation flows at the TAMC stream until after at least two additional populations have been conclusively established in other areas on O'ahu. Recent genetic work on M. xanthomelas indicates that populations between the various Hawaiian islands can be quite different (S. Jordan pers. comm.). Although the mitochondrial haplotype obtained from TAMC 87 specimens has also been found on Molokai and Lanai, other loci might reveal significant differences between the O'ahu TAMC population and other islands, however, more work needs to be done. The O'ahu population also has much less genetic diversity than populations on other islands, probably due to genetic bottlenecks (S. Jordan pers. comm.). Therefore it is critical that habitats and populations from O'ahu be maintained to preserve the distinct O'ahu race of this damselfly. In conclusion, it is recommended that a collaboration of specialists from the Bishop Museum, University of Hawai'i, U.S. Fish and Wildlife Service, and U.S. Army environmental staff continue to monitor the TAMC population of the M. xanthomelas. Additionally, this collaborative team should work on establishing at least two separate populations on the island of O'ahu. Until other populations have been established, monitoring on at least a quarterly basis at the TAMC stream could identify potential threats and ensure proper management actions are taken to allow this rare damselfly to survive. 88 REFERENCES Englund, R. A. 1999. The impacts of introduced poeciliid fish and Odonata on endemic Megalagrion (Odonata) damselflies on Oahu Island, Hawaii. Journal ofInsect Conservation 3: 225-243. Englund, R. A. 1998. Response of the orangeblack Hawaiian damselfly (Megalagrion xanthomelas), a candidate threatened species to increases in stream flow. Bishop Museum Occasional Papers 56: 19-24. Evenhuis, N. L. and Cowie, R. H. 1994. A survey of the snails, insects and related arthropods in the grounds of the TripIer Army Medical Center, Honolulu, Hawaii. Bishop Museum Technical Report 3. 21 pp. Evenhuis, N. L, D. Polhemus, S. Swift, K. Arakaki and Preston, D. 1995. A study of the biology of the orangeblack Hawaiian damselfly (Megalagrion xanthomelas), with special reference to conservation of the population at TripIer Army Medical Center, Oahu. Bishop Museum Technical Report 8: 81 pp. Evenhuis, N. L. 1997. Diversity of insects and related arthropods of the Naval Magazine Lualualei, Headquarters Branch, Oahu, Hawaii. Final report prepared for the U.S. Navy. Bishop Museum Technical Report 9: 170 pp. Liebherr,1. and Polhemus, D. A. 1997. Comparisons to the century before: the legacy of the R.C.L. Perkins and Fauna Hawaiiensis as the basis for a long-term ecological monitoring program. Pacific Science 51: 490-504. 89 Pangelinan, A. A. 1997. Demography and life history of the orangeblack Hawaiian damselfly (Megalagrion xanthomelas) (Selys-Longchamps, 1876) on Oahu, Hawaii. Masters Thesis, University of Guam. Polhemus, D. A. 1996. The orangeblack Hawaiian damselfly, Megalagrion xanthomelas (Odonata: Coenagrionidae): clarifying the current range of a threatened species. Bishop Museum Occasional Papers 45: 30-53. Zoll, M. 1995. Orange-black damselfly mitigation at TripIer Army Medical Center: population enhancement via artificial ponds. Final Report to U.S. Department of Interior, U.S. Fish & Wildlife Service, Pacific Islands Ecoregion, Honolulu, Hawaii. 30 pages. 90 CHAPTER 4. THE LOSS OF NATIVE BIODIVERSITY AND CONTINUING NONINDIGENOUS SPECIES INTRODUCTIONS IN FRESHWATER, ESTUARINE, AND WETLAND COMMUNITIES OF PEARL HARBOR, 0'AHU, HAWAIIAN ISLANDS. 91 ABSTRACT The benthic organisms and fishes of the estuarine, lower stream areas, and wetlands of Pearl Harbor were sampled from 1997-1998 as a companion study to inventories conducted in the more marine areas of Pearl Harbor. This study, the first comprehensive assessment of the non-marine areas of Pearl Harbor, found that nonindigenous species comprise the dominant portion of the biota in the estuaries, wetlands, and lower stream reaches. Many species of aquatic organisms have been introduced into these areas, altering the species composition of the aquatic fauna, resulting in the lower portions of Pearl Harbor streams, springs, and wetlands being dominated by nonindigenous species. A total of 191 aquatic species in 8 phyla were identified in the estuarine reaches of Pearl Harbor. Nonindigenous species dominated with 48% of the species recorded, while only 33% were native and 19% were cryptogenic. Two new nonindigenous species to Hawai'i were found during this study--a species offang-toothed blenny (Omobranchus ferox) and an estuarine hydrobiid snail (Pyrgophorus cf. coronatus) introduced from the Caribbean. No geographic region is a predominant source of aquatic species introductions into the Pearl Harbor area, although more species come from the Americas than other areas. For example, 57% originated from the Americas, 30% from Asia and the Pacific, 5% from Australia/New Zealand, 5% have a world-wide distribution, with fewer than 3% of species originating from Africa. An increase in non-native species in the freshwater and estuarine portions Pearl Harbor will probably continue. This is because of the wide variety of sources from which introductions take place. The majority of nonindigenous species appear to have come from five major sources: 1) intentional and accidental aquarium releases, 2) intentional biocontrol releases, and 3) intentional food source releases, 4) ballast water or hull fouling releases, 5) brought in with airplanes. INTRODUCTION Numerous studies have been conducted in marine habitats of Pearl Harbor (Coles et a1. 1999), but little baseline research has been conducted in the lower reaches of streams and coastal wetlands in the harbor despite the large extent of these habitats. The large Pearl Harbor spring, coastal wetland, and riverine systems represent an ecologically important and unique natural resource and formerly contained a significant 92 endemic fish and invertebrate fauna (Titcomb 1972). However, prior to the present survey little was known about the current status of the fauna in these areas. Over the past several hundred years the Pearl Harbor watershed has undergone tremendous environmental degradation, changing from an area offishponds and taro fields with reportedly high water quality in pre European contact times (before 1778) to a highly urbanized area with poor water quality (Coles et a1. 1999). Many spring, wetland, and stream mouth areas have been channelized and filled, and the arrival of mangroves in the early part of the 20th century (Wester 1981) also considerably changed the shoreline. Nonindigenous species are an increasing threat to Hawaiian stream, wetland, estuarine, and anchialine pond ecosystems. Not only do nonindigenous aquatic species in tropical Pacific insular environments compete with and prey upon native species (Eldredge 1994), they have also brought with them a complement of diseases and parasites to which native species are not resistant (Font and Tate 1994). The severity of nonindigenous species impacts varies according to island, elevation, and watershed; with adjacent streams often having significantly different compositions of nonindigenous species (Englund et a1. 2000a). Even the most remote estuarine and anchialine habitats found in the main Hawaiian Islands contain nonindigenous aquatic species (Maciolek 1984). However, some relatively pristine stream, wetland, and anchialine pond areas can still be found on the islands of Kaua'i, Maui, Molokai, and Hawai'i, and still have robust populations offreshwater and estuarine native fish, crustaceans, mollusks, and aquatic insects (Maciolek 1984; Polhemus 1995a). Because of naturally low base flows (Nichols et a1. 1997) and watersheds being smaller and shorter, O'ahu streams and estuaries have a lower flushing out capability than the generally larger streams found on other islands such as Kaua'i and Hawai'i. Many formerly common native aquatic insect species are now absent or rare in many lower elevation O'ahu stream and estuarine areas, including Pearl Harbor, and appear to have been displaced by nonindigenous species (Polhemus 1996; Englund 1999). 93 In Hawai'i, estuarine habitats are important for a wide variety of native species including several species of culturally important food fish such as the native Mugil ceph81us and the endemic Kuhlia sandvicensis and Awaous guamensis. Native crustaceans such as Macrobrachium grandimanus and Atyoida bisulcata also require estuaries for various stages of their life histories. Hawaiian streams exhibit low endemicity among the freshwater and estuarine macrofauna because amphidromy (Meyers 1949) has lead to gene flow between islands (Zink et al. 1996). Also, the relatively small quantity of non-marine aquatic habitats found in Hawai'i compared to continental areas has led to low native diversity. Nonetheless, these habitats supported an important native fauna. By 1991, however, at least 58 intentionally and accidentally introduced freshwater species (excluding aquatic insects) had become established throughout the Hawaiian Islands (Devick 1991). The purpose of the present study was to assess the aquatic biodiversity of freshwater and brackish habitats in Pearl Harbor and identify known or probable origins and mechanisms of nonindigenous species introductions. Study Area Located in south-central O'ahu (Figure 4.1), seventy percent of the natural freshwater discharge into Pearl Harbor originates from a spring complex that is the largest and most significant in the Hawaiian Islands, with an average combined pre-development flow of about 8 m3/sec (Nichols et 81. 1997). Although the abundance offresh water in spring areas within Pearl Harbor has diminished because of upgradient groundwater pumping (Nichols et al. 1997), this area still supports one of the largest coastal spring and wetland systems in the Pacific Islands. Pearl Harbor is formed from a drowned river system that has been submerged during various glacial epochs, with oyster beds and thin coral reefs flourishing during periods of higher sea level (Stearns 1985). The lower sections of Pearl Harbor streams, wetlands, and springs now lie largely over a fill of oyster beds, reefs, gravel and mud deposits that have resulted from erosion of the upper elevation areas of the Ko'olau and 94 Waianae Mountains, from where both surface and sub-surface water for Pearl Harbor springs and streams originates. Groundwater from these mountain ranges flows down gradient in the Ko'olau basalt until coastal sediments near Pearl Harbor are encountered. The zone of springs is restricted to a narrow strip lying N w•• S 500--0 500 1000 Meters 21"22'OO"N J57'59'OO"W 1574 :56'30"W Figure 4.1. Map of Pearl Harbor with sampling locations. between the inland edge of marine sediments and the caprock at approximately 6.1 m above sea level (Visher and Mink 1964). Streams in the Pearl Harbor watershed are now eroding the cap rock that was left above sea level with artesian springs discharging from bedrock where the cap rock has been removed (Stearns and Vaksvik 1935). These springs are fed from breaks or low points in the caprock that allow escape of groundwater; this consequently results in a series of large freshwater releases (Stearns and Vaksvik 1935). Water also emerges seaward of the exposed basalt cliff, through the thin caprock of the Pearl Harbor 95 coastal plain, but in lesser quantities than at the base of the break in the Ko'olau basalt (Visher and Mink 1964). Flow is perennial only in the Ko'olau Mountain headwaters and near the mouth in the area of the basal Pearl Harbor springs, and has been reduced by approximately 50% from pre-development flows (Nichols et al. 1997). Above the areas of spring influence, Pearl Harbor streams are characterized by high flood peaks and low baseflows (Nichols et al. 1997). Streamflows are more constant downstream of the Pearl Harbor springs, and have characteristics of groundwater (low salinity, high silica, high nitrate levels) rather than surface water (Nance 1998). METHODS Nonindigenous aquatic species have been brought into Hawai'i both accidentally and intentionally. In many cases the method ofintroduction into O'ahu and geographic origin can be determined. Species of undetermined geographic origin are termed cryptogenic (Carlton 1996). The native or nonindigenous status of arthropods was ascertained from Nishida (1997), and for this study we assumed that organisms classified as probably endemic or indigenous were native species. Aquatic species introductions have been separated into the following categories: governmental biocontrol, intentional food introduction, probable ballast water or hull fouling, accidentally introduced with baitfish, aquarium release or with aquarium plants, brought in with airplanes, and unknown. Sampling began in October 1997 and ended in August 1998. Representative sampling stations were established in each major Pearl Harbor estuarine and coastal wetland and spring area (Figure 4.1) with sampling extending to areas just above the limit of tidal influence. Most sampling stations were generally at or just above sea level. Aquatic insect sampling was conducted according to Polhemus (1995a) and Englund et al. (2000b). Collections of both immature and adult specimens were made with aerial sweep nets, aquatic dip nets, seines, and benthic samplers. Bottom communities, including insects and taxa other than insects in the soft-sediment areas of streams were sampled with a Wildco Petite Ponar" 15.2 x 15.2 cm weighted dredge. Three dredge samples were collected at each stream mouth, and after collection sediments 96 were rinsed through a 1 x 1 mm sieve. The contents were preserved in 75% ethanol for laboratory analysis. Visual observations ofaquatic insects were also conducted above the waterbody. The sampling of damselflies and dragonflies (Odonata) was also emphasized as several endemic Hawaiian species are currently listed as candidate threatened or endangered species under the United States Endangered Species Act. Seine netting using a fine-mesh, 5 m long net was the main technique used to sample fish, and dip nets were also used to sample areas not accessible to seines. Experimental gill nets of varying sized mesh were also used to sample fish in areas that were too deep to seine. Efforts were also made to visually observe and collect native gobiid fish at each sampling site. Although some fish, crustacean, and mollusk species were identifiable in the field, many smaller specimens had to be preserved (in 75% ethanol) and taken to the laboratory for identification. The reference used for the scientific and common names offishes was from the American Fisheries Society (1991), crustaceans (American Fisheries Society 1989), and Nishida (1997) for insect names. Salinity was also recorded at least once at each location sampled. RESULTS A total 191 aquatic species were identified within the lower reaches ofPearl Harbor streams and wetlands (Table 4.1). A complete list of species found at each sampling site during this survey and in previous surveys can be found at http://hbs.bishopmuseum.org/lists/pearl-spp.html. Nonindigenous species dominated with 48% of the species recorded, while only 33% were native, and 19% cryptogenic (Figure 4.2). The one new species (cryptogenic) collected was an aquatic mite (Acari) in the family Ascidae. Arthropods (mainly insects) comprised nearly 61 % of the species collected and included 89 species ofaquatic insects, 26 97 Table 4.1. Summary of the native or nonindigenous status and total number (percent) of aquatic species found in Pearl Harbor estuarine habitats. Geographic All Aquatic l Aquatic Insects Fishes Crustaceans Mollusks Status Species Nonindigenous 91 (48%) 49 (55%) 18 (46%) 5 (19%) 10 (62%) Native 64 (33%) 22 (25%) 20(51%) 14 (54%) 3 (19%) Cryptogenic 35 (18%) 18 (20%) 1 (3%) 7 (27%) 3 (19%) New 1 «1%) 0 0 0 0 Tolal 191 89 39 26 16 I Miscellaneous species such as Annelida, Nematoda, and Cnidaria are included in this total. • Nonindigenous 0 Native • Cryptogenic 80 70 60 III .!!! I ~ 50 • • V) ~ 0 40 -. I I 1l • I • E 30 ::::J Z • • 20 • • - ~ 10 ~I~ ~I~ 0 ~ ---r- ~ ~ -r- ---r- E E E .J:; E ., .. E E E ~ ..E .. .. " .. .. e ...... ~ ~ !! ~ ~" § ~ i ~ E !! ! lI. % " j ~ 0< II> til Vi Vi .~ II> II> 5l .. Vi " rII> Vi Vi .. .0 :; g 0 :E ~ ~ ~ -'! w I: C" .. 1 .f! OJ" OJ" " <{ ! " ..!! li .l! .>< g .." .." ~" ~ -a a; ] :I: "~ "'iii ;;; i ~...... ~ :it ~ ~ ~.. " ~ '" '" >< ; .!!.. ~ Figure 4.2. Number of species by stream and native or nonindigenous status for combined aquatic fauna found in estuarine regions of Pearl Harbor. 98 crustaceans and 5 aquatic mites. Other phyla included vertebrates (22%), mollusks (8%) and annelid worms (7%). Other components of the fauna collectively composed only 2% of the species, and included Cnidaria, Platyhelminthes, Nematoda, and Sipuncula. Aquatic Insect Species Composition Twenty-two, or only 25%, of the aquatic insect species were known to be native. Nonindigenous species accounted for nearly 55% (49 species) of the aquatic insect species, and cryptogenic species comprised 20% (18). Of the native species of aquatic insects, approximately 59% (13) were endemic, while 41 % (9) were indigenous. Aquatic Diptera (flies) were by far the most species-rich order found and comprised 81 % (72) of all aquatic insect species. A large percentage of aquatic Diptera (28%,20 species) could not be identified to the species level, and thus were considered cryptogenic, while 46% (33 species) were nonindigenous. Odonata (dragonflies and damselflies) were the next most common (8% of species) followed by aquatic Heteroptera (true bugs, 6%) and Coleoptera (aquatic beetles, 6%). Nonindigenous caddisflies (Trichoptera) such as Cheumatopsyche pettiti were collected in lower Waikele Stream, and composed only 1% of the sampled aquatic insect fauna. Areas where the introduced fly Ephydra gracilis had previously been recorded in Pearl Harbor such as Hickam Field and Iroquois Point (Wirth 1947) were resampled. Intensive sampling of these and other suitable wetland areas found numerous species of other native and nonindigenous ephydrid flies, but did not find Ephydra gracilis. No native Megalagrion damselflies were found, while three species of nonindigenous damselflies, Ischnura posita, Ischnura ramburii, and Ena1lagma civile, were abundant. Two native and two nonindigenous species of dragonflies were also common; the indigenous species Pantala flavescens and Anaxjunius, were some of the most common native aquatic insects remaining in Pearl Harbor. Larvae and adults of A. junius were always collected from sites with many species of nonindigenous fish. The two nonindigenous dragonfly 99 species, Crocothemis servilia, which was first observed in O'ahu in 1994, and the well-established Orthemis ferruginea, were common throughout the surveyed area. The endemic marine water strider Halobates hawaiiensis, an aquatic heteropteran, were locally common. It was often found in the shelter of nonindigenous mangroves, and was always found in areas of water with> 34 ppt salinity. It is not known from Pearl Harbor prior to the introduction of mangroves. However, it was also common in areas without mangroves. Other aquatic Heteroptera found include four common, nonindigenous species in the families Corixidae (Trichocorixa reticulata), Mesoveliidae (Mesovelia amoena and Mesovelia mulsantl) and Saldidae (Micracanthia humilis). Four nonindigenous aquatic beetle species (Coleoptera) were found but no native species were collected. The most recent introduction is the small mangrove mudflat beetle, Parathroscinus cf. mmphyi, which was first recorded in Pearl Harbor in 1996 (Samuelson 1998). Parathroscinus cf. mmphyi populations have now exploded, and they were found in extremely high densities throughout Pearl Harbor mudflats, with flying adults forming thick clouds above the mud. Two other species of nonindigenous water scavenger beetles, Enochrus sayi and Tropistemus salsamentus, were common in areas of still water. Both of these species were saline tolerant, occurring in the lowest reaches of streams in areas with salinity as great as 16 ppt. Fish Species Composition Many species (nearly 44%) were Perciformes, including both native and nonindigenous species: gobies, cichlids (e.g., blackchin tilapia), blennies, and mullet. Other Orders were represented solely by nonindigenous species: Characiformes (pacu), Siluriformes (armored catfish), Cyprinodontiformes (poeciliids or mosquitofish), Cypriniformes (carp or koi), and Synbranchiformes (rice paddy eel). Larvae of Dussumieriinae (family Clupeidae) could not be determined as native or nonindigenous because of their small size. 100 A total of 39 fish species were collected with 51 % native, 46% nonindigenous, and 3% cryptogenic. Fish were found in salinities ranging from 0 to 37 ppt with fish species found in a wide range of salinities. Important exceptions were the two species ofnonindigenous South American armored catfish (Ancistrus temminckii and Hypostomus cf. watwata), which were restricted to freshwater; and three poeciliids, Poecilia reticulata (guppy), Xiphophorous helleri (green swordtail), and Xiphophorus maculatus (southern platyfish) that were restricted to waters of <3.0 ppt salinity. A general but not significant (chi-square test) trend for the percentage of native fish species to increase as salinities increased was observed, with areas of low salinity containing fewer native species (Figure 4.3). Areas of completely freshwater contained relatively few native species. Another introduced species, the fang-toothed blenny (Omobranchus ferox) was found in lower Halawa Stream in areas having salinities of 35 ppt, and only in approximately 15 m of rocky mangrove habitat. • Nonindigenou5 o Native 100% 90% 12 9 6 14 80% I/) Q) 70% .(3 ~ 60% (/) '0 50% -c: e 40% Q) a. 30% 20% 10% 0% 0-5 ppt 6-25 ppt 26-30 ppt > 30 ppt Figure 4.3. Native or nonindigenous status of fish species and total numbers found at different salinity levels in Pearl Harbor estuaries. 101 The large (> 30 ha) spring complex including Kalauao and Waiawa Springs had low salinity levels (1 to 4 ppt) and was almost entirely dominated by high densities of nonindigenous fish such as blackchin tilapia (Sarotherodon melanotheron) and livebearers (Gambusia affinis, Poecilia latipinna, Poecilia mexicana, etc.). Along with Waimano-Waiau Springs, the Waikele Springs area was completely freshwater (0 ppt salinity), and no native fish species were observed in these areas. The nonindigenous goby Mugilogobius cavifrons was abundant at most sampling stations. The native goby A waous guamensis was observed or collected in low numbers at only two sampling areas: Waikele and Waimalu Streams. Other native gobiid estuarine fish found include Oxyurichthys lonchotus, and Stenogobius hawaiiensis. Eleotris sandwicensis was the most common native stream fish, occurring widely in a variety of habitats. The endemic native gobiid Stenogobius hawaiiensis was less common and was found in only 6 of the 15 sampling stations. Blackchin tilapia (Sarotherodon melanotheron) was the dominant inshore fish, occurring in high densities at every sampling location. In many enclosed wetland areas, such as at the Pearl Harbor National Wildlife Refuge, blackchin tilapia appeared stunted (only 7 to 10 cm in length but in breeding colors). In comparison, tilapia were generally larger in areas where they had direct access to Pearl Harbor such as in Halawa Stream, and adults ranged in size from 20 to 30 cm. Crustacean Species Composition Twenty-six taxa were distinguished; all were identified to order but some could not be identified to species. Of those identified to species level, 54% (14 species) were native, 19% (5) nonindigenous, and 27% (7) cryptogenic. Nearly 60% were decapods, while isopods (15%) and amphipods (11 %) comprised the next most abundant taxa. Less species-rich orders included mysids and copepods, although large numbers of individuals of these orders were found in some sampling areas. The number of crustacean species found was highest in Halawa Stream and lowest in Pouhala Marsh. 102 Native estuarine decapods that were abundant and found at most sampling sites included Periclimenes cf. grandis, Pa1aemon debilis, and Tha1amita crenata; Macrobrachium grandimanus was relatively common but was restricted to more freshwater habitats. Two nonindigenous species, Macrobrachium 1ar and Procambarus clarkii, were also common and found exclusively in freshwater. Two isopod crustaceans were identified to the species level, the endemic Ligia hawaiiensis, a common marine shoreline species, and Porcellio 1aevis, a widespread nonindigenous species. A nonindigenous freshwater shrimp was identified as Neocaridina denticu1ata sinensis, a subspecies previously known only from the Chinese mainland and Taiwan; it was abundant in lower Waikele Stream in 1998 but absent from the same location in 1993 (Englund and Cai 1999). Mollusk Species Composition Sixteen species of mollusks were collected but these included no native freshwater or estuarine species. Three (19%) of the 16 were native marine mollusks often found in estuarine regions; 10 (63%) were nonindigenous freshwater/brackish species and three (19%) were cryptogenic species. Two ofthe cryptogenic species could only be identified to the family level (Terebridae and Thiaridae), while an undetermined bivalve was the third cryptogenic species. The number of mollusk species found at each sampling site ranged from one species to a high of eight species collected at Halawa Stream. Nonindigenous species predominated within all stream and estuary areas, with the exception of two common native marine species (Cerithium nesioticum and Ceritihium cf. zebrum) found in the lowest reaches of Halawa Stream. Two especially significant species of nonindigenous snails were found. Apple snails (Pomacea canalicu1ata) were recorded for the first time in a Pearl Harbor drainage (Lach and Cowie 1999). High densities of apple snails were also observed in taro fields in a spring within 25 m of lower Waikele Stream; however, they were not observed in the stream channel. The presence of apple snails so close to the stream (and in the floodplain) means that it is highly likely they will soon be in Waikele Stream itself. A new Pacific Ocean record was also established for a species of a 103 hydrobiid snail, Pyrgophorus cf. coronatus, found in Pouhala Marsh and Waiawa Springs. The genus Pyrgophorus originates in the Caribbean region, and these snails are found in fresh-to-brackish-marine waters in streams and wetlands in their native regions (Cowie 1999). Pyrgophorus cf. coronatus was found in water ranging from 1 to 9 ppt salinity and always on a silty mud bottom. Densities of this newly introduced species were high, with numerous individuals incidentally captured in a single fish haul seine. Miscellaneous Species Three species ofannelids, one species of cnidarian, one species of nematode, and one species of platyhelminthes were also collected. Most of these species were collected from sediment samples taken with an Ekman dredge, while leeches (Hirudinea), aquatic earthworms (Oligochaetes), and flatworms (Platyhelminthes) were found during general collections. Nematodes were also found inside fish guts. Because of the cosmopolitan nature of many of these sediment dwelling species, with two exceptions (Eldredge and Miller 1997), their geographic status is uncertain. Myzobdella lugubris is known to be nonindigenous; it was restricted to freshwater and was commonly observed attached to both native and nonindigenous fish species. The indigenous leech Aestabdella abditovesiculata was also common, mainly on marine fish. DISCUSSION Origins and Modes ofIntroductions of Nonindigenous Species Found in Pearl Harbor Estuarine Areas Invasions in a number of other estuarine areas of the world have been examined, for example, San Francisco Bay (Cohen and Carlton 1995; 1998), Chesapeake Bay (Smith et a1. 1999), and the Baltic Sea (Olenin and Leppakoski 1999). High percentages of nonindigenous species were found in the San Francisco Bay area (Cohen and Carlton 1995) similar to the findings of this study in the estuarine regions of Pearl Harbor. However, most studies on the biodiversity of estuarine invasions have been in cool temperate regions, not tropical waters as in the present study. Furthermore, the present study is a faunal wide assessment that included aquatic insects, a major component of the estuarine biota that was not assessed in other estuarine 104 invasion studies. In the present study, 89 species of aquatic insects were recorded (native and introduced), whereas insect invasions associated with or proximal to estuarine waters were not examined in other studies (e.g., San Francisco Bay or the Hudson River (Mills et a1., 1996) (J. T. Carlton, pers. comm, September 2000), further limiting comparisons. This study found that nonindigenous species comprise the dominant portion of the biota, and the lower portions of Pearl Harbor streams, springs, and wetlands are now dominated by nonindigenous species. The probable origins and mode of introductions of these established nonindigenous species are shown in Table 4.2, and only organisms identified to the species level were included in this table (76 species). Determining the mode of transport of aquatic insects into Hawai'i is often difficult because of the inconspicuous nature of both the insects and how they are introduced. The Pearl Harbor area has grown from a small, shallow harbor in the early 20th century to a large military port with adjacent civilian and military airports, with traffic arriving from all over the world (Coles et a1. 1997). No geographic region is a predominant source of aquatic species introductions into the Pearl Harbor area (Table 4.2), although more species come from the Americas than other areas. For example, 57% (43 species) originated from the Americas, 30% (23) from Asia and the Pacific, 5% (4) from AustraliafNew Zealand, 5% (4) have a world-wide distribution, and fewer than 3% (2) of species originated from Africa. It is not surprising that these introductions come from a wide range of areas, and illustrates with modern transportation how easily nonindigenous organisms become established in vulnerable insular tropical island environments. The findings of this study further illustrate that Pearl Harbor is the "crossroads of the Pacific Ocean" for nonindigenous species introductions Coles et ai. (1997). Post-Introduction Spread of Nonindigenous Species in Lower Pearl Harbor Watersheds It is highly likely that once an organism is introduced into a stream or adjacent wetland on O'ahu it will spread to other aquatic habitats throughout the island, and potentially to the other Hawaiian Islands. The 105 nonindigenous goby Mugilogobius cavifrons was first observed in 1987 in Pearl Harbor (Randall et a1. 1993) and is now common in estuarine areas throughout windward and leeward O'ahu (Englund et al. 106 Table 4.2. Geographic source (year of introduction) and known (or probably known) mode of introduction of nonindigenous species of aquatic macrofauna found in Pearl Harbor streams and estuaries. Native Region (Year first released or found in Mode of Introduction3 Taxa Hawaii) Aquatic Insects Coleoptera-Hydrophilidae Enochrus sayi Eastern North America (1931) Airplane Tropistemus lateralis humeralis North America (Pacific Coast) Aquarium Release or with Aquarium Plants (1948) Tropistemus salsamentus North America (California) (1968) Ballast Water Coleoptera-Liminchidae Parathroscinus cf. murphyi Southeast Asia (1996) Ballast Water Diptera- Canacidae Canaceiodes angulatus N. and S. America- West Coasts Cryptogenic (Unknown) (1922) Procanace wi1liamsi Oriental region? (1944) Airplane Diptera-Ceratopongonidae Atrichopogon jacobsoni Oriental and Pacific Regions (1958) Airplane Diptera-Chironomidae Chironomus crassiforceps Oriental and Pacific Regions (1944) Airplane Cricotopus bicinctus Holarctic (1955) Aquarium Release or with Aquarium Plants Goeldichironomus holoprasinus N. and S. America (1969) Cryptogenic (Unknown) Diptera-Culicidae Aedes albopictus Mexico (1826) Ballast Water Diptera-Dolichopodidae Chrysotus longipalpus West Indies (1930) Cryptogenic (Unknown) Condylostylus longicomis Neotropics to French Polynesia Cryptogenic (Unknown) (1996) Pelastoneurus lugubris North America (1994) Cryptogenic (Unknown) Syntonnon flexible Taiwan, Australia (1917) Cryptogenic (Unknown) Tachytrechus angustipennis N. and S. America (1993) Cryptogenic (Unknown) Thinophilus hardyi Australasia (1996) Cryptogenic (Unknown) Diptera-Empididae Hemerodromia stellaris Southwest United States (1982) Cryptogenic (Unknown) Diptera-Ephdridae Brachydeutera ibari Oriental region? (1980) Cryptogenic (Unknown) Ceropsilopa coquilletti Nearctic (1946) Cryptogenic (Unknown) Clasiopella uncinata West Indies? (1946) Airplane Discocerina mera Pacific (1948) Cryptogenic (Unknown) Donaceus nigronotatus Oriental Region (1958) Airplane Ephydra milbrae North America (west coast) (1950) Cryptogenic (Unknown) Hecamede granifera Pacific Region (1923) Cryptogenic (Unknown) Hydrellia willamsi Australia/New Zealand (1931) Aquarium Release or with Aquarium Plants Lytogaster gravida North America (1937) Cryptogenic (Unknown) Mosi11us tibialis North America (1944) Cryptogenic (Unknown) Ochthera circularis Oriental and E. Palaearctic (1982) Cryptogenic (Unknown) Paratissa pollinosa Neotropics (1945) Cryptogenic (Unknown) PlacopsideIla marquesana Pacific (1951) Cryptogenic (Unknown) Psilopa girschneri Holarctic (1952) Cryptogenic (Unknown) ScateIla stagnalis Holarctic (1946) Cryptogenic (Unknown) 107 Table 4.2 (cont.). Geographic source (year of introduction) and known (or probably known) mode of introduction of nonindigenous species of aquatic macrofauna found in Pearl Harbor streams and estuaries. Native Region (Year first released or found in Mode of Introduction3 Taxa Hawaii) Diptera-Tethinidae Tethina variseta North America (1946) Cryptogenic (Unknown) Diptera-Tipulidae Stryringomyia didyma Australasia (1896) Cryptogenic (Unknown) Symplecta piIipes Tropical Widespread species (1892) Cryptogenic (Unknown) Heteroptera-Corixidae Trichocorixa reticulata N. and S. America (1878) Cryptogenic (Unknown) Heteroptera-Mesovellidae Mesovelia amoena N. and S. America (1971) Cryptogenic (Unknown) MesoveIia mulsanti N. and S. America (1933) Aquarium Release or with Aquarium Plants Heteroptera-Saldidae Micranthia humiIis N. America (1988) Aquarium Release or with Aquarium Plants Odonata-Coenagrionidae Enallagma civile Western North America (1936) Aquarium Release or with Aquarium Plants Ischnura posita N. and Central America (1936) Aquarium Release or with Aquarium Plants Ischnura ramburii North America (1973) Airplane Odonata-Libellulidae Crocothemis serviIia Middle East, Asia to Australia Aquarium Release or with Aquarium Plants (1994) Orthemis ferruginea South America to Florida (1977) Aquarium Release or with Aquarium Plants Trichoptera-Hydropsychidae Cheumatopsyche anaIis Western North America (1965) Aquarium Release or with Aquarium Plants Crustaceans Macrobrachium lar Guamrrahiti (1957) IntentionlliFoodIntroduction Neocaridina denticulata sinensis China-Taiwan (1991) Aquarium Release or with Aquarium Plants Panopeus lacustris Northwest Atlantic (1947) Ship fouling (hull or seachest) Panopeus pacificus Philippines (1929) Ship fouling (hull or seachest) Procambarus clarkii North America (1923) Intentional Food Introduction Mollusks Cipangopaludina chinensis Southeast Asia (1900) Aquarium Release or with Aquarium Plants Corbicula flurninea Asia (1981) Intentional Food Introduction Planorbella duryi North America (1994) Aquarium Release or with Aquarium Plants Pomacea canilliculata South America (1989) Intentional Food Introduction Fishes Ancistrus cf. temminckii South America (1985) Aquarium Release or with Aquarium Plants Clarias fuscus Asia «1900) Intentional Food Introduction Colossoma macropomum1 South America (1987) Aquarium Release or with Aquarium Plants Cyprinus carpio Asia «1900) Intentional Food Introduction Gambusia affinis Texas (1905) Intentional Biocontrol Hemichromis elongatus Africa (1991) Aquarium Release or with Aquarium Plants Hypostomus cf. watwata South America (1984) Aquarium Release or with Aquarium Plants Limia vittata Cuba (1950) Aquarium Release or with Aquarium Plants Monopterus albus Asia « 1905) Intentional Food Introduction Moolgarda engeIi Marquesas (1955) Accidentlli with Baitfish Mugilogobius cavifrons Western Pacific (Japan to Ballast Water Indonesia) (1987) Omobranchus ferox South Pacific (Philippines to Ballast Water Madagascar) (1998) 108 Table 4.2 (cont.). Geographic source (year of introduction) and known (or probably known) mode of introduction of nonindigenous species of aquatic macrofauna found in Pearl Harbor streams and estuaries. Native Region (Year first released or found in Mode of Introduction3 Taxa Hawaii) Poecilia latipinna Texas (1905) Intentional Biocontrol Poecilia mexican!? North America(1940-1950) Aquarium Release or with Aquarium Plants Poecilia reticulata South America (1922) Intentional Biocontrol Tilapia (Sarotherodon) Africa (1951) Intentional Biocontrol melanotheron Xiphophorus helleri Central America (1922) Intentional Biocontrol Xiphophorus maculatus Central America (1922) Intentional Biocontrol Amphibians Bufo marinus South America (1932) Intentional Biocontrol Rana catesbeiana North America (1902) Intentional Biocontrol Rana rugosa Japan (1896) Intentional Biocontrol Summary of Introduction Pathways Percent (number) Aquarium Release or with 25% (19) Aquarium Plants Intentional Biocontrol 12% (9) Ballast Water 5% (4) Ship Fouling (hull or seachest) 3% (2) Intentional Food Introduction 9% (7) Accidental with Baitfish 1% (1) Airplane 11 % (8) Cryptogenic (Unknown) 34% (26) lBreeding populations of Colossoma macropomumnot yet known to be established, only large live adults observed. 2Poecilia mexicana possibly hybridized before introduced to Hawaii, however, the source of these fish is thought to be from Mexico or the southern U.S. Randall (1987) believed they were released before 1950, and this probably occurred after 1930's surveys conducted by G.B. Mainland (1939). \References: Van Dine 1907, Brock 1960, Edmondson 1962, Randall 1987, Evenhuis 1989, Devick 1991, Cowie 1997, Englund et al. 2000a, Yamamoto and Tagawa 2000). 2000a). This small « 50 mm) estuarine goby species is cryptically colored and is not found in the aquarium trade, nor used as a potential food source. It is likely only a matter of time before M. cavifrons spreads to the other Hawaiian Islands. Aquarium observations ofM. cavifrons also indicate this species is carnivorous and will consume most prey items smaller than itself. The impacts of predation by M. cavifrons on native biota are unknown, but the fact that it will prey on native species is cause for concern, especially as it was much more abundant than any native stream goby. Also, a nonindigenous fang-toothed blenny (Omobranchus ferox) first detected in the Hawaiian Islands during this study in 1998 had by 2000 spread 24 km away from Pearl Harbor (Yamamoto and Tagawa 2000). Sampling conditions in Pearl Harbor were sometimes difficult, and poor water visibility and deep silt sometimes hampered fish collection efforts in many areas, and partially saline conditions prevented the effective use of electrofishing. However, similar 109 to the native goby species, Mugilogobius cavifrons is cryptic and benthic in nature, and our routine capture of this species but not A waous guamensis indicates gear limitations may not have been the reason that native A waous guamensis were so uncommon during this study. The introduced dragonfly Crocothemis servilia was first observed on O'ahu in 1994 (Polhemus 1995b) and by 1999 had spread to the island of Kaua'i (Englund, personal observation). Nonindigenous aquatic insects fIrst sighted in Pearl Harbor before spreading to other areas in Hawai'i include the non-biting midge Chironomus crassiforceps first recorded in 1944 (Van Zwaluwenburg 1945), Cricotopus bicinctus first recorded in 1955 (Hardy 1956), and the ephydrid flies Ephydra gracilis and Clasiopella uncinata, both first recorded in 1946 (Wirth 1947; Adachi 1952). Cricotopus bicinctus is one ofthe most widespread aquatic insect species in low-elevation areas in the Hawaiian Islands and now comprises a substantial portion of invertebrate drift in Hawaiian streams (Englund et al. 2000b). New Invasions Detected in this Study An average of four species of terrestrial and aquatic snails became established, per decade, in Hawai'i during the twentieth century (Cowie 1998). At least 58 species of freshwater organisms (excluding aquatic insects) became established in the Hawaiian Islands from 1900 to 1991 (Devick 1991), i.e. nearly 6.4 species per decade. The present study suggests that these trends are continuing. For example, three new nonindigenous species colonized the lower Waikele Stream estuary area between 1993 and 1998: a dragonfly (Crocothemis servilia), an atyid shrimp (Neocaridina denticulata sinensis), and the apple snail (Pomacea canaliculata). At least two of these species, the atyid shrimp and apple snail could have colonized Pearl Harbor watersheds through human-mediated actions, or perhaps through accidental transport by birds or flooding (Lach et al. 2000). A species of African cichlid (Hemichromis elongatus), previously unknown in Pearl Harbor watersheds was common in the lower portions of Waiawa Stream. A new and potentially harmful species of fish was found 110 during these surveys. Native to the Philippines and South China Sea region, the fang-toothed blenny (Omobranchus ferox) appears to have recently become established in Pearl Harbor. Our collection of a wide range of size classes indicates that it is successfully reproducing. In its native habitat in the Philippines, Omobranchus ferox inhabits a wide range of shallow estuarine and freshwater habitats, ranging from mangrove swamps to rivers and freshwater lakes (Springer and Gomon 1975). It thus represents a potential threat not only to the ecologically similar indigenous Oxyurichthys lonchotus but also to other native freshwater and estuarine fish and invertebrates. The freshwater shrimp Neocaridina denticulata sinensis was found in high densities in the Waikele Springs area. The finding of N d. sinensis on O'ahu is also the first Pacific Island record for this species. Unlike the native freshwater atyid shrimp Atyoida bisulcata, N d. sinensis does not have an obligate marine phase (Hung et a1. 1993) and is restricted to freshwater. Neocaridina denticulata sinensis could have spread into separate watersheds by repeated human introductions, through flooding, or perhaps through other agents such as birds. It is possible that N d. sinensis will compete for food and space with Atyoida bisulcata, as they occupy similar habitats and have overlapping elevational distributions (Englund and Cai 1999). Its native range includes Japan, Taiwan, the Ryukyu Islands, Korea, mainland China, and Vietnam (Hung et a1. 1993). Another new Pacific Island record was established for a species of hydrobiid snail, Pyrgophorus cf. coronatus, that was found in several estuarine wetland areas. This species was found in high densities in wetland habitats, and its impacts on native species are unknown. The Distribution ofNative versus Introduced Taxa In high elevation areas, Pearl Harbor streams still contain significant reservoirs of native aquatic species, in contrast to the dominance of nonindigenous species found during the present study ofestuarine areas (Figure 4.4). That high-elevation reaches of Pearl Harbor watersheds are still relatively unimpacted by introduced aquatic species may be attributed to a combination of factors including the lack of urbanization in upper 111 elevation areas, the lack of water diversions, or that invasive species may be more adapted to disturbed conditions (Courtenay 1997) or may be more generalist in nature (e.g., Lach et al. 2000). o Native. Nonindigenous • Cryptogenic 110 .,------, 100 90 fJ) ~ 80 Q) [Ji 70 2c: 60 ~ 50 ~ 40 30 20 10 o HalllWll Hal_ WlIike...... ikete (0-1 m) ""'ik."(0-1 m) (290-315 m) (370-390m) (480-490 m) Figure 4.4. Native or nonindigenous status of aquatic insects at varying elevations on O'ahu: upper Halawa data from Polhemus (1994), upper Waikele data from Englund (1993). The difference in the abundance of native species of aquatic insects between lower Waikele Stream (10%) and Halawa Stream (38%) is of interest and may result from differences in salinity between the two sites. Waikele Stream empties a large amount of water, whereas lower Halawa Stream is generally dry or contains a minimal amount of waterflow. Halawa Stream is more marine in character (30 to 37 ppt salinity), and is often intermittent in its lower reaches because of stream diversions. This is in contrast to a large freshwater plume of <15 ppt salinity in Waikele Stream that extended well out into Pearl Harbor (Nance 1998). The low salinity conditions in Waikele Stream may better foster biological invasions. This could be because human-mediated introductions of freshwater organisms appear more likely to occur than marine ones. For example, freshwater introductions, such as apple snails, the shrimp Neocaridina denticulata sinensis, or 112 aquatic insects coming in with aquarium plants, occur on a regular basis. It also may be that marine environments are somewhat more resistant to invasions than freshwater environments, but the human factor of introducing freshwater species for sport, through aquarium escapes, and for additional food sources is more likely to be the case than with marine species. Although the freshwater areas surveyed in this study contained relatively fewer native species than the more marine areas (Figure 4.3), this difference was not statistically significant. The native or nonindigenous status of many species of Hawaiian aquatic insect species in disturbed lowland habitats such as Pearl Harbor is not yet known. For example, 20 species of aquatic Diptera (flies) could not be identified to the species level, thus rendering determination of geographic origin impossible. The native Hawaiian insect fauna in accessible lowland areas has been relatively well studied since the 1880s, starting with early collectors such as Blackburn and Perkins (Liebherr and Polhemus 1997; Englund 1999), and it would seem reasonable to assume that most native aquatic insect species in these lower elevation areas of Pearl Harbor have been described. Ifit is assumed that most, if not all, cryptogenic aquatic insect species found in Pearl Harbor are new introductions, then 75% of the aquatic insects found in Pearl Harbor estuarine regions are nonindigenous species. Additionally, extensive surveys in upper elevation areas of the Hawaiian Islands in the 1990s have generally yielded either known native and nonindigenous species, or undescribed native species of aquatic insects (Evenhuis and Polhemus 1994; Evenhuis 1997b, Evenhuis 2000). For example, only one cryptogenic aquatic insect was found during extensive surveys of the nearly pristine stream systems in the upper Alaka'i plateau on Kaua'i between 1997 to 1999 (Englund et a1. 2000b). Ecological Impacts of Invasions The ecological impacts ofinvasions into Hawaiian estuarine and freshwater appear to be easily observable in the field, for example, recovering only introduced blackchin tilapia and other introduced fish when seining in a wetland. However, with the exception of the documentation of the extinction of certain taxa from these habitats (Polhemus 1996; Englund 1999) no ecological data exist documenting these impacts; additionally, 113 the lack (or extirpation) of native species in these areas also currently makes establishing cause and effect difficult. Recent food introductions, such as the apple snail and Asiatic clam (Corbicula fluminea) (Eldredge 1994), are now causing great ecological and economic damage. Apple snails (Pomacea canaliculata), first introduced illegally as a food source in 1989, are now a major pest threatening the cultivation of an important Hawaiian staple food, taro (Cowie, in press). The Asiatic clam has caused enormous economic losses on the u.s. mainland similar to those caused by zebra mussel infestation of the Great Lakes (U.S. Congress 1993). In Hawai'i, Asiatic clams have clogged irrigation pipes with resulting economic damage in Maui and elsewhere (Devick 1991). The ecological impacts of these species on native aquatic biota are unknown. Blackchin tilapia (Sarotherodon melanotheron) were first introduced in 1951 for aquatic weed control and as baitfish, and may impact the abundance and distribution ofnative Hawaiian waterbirds by leaving little invertebrate forage for them, although no studies exist demonstrating these impacts. Aquarium trade and subsistence food introductions are some of the most serious threats facing native aquatic species in Hawai'i, with ballast water and hull fouling representing another potential introduction pathway, although only 6 (8%) of the 76 species listed in Table 4.2 are considered likely to have come from the latter sources. Since 1987, for example, a goby (Mugilogobius cavifrons), the fang-toothed blenny (Omobranchus ferox), and the small mangrove mudflat beetle (Parathroscinus ef. murphyl) were all first observed or collected in Pearl Harbor. None of these small species are food introductions, and all of these species are also unlikely aquarium introductions. Evidence For Extirpation or Reduction of Native Biota Poecilia latipinna, Fundulus grandis, and Gambusia affinis were the first recorded introductions of nonindigenous species into Pearl Harbor waters (Van Dine 1907), although unrecorded aquatic species introductions undoubtedly occurred earlier. Native damselflies (Megalagrion spp.) were formerly common in the Pearl Harbor area (Polhemus 1996; Liebherr and Polhemus 1997) but are now absent. Poeciliid fish 114 may be a major cause of the extinction of Megalagrion damselflies in low-elevation areas of Hawaiian streams and wetlands (Englund 1999). The well-documented loss in Pearl Harbor of major taxa such as the native damselflies suggests that much has changed because of nonindigenous species introductions. Although habitats have been changed because of urbanization in this watershed, large amounts of brackish to freshwater spring and wetland habitats still remain yet are dominated by alien species. With the exception of a few taxa such as the native damselflies, little information can be found on conditions in Pearl Harbor prior to the massive influx of nonindigenous species into the freshwater to estuarine areas that started in the early part of the 20th century. Other native groups that have apparently been lost from Pearl Harbor but are not as well documented include the water bugs in the family Saldidae, aquatic beetles (Coleoptera), and freshwater mollusks; none were found during this study. For example, Saldidae are some of the most common native aquatic insects in the Hawaiian Islands, with as many as three native species found in a single stream (Polhemus 1995a). However, in the lower areas of Pearl Harbor streams and wetlands only the nonindigenous saldid Micracanthia humilis was found. Four species of introduced aquatic beetles were found during this study, but none of the 9 native species (Nishida 1997) were collected. The decline of native species in Pearl Harbor will likely continue as more introductions occur. The decline will also be influenced by the environmental degradation that has occurred, which provides more favorable habitat for invading aquatic species than for native species. The low percentage of native aquatic insects, the absence of native freshwater mollusks, and the scarcity of native fish in the lower stream regions are evidence of this decline. There are no comparable studies of other large Hawaiian estuarine systems, so it is not possible to ascertain whether this is an archipelago wide trend, or a phenomenon restricted to Pearl Harbor. However, it is likely that most Hawaiian estuaries have experienced similar alterations in the composition of native aquatic fauna, as most of the introduced species found in this study are vagile, and as 115 discussed earlier, many species first recorded in Pearl Harbor have now spread to other areas ofO'ahu and throughout the Hawaiian Islands. 116 REFERENCES Adachi, M. S. 1952. Notes. Proceedings ofthe Hawaiian Entomological Society 14: 353-354. American Fisheries Society. 1989. Common and scientific names of aquatic invertebrates from the United States and Canada: decapod crustaceans. American Fisheries Society Special Publication 17. Bethesda. American Fisheries Society. 1991. Common and scientific names of fishes from the United States and Canada, Fifth Edition. American Fisheries Society Special Publication 20. Bethesda. Brock, V. E. 1960. The introduction of aquatic animals into Hawaiian waters. Intemationale Revue der gesamten und Hydrographie 45: 463-480. Carlton, 1. T. 1996. Biological invasions and cryptogenic species. Ecology 77: 1653-1655. Cohen, A. N. and 1. T. Carlton. 1995. Nonindigenous aquatic species in a United States estuary: a case study of the biological invasions of the San Francisco Bay and Delta. U.S. Fish and Wildlife Service, Washington D.C. Cohen, A. N. and J. T. Carlton. 1998. Accelerating invasion rate in a highly invaded estuary. Science 279: 555-557. Coles S. L., R. C. DeFelice, L. G. Eldredge and J. T. Carlton. 1997. Biodiversity of marine communities in Pearl Harbor, Oahu, Hawaii with observations on introduced exotic species. Bishop Museum Technical Report 10. 117 Coles, S. L., R. C. DeFelice, L. G. Eldredge and I. T. Carlton. 1999. Historical and recent introduction of nonindigenous marine species into Pearl Harbor, Oahu, Hawaiian Islands. Marine Biology 135: 147-158. Courtenay, W. R. If. 1997. Nonindigenous Fishes. In: D. Simberloff, D. C. Schmitz, and T. C. Brown (eds.), Strangers in Paradise: Impact and Management of Nonindigenous Species in Florida. Island Press, p. 109-122. Washington D. C. Cowie, R. H. 1997. Catalog and bibliography of the nonindigenous nonmarine snails and slugs of the Hawaiian Islands. Bishop Museum Occasional Papers 50: 66 pp. Cowie, R. H. 1998. Patterns of introduction of non-indigenous non-marine snails and slugs in the Hawaiian Islands. Biodiversity and Conservation 7: 349-368. Cowie, R. H. 1999. New records of alien nonmarine mollusks in the Hawaiian Islands. B Bishop Museum Occasional Papers 59: 48-50. Cowie, R. H. In Press. Apple snails as agricultural pests: their biology, impacts and management. In: Molluscs as Crop Pests (ed. G.M. Barker). CABI, London. Devick, W. S. 1991. Patterns of introductions of aquatic organisms to Hawaiian freshwater habitats. In: New directions in research, management, and conservation of Hawaiian freshwater stream ecosystems, Proceedings ofthe 1990 Symposium on Freshwater Stream Biology and Management, State of Hawaii, pp. 189-213. Hawaii Division of Aquatic Resources, Honolulu. 118 Ducks Unlimited, Inc. 1998. Environmental and enhancement plan for Pouhala Marsh, Oahu, Hawaii. Prepared for: Hawaii Division ofForestry and Wildlife, U.S. Fish and Wildlife Service, City and County of Honolulu. Edmondson, C. H. 1962. Xanthidae of Hawaii. Bishop Museum Occasional Papers 22: 215-309. Eldredge, L. G. 1994. Perspectives in aquatic exotic species management in the Pacific Islands: Volume I, Introductions of commercially significant aquatic organisms to the Pacific Islands. South Pacific Commission, Noumea, New Caledonia. Eldredge, L. G. and S. E. Miller. 1997. Numbers of Hawaiian species: supplement 2, including a review of freshwater invertebrates. Bishop Museum Occasional Papers 48: 3-22 Englund, R. A. 1993. A survey of the fish and aquatic insect fauna of the Waikele/Kipapa streams, Oahu, Hawaii. BHP Environmental Technologies. Consultants report prepared for Halekua Development Corp., Honolulu. Englund, R. A. 1999. The impacts of introduced poeciliid fish and Odonata on endemic Megalagrion (Odonata) damselflies on Oahu Island, Hawaii. Joumal ofInsect Conservation 3: 225-243. Englund, R. A and Y. Cai. 1999. The occurrence and description of Neocaridina denticulata sinensis (Kemp, 1918) (Crustacea: Decapoda: Atyidae), a new introduction to the Hawaiian Islands. Bishop Museum Technical Report 58: 58-65. Englund, R. A. and D. 1. Preston. 1999. Biological Assessment of the Lower Hamakua Ditch on the Hawaiian Stream Fly (Sigmatineurum meaohl) and other aquatic insects. Hawaii Biological 119 Survey, Bishop Museum report for USDA The Natural Resources Conservation Service, 300 Ala Moana Boulevard, Honolulu, Hawaii. Englund, R. A, K. Arakaki, D. 1. Preston, S. L. Coles and L. G. Eldredge. 2000a. Nonindigenous freshwater and estuarine species introductions and their potential to affect sportfishing in the lower stream and estuarine regions of the south and west shores of Oahu, Hawaii. Bishop Museum Technical Report 17. Englund, R. A, D. A Polhemus and D. 1. Preston. 2000b. Assessment of the impacts of rainbow trout predation on native aquatic invertebrate species within Kokee State Park streams, Kauai, Hawaii. Bishop Museum Technical Report 18. Evenhuis, N.L. 1989. Catalog of the diptera of the Australasian and Oceanian regions. Bishop Museum Special Publication 86. Bishop Museum Press and EJ. Brill. 1155 pp. Evenhuis, N. L. 1997. The genus Sigmatineurum Parent in Hawaii (Diptera: Dolichopodidae), with a revised key to species. Bishop Museum Occasional Papers 47: 66-73. Evenhuis, N. L. 2000. A new species of Sigmatineurum (Diptera: Dolichopodidae) from seeps in Waimanu Valley on the Big Island of Hawaii. Bishop Museum Occasional Papers 63: 31-33. Evenhuis, N. L. and D. A Polhemus. 1994. Review of the endemic Hawaiian genus Sigmatineurum Parent (Diptera: Dolichopodidae). Bishop Museum Occasional Papers 37: 1-19. Font, W. F. and D. C. Tate. 1994. Helminth parasites of native Hawaiian freshwater fishes: an example of extreme ecological isolation. Journal ofParasitology 80: 682-688. 120 Hardy, D. E. 1956. Notes. Proceedings ofthe Hawaiian Entomological Society 16: 15. Hung, M. S., T. Y. Chan, and H. S. Yu. 1993. Atyid shrimps (Decapoda: Caridea) of Taiwan, with descriptions of three new species. Journal ofCrustacean Biology 13: 481-503. Lach, L., D. K. Britton, R. J. Rundell, and R. H. Cowie. 2000. Food preference and reproductive plasticity in an invasive freshwater snail. Biological Invasions 2: 279-288. Lach, L. and R. H. Cowie. 1999. The spread of Pomacea canaliculata (Lamarck) (Gastropoda: Ampullariidae), an introduced freshwater apple snail, on Oahu, Hawaii. Bishop Mus. Occas. Pap. 58:66-72. Liebherr, J. and D. A. Polhemus. 1997. Comparisons to the century before: the legacy of the R.C.L. Perkins and Fauna Hawaiiensis as the basis for a long-term ecological monitoring program. Pacific Science 51: 490-504. Maciolek, J. A. 1984. Exotic fishes in Hawaii and other islands of Oceania. In: Distribution, Biology, and Management of Exotic Fishes, Ed. By W.R. Courtenay, Jr. and J.R. Stauffer, Jr., John Hopkins University Press, Baltimore and London, pp. 131-161. Meyers, G. S. 1949. Usage of anadromous, catadromous and allied terms for migratory fishes. Copeia 1949:87-88. Mills, E. L., D. L. Strayer, M. D. Scheuerell, and J. T. Carlton. 1996. Exotic species in the Hudson River basin: a history of invasions and introductions. Estuaries 19:814-823. 121 Nance, T. 1998. Effect of the proposed use of the WP-18 pump station on Waikele Stream and West Loch. Consultants report prepared by Tom Nance Water Resource Engineering, 680 Ala Moana Boulevard, Suite 406. Nichols, W. D., P. J. Shade and C. D. Hunt. 1997. Summary of the Oahu, Hawaii, regional aquifer-system analysis. U.S. Geological Survey Professional Paper 1412-A. Nishida, G. M. 1997. Hawaiian terrestrial arthropod checklist, 3rd Edition (Searchable database on the internet at http://www.bishopmuseum.org/bishop/ento/entodbhome.html). Hawaii Biological Survey. Bishop Museum Technical Report 12. 263 pp. Olenin, S. and E. Leppakoski. 1999. Non-native animals in the Baltic Sea: alteration of benthic habitats in coastal inlets and lagoons. Hydrobio1ogia 393: 233-243. Polhemus, D. A. 1994. The aquatic insect fauna of North Halawa, Oahu, Hawaii. In F.G. Howarth et aI., A natural history survey of North Halawa Valley, Oahu, Hawaii. Bishop Museum report to Hawaii State Department ofTransportation, H-3 Project. pp. 75-85. Bishop Museum, Hawaii Biological Survey. Honolulu, Hawaii. 112 pp. Polhemus, D. A. 1995a. A survey of the aquatic insect faunas of selected Hawaiian streams. Report for the Commission on Water Resource Management, Department of Land and Natural Resources, State of Hawaii. Department of Natural Sciences, Bishop Museum. Hawaii Biological Survey No. 1995- 008. 122 Polhemus, D. A. 1995b. New Heteroptera and Odonata (Insecta) records and range extensions in the Hawaiian Islands. Bishop Museum Occasional Papers 42: 42-43. Polhemus, D. A. 1996. The orangeblack Hawaiian damselfly Megalagrion xanthomelas (Odonata: Coenagrionidae): clarifying the current range of a threatened species. Bishop Museum Occasional Papers 45: 30-53. Randall, J. E., J. L. Earle, T. Hayes, C. Pittman, M. Severns, and R. J. F. Smith. 1993. Eleven new records and validations of shore fishes from the Hawaiian Islands. Pacific Science 47: 222-239. Samuelson, G. A. 1998. New records for Hawaiian Coleoptera. Bishop Museum Occasional Papers 56: 27 33. Smith, L. D., M. J. Wonham, L. D. McCann, G. M. Ruiz, A. H. Hines, and J. T. Carlton. 1999. Invasion pressure to a ballast-flooded estuary and an assessment of inoculant survival. Biological Invasions 1: 67-87. Springer, V. G. and M. F. Gomon. 1975. Revision of the Blenniid fish genus Omobranchus with descriptions of three new species and notes on other species of the tribe Omobranchini. Smithsonian Contributions to Zoology, Number 177. Steams, H. T. 1985. Geology of the State of Hawaii. Pacific Books, Palo Alto, California. Steams, H. T. and K. N. Vaksvik. 1935. Geology and ground-water resources of the island of Oahu, Hawaii. Territory of Hawaii, Department of Public Lands, Division of Hydrography, Bulletin 1. 123 Titcomb, M. 1972. Native use of fish in Hawaii. Second Edition. University of Hawaii Press, Honolulu, Hawaii. U.S. Congress, Office of Technology Assessment. 1993. Harmful non-indigenous species in the United States. OTA-F-565. Washington, DC: U.S. Government Printing Office, September 1993. Van Dine, D. L. 1907. The introduction of top-minnows (natural enemies of mosquitoes) into the Hawaiian Islands. Hawaiian Agricultural Experiment Station Press Bulletin 20: 1-10. Van Zwaluwenburg, R. H. 1945. Notes. Proceedings ofthe Hawaiian Entomological Society 12: 228. Visher, F. N. and J. F. Mink. 1964. Ground-water resources in southern Oahu, Hawaii. Geological Survey Water-Supply Paper 1778. 133 pp. Wester, L. 1981. Introduction and spread of mangroves in the Hawaiian Islands. Association ofPacific Coast Geography Yearbook. 43:125-137. Wirth, W. W. 1947. Ephydra gracilis Packard, a recent immigrant fly in Hawaii (Diptera: Ephydridae). Proceedings ofthe Hawaiian Entomological Society 13: 141-142. Yamamoto, M. N. and A. W. Tagawa. 2000. Hawaii's native and exotic freshwater animals. Mutual Publishing, Honolulu, Hawaii. Zink, R. M., J. M. Fitzsimons, D. L. Dittmann, D. R. Reynolds, and R. T. Nishimoto. 1996. Evolutionary genetics of Hawaiian Freshwater Fish. Copeia 1996:330-335. 124 CHAPTER 5. FLOW RESTORAnON AND PERSISTENCE OF INTRODUCED SPECIES IN WAIKELE STREAM,O'AHU 125 ABSTRACT Unintentional stream flow restoration in Waikele Stream, O'ahu, Hawai'i resulted from the demise of sugar cane cultivation on O'ahu and subsequent cessation of direct surface water diversions in 1989. Previous artificial stream studies in Hawai'i have suggested that increases in the base flow of a diverted stream would displace or reduce introduced fish populations. Surveys ofWaikele Stream, conducted in 1993 and 1997 1998 from the Waikele Springs area downstream to the beginning of the tidal reach found that despite an increase in stream flow, introduced fish remained abundant and native species appeared to have declined. In fact, two new introduced aquatic taxa, a dragonfly and a shrimp, had appeared. These results indicate that although restoring hydrological conditions is an important first step in overall restoration of degraded aquatic ecosystems, flow restoration alone is not a panacea, especially in O'ahu streams with naturally low discharge rates. For stream and wetland restoration to fully succeed, introduced fish and other alien aquatic species must be eradicated by methods other than simply increasing stream base flows. INTRODUCTION Hawaii has a significant and endemic freshwater fauna that is now seriously threatened. Native Hawaiian stream animals have been adversely affected by the accelerated introduction of non-native species (Devick 1991a), urban development, stream diversions, and stream channelization (Norton et al. 1978, Timbol and Maciolek 1978). In O'ahu streams, escaped ornamental species are increasingly displacing native organisms. Although the adverse effects associated with introduced freshwater species have been well documented in the Pacific region (Maciolek 1984, Arthington & Lloyd 1989, Crowl et al. 1992, Eldredge 1994), reliable solutions to the problem have not been developed. The Hawaii Division of Aquatic Resources has attempted to limit the spread of new introductions through educational advertisements in the media on the harmful effects of introduced species. However, limited efforts have been made to decrease the effects of introduced species already in Hawaiian streams, and success in eliminating or reducing introduced fish in freshwater habitats in Hawai'i has been equivocal. 126 One method, currently under discussion, to reduce the numbers of introduced species in Hawaiian streams involves restoring water flow to natural levels through the removal of agricultural water diversions. Flow restoration has been suggested as an effective way ofeliminating or reducing the abundance of introduced animals in historically diverted Hawaiian streams (Fitzsimons & Nishimoto 1996, Fitzsimons et a1. 1997). These results were based on laboratory studies conducted in artificial streams. However, it is not obvious how applicable the artificial stream results are to field situations, such as unchannelized streams. Increases in water velocity associated with increases in flow are believed to displace non-native organisms that have evolved in slow-water environments (Fitzsimons et a1. 1997). An obvious problem with this argument, however, is that regardless of flow regime, a stable unchannelized stream will always contain some slow-water habitats (Platts et a1. 1983, Helm et a1. 1985). In this study we examine Waikele Stream, an O'ahu catchment that was diverted from prior to 1931 until 1989. Species composition was assessed in this stream four and nine years after the cessation of large-scale agricultural water diversions in 1989. The purpose of this study was to document species composition of native and introduced aquatic animals in the area downstream of a hydrologically restored O'ahu stream, and to assess the success of flow restoration in regard to the removal or reduction of introduced species. STUDY AREA Waikele Stream drains the leeward slope of the Ko'olau Mountains and the windward slope of the Waianae Mountains in central O'ahu. From its origin in the Ko'olau range, the stream flows for 28 Ian to Pearl Harbor. The Wai'anae Mountain tributaries ofWaikele Stream are intermittent, and flow only following heavy rains. The main channel of Waikele Stream consists of alternating sections of flowing and dry stream except during and after periods of heavy precipitation. This is a natural condition caused by percolation of water into the alluvium. Flow is perennial only in the Ko'olau Mountain headwaters and near the mouth in the area of Waikele Springs. As in many O'ahu streams, streamflow is characteristically flashy, with high flood peaks and low baseflows (Nichols et a1. 1997). 127 The study area extended from approximately 250 m above Waikele Springs to a concrete weir spanning lower Waikele Stream 1.3 km downstream of the springs (Figure 5.1). This weir completely separates Figure 5.1. Waikele Stream study area. freshwater from the tidal reach on the downstream side, and the difference in stream levels can range from 1 1.5 m (Nance 1998). The terminal reach ofWaikele Stream is mainly fed by a series of large basal springs that are collectively called Waikele Springs. Waikele Springs issue from low points in the cap rock that release groundwater (Stearns and Vaksvik 1935). At base flow, Waikele Springs contribute approximately 80% of the water to the lower Waikele Stream (Nance 1998). 128 Unintentional stream flow restoration resulted from the demise of sugar cane cultivation on O'ahu and subsequent cessation of direct surface water diversions in 1989, and groundwater well pumping between 1994 and 1995 (Nance 1998). The earliest recorded diversion ofWaikele Springs was conducted in 1931 by the 0'ahu Sugar Company and consisted of0.1 m3Is (Stearns & Vaksvik 1935). Although it is known that diversions occurred prior to 1931, records are not available for these early diversions. From 1951 to 1989 flow diversion averaged 0.19 m3/s (Nance 1998). Diversions greater than 0.6 m3/s frequently occurred during droughts. Direct diversions from the Waikele Springs ceased in July 1989. The last groundwater pumping that could have affected flow in the Waikele Springs occurred in 1995 (Nance 1998). It is not known to what extent these wells affected Waikele Stream flow, but it appeared to be insignificant compared to direct diversions that occurred prior to 1989 (T. Nance, personal communication, 1998). During this study, habitat downstream of the Waikele Springs input consisted of high-velocity runs (0.3 0.6 m in depth) connecting shallow pools (0.2 -0.6 m in depth) with stream widths of 4-8 m. Unlike many low-elevation sections of O'ahu streams, Waikele Stream is not lined with concrete or channelized in the area ofWaikele Springs. The natural stream channel is maintained from above Waikele Springs downstream for 1.3 km until the stream is impounded by the concrete weir. METHODS Stream animals were sampled during two periods: March 1993, and December 1997 to April-August 1998. Sampling effort in 1997 and 1998 was close enough in time to be considered one sampling period. In 1993, sampling methods consisted of snorkeling, netting, and above-water observations. Quantitative fish sampling was not conducted in 1993, but relative abundance and fish species composition were noted. In 1998, one quantitative seine haul was used to characterize the relative abundance of introduced fish immediately below Waikele Springs. 129 A fine-mesh, 5 m long seine was used to sample stream animals and assess species composition. Snorkeling and above-water observation were also used, especially in fast-water habitats. During both periods aquatic insect collection focused on dragonflies and damselflies (Odonata). Odonata were captured with both aerial and dip nets, and dip nets were also used to sample areas not accessible to seines. Water velocity was measured with a Swoffer 2100 current meter in the main channel downstream of Waikele Springs. Velocity measurements were collected at 0.6 to 1.2 m intervals along transects that were perpendicular to stream flow. Three transects were established in areas of relatively laminar flow: 4, 18, and 33 m downstream of the beginning of the Waikele Springs input. Water velocities were measured at the water's surface and six tenths of total stream depth below the stream surface. The latter measurement location corresponds to the location of average water velocity in an ideal channel (Nielsen & Johnson 1983). Sampling effort in March 1993 consisted of approximately 50 hours of observations and organism collection in the study area. In 1997-1998, sampling effort increased to approximately 100 hours, with similar proportions of time spent snorkeling and netting. In both sampling periods observations were also made directly below the weir. RESULTS Water velocity measurements for each transect are shown in Table 5.1. In 1998, relatively high average water velocities were encountered downstream of the restored Waikele Springs. Water velocities as low as 10 cm/s were recorded near the streambanks. In 1997-1998, two previously unrecorded introduced species were collected in Waikele Stream (Table 5.2). Only three of the five freshwater fish species known from Hawaiian streams (Stenogobius hawaiiensis, A waous guamensis, and Eleotris sandwicensis) were captured. Most observations of native fish species 130 Table 5.1. The range and mean water velocities (:I: standard error) recorded in transects downstream of Waikele Springs. Distance from start of restored Waikele Springs outlet Surface Velocity Mid-Channel Velocity Mean Mid-Channel (stream width-m) (range-cm/s) (range-cm/s) Velocity (cm/s) 4 m (7.9 m wide) 10-67 10-67 33 :t 7 18 m (7.0 m wide) 25-92 16-93 52 :t 7 33 m (5.2 m wide) 10-80 13-65 48 :t 6 occurred in the tidal reach below the 1.5 m high concrete weir in lower Waikele Stream. Native fish such as Kuhlia sandvicensis, Mugil cephalus, and Eleotris sandwicensis, were common but found only downstream of this weir which precluded their upstream movement. The weir did not prevent upstream movement by native stream gobies. However, neither Sicyopterus stimpsoni nor Lentipes concolor were encountered during this or previous surveys. The stream gobies A. guamensis and S. hawaiiensis appear to have declined since 1993 in the study area. In 1993, high densities ofpost-larvae ofboth of these species were observed and collected below the weir (Figure 5.1). Although adults of both of these species were observed in 1998, post-larvae were not observed in that same location during extensive sampling conducted in 1997-1998. While not common, S. hawaiiensis was collected both up and downstream in the vicinity of the weir, but was not observed upstream near Waikele Springs in 1998, in contrast to 1993. Despite greater sampling effort than in 1993, only a few A. guamensis were observed just below the concrete weir in 1997-1998. Additionally, native fish were not observed in the vicinity of Waikele Springs in 1998, despite intensive sampling effort. In 1993, native gobies were common at the Waikele Springs area, with 12 A. guamensis in a wide range of size classes netted, including relatively recent recruits as small as 24 mm total length (Englund 1993). 131 Table 5.2. Introduced and native species found in Waikele Stream, Oahu in 1993 and 1997-1998 from 250 m above Waikele Springs downstream to concrete weir. Oahu introduction dates from Beardsley (1980), Devick (1991a), Cowie (1995), Polhemus & Asquith (1996), Randall (1996), Cowie (1998). Biogeographic Status Year of Oahu Introduction Taxon 1993 1998 or Discovery Fish A waous guamensis X X Indigenous Stenogobius hawaiiensis X X Endemic E1eotris sandwicensis X X Endemic Mugi1ogobius cavifrons X X Introduced Kuh1ia sandvicensis X X Endemic Mugil cepha1us X X Indigenous Moo1garda engeli X X Introduced 1955 Ancistris cf. temminckii X X Introduced 1985 Sarotherodon me1anotheron X X Introduced 1951 Poecilia reticu1ata X X Introduced 1922 Poecilia mexicana X X Introduced 1960 Gambusia affinis X X Introduced 1905 Xiphorphous helleri X X Introduced 1922 Crustaceans Macrobrachium grandimanus X X Endemic Procambarus c1arkii X X Introduced 1923 Macrobrachium 1ar X X Introduced 1957 Neocaridina denticu1ata X New Introduction 1991 Mollusks Corbicu1a fluminea X X Introduced 1988 Pomacea canalicu1ata X New Introduction! 1990 Tarebia granifera X X Introduced 1856 Damselflies/dragonflies (Odonata) Ischnura ramburii X X Introduced 1973 Ischnura posita X X Introduced 1936 Enallagma civile X Introduced 1936 Pantala flavescens X X Indigenous Anaxjunius X X Indigenous Crocothemis servilia X New Introduction 1994 Orthemis ferruginea X X Introduced 1977 Amphibians Bufo marinus XX Introduced 1932 Rana catesbeiana XX Introduced 1867 !Apple snails not directly found in stream, but in taro fields 25 m away from stream channel 132 Introduced fish were common throughout the study area. Tilapia (Sarotherodon melanotheron) were found both upstream and downstream of Waikele Springs. Bristle-nosed or armored catfish (Ancistris cf. temminckiI) were extremely abundant in runs and riffles downstream of Waike1e Springs. Two species of alien fish known to inhabit lower Waikele Stream, Chinese catfish (Clarias fuscus) and the rice paddy eel (Monopterus albus) (Hawaii Division of Aquatic Resources, personal communication, 1998), were not observed at this location in 1993 or 1997-1998. This is likely due to their wariness, or to gear limitations, since electrofishing is the most effective means of sampling these species. As in 1993, large numbers of introduced poeciliids such as Gambusia affinis, Poecilia mexicana, P. reticulata, and Xiphophorus helleri were observed in the Waikele Springs area downstream to the tidal reach. Densities of poeciliids remained high in 1997-1998. Using a haul seine (in a pool) in 1998 we found densities of 2.2/m2 for G. affinis, 1.5/m2 for P. reticulata, 0.4/m2 for P. mexicana, and O.2/m2 for X. helleri, for a total of 4.3/m2 for all poeciliids combined. Introduced dragonflies and damselflies dominated the aquatic insect fauna of Waikele Stream. All damselfly species were introduced (Table 5.2). Native Megalagrion damselflies were not observed in lower Waikele Stream in 1993 or 1998. The indigenous dragonfly Anaxjunius was common around Waikele Springs, and the introduced dragonfly Crocothemis servilia was absent in 1993, but common in 1998. The introduced freshwater shrimp, Neocaridina denticulata sinensis, was abundant in 1998 but was not found in 1993. Introduced apple snails (Pomacea canaliculata) were observed in taro fields in a separate and lower spring area that is adjacent to the weir below Waikele Springs. This area was within 25 m of the stream, but apple snails were not observed within the stream channel. 133 DISCUSSION In the last five years, two new species, a dragonfly and a shrimp, have become established in lower Waikele Stream, and an additional species of introduced aquatic snail was found within 25 m of the stream. At the same time, native stream animals have become less common despite flow restoration. The dragonfly C. servilia was fIrst collected around taro fields in Waiahole Stream, O'ahu in 1994 (Polhemus 1995). The rapid spread of this dragonfly across O'ahU was expected because of its vagility, thus its appearance at Waikele Stream is not surprising. Moreover, this dragonfly is suited to the disturbed, lowland aquatic habitats common on O'ahu. The long-term effects of this introduced dragonfly on native aquatic organisms are unknown. However, its distribution overlaps with the native dragonfly Anaxjunius, suggesting a potential for negative interactions. The freshwater shrimp, N. denticulata sinensis, was probably introduced to O'ahu streams as an escaped or released ornamental species. Its native range includes Taiwan, the Ryukyu Islands, Korea, mainland China, and Vietnam (Hung et a1. 1993). Recently this species has been found in several widely separated windward and leeward O'ahu streams (Devick 1991b). However, the Waikele record is the fIrst time it has been found in a Pearl Harbor stream. Previously this shrimp was incorrectly identified from Nu'uanu Stream, O'ahu, as Caridina weberi (Devick 1991b). It is possible that N. denticulata sinensis could compete with the native atyid shrimp Atyoida bisulcata. High densities of adult and immature apple snails were seen in an area of taro fields less than 25 m from Waikele Stream and within the floodplain. The presence of apple snails within 25 m of the stream means that it is highly likely they will soon be in Waikele Stream itself. Introduced poeciliids were abundant in the unchannelized and restored flow areas ofWaikele Stream. Total poeciliid densities in Waikele Stream were equal to or greater than those found in other similarly degraded 134 O'ahu streams. For example, poeciliid densities in low elevation areas of Kawa Stream (0-1.5 m above sea level) ranged from 1.6 to 2.9Im2 (Filbert and Englund 1995), compared to 4.3/m2 found in Waikele Stream. The high poeciliid densities appear not to support predictions made from artificial stream research. Using species found in Hawai'i, Fitzsimons et a1. (1997) found strong water flows displaced introduced poeciliids in an artificial stream, and then applied these displacement velocities to natural stream channels. Fitzsimons et a1. (1997) concluded that, " ...a stream with a base flow of 20 em/second or greater will be ideal for native fishes and will eliminate or at least suppress non-native poeciliids and the copepod [parasite] intermediate hosts...". Water velocities in the restored Waikele Stream ranged from 33 to 52 em/second (Table 5.1). Although the average mid-column and surface water velocities in Waikele Stream (below Waikele Springs) now far exceed 20 cm/s, poeciliids and other alien fish remain abundant in areas of high water velocities. Armored catfish (A. cf. temminckil} appear to preferentially select areas of highest water velocities, and were less common in pool habitats. The likely reason that introduced poeciliids were not displaced by the> 20 cm/s water velocities in Waikele Stream may be due to the habitat complexity found in natural, unchannelized streams. Large rocks, downed trees, side channels, and aquatic vegetation all offer a velocity refuge to introduced fish (such as poeciliids or tilapia) which favor slower water velocities than do the native stream gobies. Additionally, many O'ahu streams are small, and naturally have low water velocities at the stream mouth or in other low-elevation sections of the stream. The negative effects of introduced poeciliids on other vertebrates and invertebrates have been widely documented (Hurlbert et a1. 1972, Meffe & Snelson 1989). For example, G. affinis_prey upon eggs, larvae, and fry of sportfish and native fish in areas outside of their native habitat (Courtenay & Meffe 1989). Predation by introduced poeciliids was believed to be a significant cause of extirpation of native fish in Nevada (Courtenay & Meffe 1989) and invertebrates in Australia (Arthington & Lloyd 1989). In south western Australia, Morgan et a1. (1998) found fin nipping by G. affinis holbrooki, the eastern mosquitofish, caused extensive caudal fin damage to native fish species. 135 The persistence of poeciliids in Waikele Stream after flow restoration will also likely prohibit the recolonization of this area by native stream breeding Megalagrion damselflies. Polhemus & Asquith (1996) believed the presence of introduced poeciliids was responsible for the absence of native Megalagrion species in areas where they co-occurred. These authors found a complete absence of native damselflies in low elevation areas similar to Waikele Stream where introduced poeciliids were found. Additionally, bristle-nosed catfish densities were high, and may also be contributing to the absence of A. guamensis in the study area in 1998. Native stream gobies are undoubtedly adversely affected by loricariid catfish through competition for food and space. The species ofloricariid catfish found on 0'ahu are primarily algivores, but will also readily consume fish eggs (1. Armbruster, personal communication, 1998). Native stream gobies are rare in low elevation areas of O'ahu streams containing very high densities of introduced armored catfish (Kawa Stream, Filbert & Englund 1995; Manoa Stream, Bishop Museum unpublished database). Flow restoration between 1989 and 1998 appears to have had little beneficial effect on native stream animals. Our surveys in 1993 and 1997-1998 indicate native animals are rare to non-existent and introduced species are more abundant. This suggests that elements other than flow may be important to the rehabilitation of native stream communities in Hawai'i. However, there may be situations in which flow restoration alone has had a beneficial effect on native organisms. Flow restoration in an unnamed stream at TripIer Army Medical Center (O'ahu) led to an increased abundance of a rare native damselfly (Megalagrion xanthomelas). After ten months of flow restoration, adult damselfly observations increased from 17 to 162 adults per monitoring period (Englund 1998). The absence of introduced fish in this stream likely explains the success offlow restoration in this case. In the presence of an introduced fish damselfly abundance would likely not have increased (Polhemus & Asquith 1996). 136 The results of this study corroborate other field observations that flow restoration alone will not reduce the numbers of alien species. On O'ahu, recovery of native freshwater vertebrates and invertebrates will not occur until it is understood that alien species now dominate the system in low elevation aquatic habitats. For stream and wetland restoration to succeed, introduced fish and other harmful aliens must first be eradicated. Introduced poeciliids occur in almost every major wetland in the Hawaiian archipelago. Even if it were possible to flush introduced fish out of an Hawaiian stream, adjoining wetlands or side-channel habitats would still provide low velocity refugia. This would limit the recolonization of streams by native aquatic insects such as Megalagrion damselflies. Introduced freshwater fish now occur in many parts ofthe world, including Australia, New Zealand, and on most Pacific islands with freshwater habitats (Maciolek 1984, Eldredge 1994). They threaten the biodiversity of aquatic ecosystems throughout the Pacific region. Restoring hydrological conditions is an important first step in restoring degraded aquatic ecosystems but flow restoration is not a panacea, especially in areas having streams with a naturally low baseflow discharges such as those on O'ahu. To preserve native fish and invertebrate biodiversity in Hawai'i and the Pacific region, creative management solutions must be found, including the elimination of introduced species. Every step should also be taken to ensure that new species introductions do not occur in pristine aquatic habitats. 137 REFERENCES Arthington, A H. and L. N. Lloyd. 1989. Introduced Poeciliids in Australia and New Zealand. In G.K. Meffe and F.F. Snelson, Jr. (eds.), Ecology and evolution of livebearing fishes (Poeciliidae), pp. 333-348. Prentice Hall. Beardsley, 1. W. 1980. New record of Orthemis ferruginea (Fabricius) from Oahu. Proceedings ofthe Hawaiian Entomological Society 23: 83. Courtenay, WK, & G.K. Meffe. 1989. Small fishes in strange places: a review of introduced poeciliids. In: G.K. Meffe and F.F. Snelson, Jr. (eds.), Ecology and evolution of livebearing fishes (Poeciliidae), pp. 319-331. Prentice Hall. Crowl, T. A, A R. Townsend and A. R. McIntosh. 1992. The impact of introduced brown and rainbow trout on native fish: the case of Australasia. Reviews in Fish Biology and Fisheries 2: 217-241. Cowie, R. H. 1995. Identity, distribution and impacts ofintroduced Ampullariidae and Viviparidae in the Hawaiian Islands. Journal ofMedical and Applied Malacology 5: 61-67. Cowie, R. H. 1998. Patterns of introductions of non-indigenous non-marine snails and slugs in the Hawaiian Islands. Biodiversity and Conservation 7: 349-368. Devick, W. S. 1991a. Patterns of introductions of aquatic organisms to Hawaiian freshwater habitats. In: New directions in research, management, and conservation of Hawaiian freshwater stream ecosystems, Proceedings ofthe 1990 Symposium on Freshwater Stream Biology and Management, State of Hawaii, pp. 189-213. Hawaii Division of Aquatic Resources. 138 Devick, W. S. 1991b. Job progress report F-14-R-15 [Fresh water fisheries and surveys]. Disturbances and fluctuations in the Wahiawa Reservoir Ecosystem. Hawaii Division of Aquatic Resources. Eldredge, L. G. 1994. Perspectives in aquatic species management in the Pacific islands. Volume 1, introductions ofcommercially significant aquatic organisms to the Pacific Islands. South Pacific Commission, Noumea, New Caledonia. 127 pp. Englund, R. A. 1993. A survey of the fish and aquatic insect fauna of the Waikele/Kipapa streams, Oahu, Hawaii. BHP Environmental Technologies report prepared for Halekua Development Corp., Honolulu. 20 pages. Englund, R. A. 1998. Response of the Orangeblack Hawaiian Damselfly (Megalagrion xanthomelas), a Candidate Threatened Species, to increases in stream flow. Bishop Museum Occasional Papers 56: 19-24. Filbert, R. B. and R. A. Englund. 1995. Assessment of the freshwater macrofauna of Kawa Stream, Oahu. Pacific Aquatic Environmental report prepared for Pacific Atlas Hawaii, Inc. 19 pages. Fitzsimons, J. M. and R. T. Nishimoto. 1996. Recovery of three Kauai streams from Hurricane Iniki and implications for the restoration and regeneration of freshwater ecosystems in Hawaii. In: Will Stream Restoration Benefit Freshwater, Estuarine, and Marine Fisheries? Proceedings of the October 1994 Hawaii Stream Restoration Symposium. Fitzsimons, J. M., H. L. Schoenfuss, & T. C. Schoenfuss. 1997. Significance of unimpeded flows in limiting the transmission of parasites from exotics to Hawaiian stream fishes. Micronesica 30: 117-125. 139 Helm, W. T., P. Brouha, M. Aceotima, C. Armour, P. Bisson, J. Hall, G. Holton, & M. Shaw. 1985. Glossary of stream habitat terms. Western Division, American Fisheries Society. 34 pp. Hung, M. S., T. Y. Chan, & H. S. Yu. 1993. Atyid shrimps (Decapoda: Caridea) of Taiwan, with descriptions of three new species. Journal ofCrustacean Biology 13: 481-503. Hurlbert, S. H., J. Zedler, & D. Fairbanks. 1972. Ecosystem alteration by mosquitofish (Gambusia afflnis) predation. Science 175: 639-41. Maciolek, J. A. 1984. Exotic fishes in Hawaii and other islands of Oceania. In: W.R. Courtenay, Jr. and J.R. Stauffer, Jr. (eds.), Distribution, biology, and management of exotic fishes, pp. 131-161. Johns Hopkins University Press, Baltimore. Meffe, G. K. & F. F. Snelson. 1989. An ecological overview of poeciliid fishes. In G.K. Meffe and F.F. Snelson, Jr. (eds.), Ecology and evolution of livebearing fishes (Poeciliidae), pp. 13-31. Prentice Hall. Morgan, D. L., H. S. Gill, & I. C. Potter. 1998. Distribution, identification and biology of freshwater fishes in south-western Australia. Records ofthe Westem Australian Museum (supplement 56) 97 pp. Nance, T. 1998. Effect of the proposed use ofthe WP-18 pump station on Waikele Stream and West Loch. Tom Nance Water Resource Engineering, 680 Ala Moana Boulevard, Suite 406. 39 pages + appendices. 140 Nichols, W. D., Shade, P. J. & C. D. Hunt, Jr. 1997. Summary of the Oahu, Hawaii, regional aquifer system analysis. U.S. Geological Survey Professional Paper 1412-A. 61 pp. Nielsen, L. A., & D. L. Johnson. 1983. Fisheries Techniques. American Fisheries Society. 468 pp. Norton, S. E., Timbol, A. S., & J. D. Parrish. 1978. Stream channel modification in Hawaii. Part B: effect ofchannelization on the distribution and abundance offauna in selected streams. FWS/OBS 78117. USFWS National Stream Alteration Team, Columbia, Missouri. 47 pp. Platts, W. S., Megahan, W. E, & G. W. Minshall. 1983. Methods for evaluating stream, riparian, and biotic conditions. Intermountain Forest and Range Experiment Station. U.S. Forest Service, Ogden, Utah. 70 pp. Polhemus, D. A. 1995. New Heteroptera and Odonata (Insecta) records and range extensions in the Hawaiian Islands. Bishop Museum Occasional Papers 42: 42-43. Polhemus, D.A. & A. Asquith. 1996. Hawaiian damselflies: a field identification guide. Bishop Museum Press. 122 pp. Randall, J.E. 1996. Shore Fishes of Hawaii. Natural World Press. 216 pp. Stearns, H. T., and K. N. Vaksvik. 1935. Geology and ground-water resources of the island of Oahu, Hawaii. Territory of Hawaii, Department of Public Lands, Division of Hydrography, Bulletin 1, 479 p. 141 Timbol S. A. and J. A. Maciolek. 1978. Stream channel modification in Hawaii, Part A: Statewide inventory of streams, habitat factors and associated biota. FWS/OBS-78-16. USFWS National Stream Alteration Team, Columbia, Missouri. 157 pp. 142 CHAPTER 6: INVASIVE SPECIES THREATS TO NATIVE AQUATIC INSECT AND ARTHROPOD BIODIVERSITY IN HAWAI'I, THE PACIFIC AND OTHER RELEVANT AREAS WITH DISCUSSION OF CONSERVATION MEASURES INTRODUCTION 143 The conservation status of the native aquatic insect faunas of tropical insular regions, especially in highly diverse groups such as Odonata, Diptera, and Heteroptera, is poorly known. A notable exception is Hawai'i, where extinctions in some groups have been documented (Liebherr and Polhemus 1997; Englund 1999). Restricted habitats, small population sizes, and a lack of defenses against invasive species make tropical insular species especially vulnerable to disturbance and extinction (Simberloff 1986, 1995; Paulay 1994). Conservation biologists working on isolated oceanic islands such as Hawai'i and other areas of Polynesia have long recognized that introduced species represent the most pervasive and persistent threat to the survival of these insular biotas (Elton 1958, Vitousek 1988; Meyer and Florence 1997; Loope et aI. 2001; Staples and Cowie 2001). Within the Pacific region, the Hawaiian Islands have received the greatest amount of attention related to the spread of invasive species and their impacts on native aquatic biota (Eldredge 1994), although a few other tropical regions including Fiji (Andrews 1985), French Polynesia (Polhemus et a1. 2000; Keith et a1. 2002; Englund 2003) and New Caledonia (Gargominy et a1. 1996; Marquet et aI. 2003) have had limited research devoted to this problem. This review assesses the susceptibility of native arthropods to invasive species in island or island-like environments. A brief review of the impacts of invasive species on native insects in other tropical and temperate warm regions is provided, as well as a synthesis of the invasive species problem facing freshwater insects and other terrestrial arthropods in Hawai'i, the Pacific, and other relevant areas. Some terrestrial systems are included as they provide many relevant examples of invasive species impacts, and remedial practices that have been effective in some cases. It may be possible to extend findings from terrestrial systems to aquatic systems for effective conservation of insects. The Hawaiian example of invasive species impacts on native insects and other arthropods such as certain spiders has many parallels to insect biotas in other vulnerable island and island-like habitats, and therefore forms a central component of this overview. Finally, drawing on a mixed record of past mistakes and successes in Hawai'i and elsewhere, some practical conservation measures intended to preserve and restore endemic aquatic insects are proposed. 144 In Hawai'i, the first record of any invasive insect species in the literature is the 1826 introduction of the mosquito Culex quinquefasciatus Say (Hardy 1960). The introduction of this mosquito species and its subsequent vectoring of avian malaria have resulted in catastrophic impacts to much of the native Hawaiian forest bird fauna (Van Riper et a1. 1986; Van Riper and Scott 2001). While extinctions among Hawaiian forest birds and the current restriction of their remaining populations to elevations above the range of mosquitoes has received considerable worldwide attention, a less publicized but nevertheless ongoing pattern of species extirpation has also been occurring among native Hawaiian insects, both terrestrial and aquatic. In both cases the common thread is invasive species. In addition to the problems caused by invasive species, lack of taxonomic resolution remains one of the greatest problems related to arthropod conservation on oceanic islands (Gillespie 1999; Nishida and Evenhuis 2000). This is particularly true for aquatic insects, as relatively few species have been collected or described from the insular tropical Pacific. While the Pacific region from New Guinea to Hawai'i probably contains at least 900,000 arthropod species, amounting to an estimated 15% of the world's total (Allison and Englund, in press), a lack of basic knowledge prevents effective conservation measures from being established. The remote nature and consequent difficulty in accessing oceanic islands has also been a major obstacle in providing basic taxonomic assessments for those areas. Outside of Hawai'i, only the most economically important aquatic insects such as the biting blackflies and sandflies (Simuliidae and Ceratopogonidae respectively) and mosquitoes (Culicidae) have received any great amount ofecological or taxonomic treatment. For example, other than the research conducted on the anthropophilic black flies of French Polynesia (Craig 1997,2001,2003; Craig and Currie 1999; Craig and Joy 2000; Craig et a1. 1995; Craig et a1. 2001), biting Ceratopogonidae (Macfie 1935), and Dolichopodidae (Bickel 1994; Evenhuis 1999), little is known about the aquatic insect fauna of the Society Islands or French Polynesia (e.g., Paulian 1998; Sechan 1998). 145 The lack of attention to aquatic systems throughout the insular Pacific region may have resulted from the fact that initial biological assessments of freshwater ecosystems were generally haphazard and secondary to terrestrial arthropod assessments. Although thousands of new terrestrial arthropod species were collected during the Pacific Entomological Surveys ofFrench Polynesia and elsewhere by Bishop Museum and Hawaii Sugar Planters' Association staff during the late 1920s and 1930s (Zimmerman 1935; Adamson 1936, 1939) only a small portion of these were aquatic insects, consisting mostly of a few aquatic species of Tipulidae (Alexander 1932,1933,1935), Dolichopodidae (Lamb 1933), and Odonata (Mumford 1935, 1936; Needham 1942). Additionally, the biting flies that created an obvious nuisance to researchers (Simuliidae and Ceratopogonidae) (Edwards 1932, 1933a, 1933b) and a few shorefly species (Ephydridae) (Malloch 1935) were also described from these early surveys. The overall dearth of Pacific aquatic insect collections in general contrasts greatly with the treatment of Hawaiian Odonata, which attracted the attention of both professional and amateur naturalists from the first stages of European exploration, probably due to their large size and stunning appearance. The first native aquatic insect described from Hawai'i was Anax strenuus (Hagen 1867), collected during the expedition of the Danish corvette Galathea in 1846 (Bille 1851). The first native damselfly collected and described from Hawai'i was Megalagrion xanthomelas (Selys Longchamps 1876) taken by G.F. Matthew of the Royal Navy some time prior to 1876 (the specimens were apparently labeled only "Sandwich Islands") (Polhemus 1996). A few years later, McLachlan (1883) described five endemic damselfly and one dragonfly species from specimens collected in Hawai'i by Reverend Thomas Blackburn. In the 1880s, Blackburn made further collections of native aquatic insects in Hawai'i, describing three Megalagrion damselflies from O'abu, Uina'i and Maui (Blackburn 1884). Just a few years after Blackburn's collections, the most comprehensive historical collections of Hawaiian aquatic insects began in 1892 with the formation of the British Association for the Advancement of Science's Sandwich Islands Committee, which sent R.C.L. Perkins to Hawai'i to collect and catalog the 146 islands' fauna (Juvik 2001). Perkins's work led to the publication of the Fauna Hawaiiensis, which included descriptions ofmany new aquatic insect species (Perkins 1913). The aquatic insect specimens collected by Perkins and later collectors such as F.X. Williams (Williams 1936) provided a wealth of historical information on aquatic insect distributions prior to large-scale perturbations from urbanization and the introduction of the majority of non-native species into Hawai'i. By contrast, relatively few aquatic insect species were described or documented from the Pacific Entomological Surveys of the 1920s and 1930s, or surveys such as the St. George Expedition to Tahiti in the 1920s. For example, two new species of Veliidae (Cheesman 1926), a few widespread dragonflies and one dolichopodid species (Parent 1934) were collected by Evelyn Cheesman during the St. George Expedition (Cheesman 1927). The historical record of Hawaiian aquatic insects largely provided by Perkins and Williams is thus lacking from most other Pacific island areas. CASE STUDIES OF IMPACTS AND DOCUMENTED EXTINCTIONS CAUSED BY INVASIVE SPECIES Hawai'i Because surveying an entire island or all of its suitable habitats in typically rugged Pacific island terrain is difficult, extinction is often a complicated matter for researchers to verify, especially in the case of smaller and less conspicuous arthropods. Extinctions ofestablished populations are the result of these species being unable to adapt to perturbations or changes in their environment to which they have not been previously exposed, or if the population drops below a level necessary for its continued survival (Howarth and Ramsay 1991). In Hawai'i, there are many notable cases of presumed arthropod extinctions. However, with the exception of a small percentage of the insect biota, recent systematic surveys to verify the status of a particular group of insects have rarely been conducted. There are 5,368 described species of native insects in the Hawaiian Islands, accounting for 26% of the 20,440 species of native plants and animals found in the 147 archipelago (Eldredge and Evenhuis 2003). Ninety-eight percent (5,233) of these insects are endemic to Hawai'i (Eldredge and Evenhuis 2003) with many species found only on a single island. While definitive figures on the conservation status of the entire Hawaiian insect fauna are not available, extrapolations can be made from groups that have been the subject of detailed study since the early 1990s (Table 1). These groups include all of the Odonata, some Diptera (Foote and Carson 1995, Evenhuis and Polhemus 1994, Evenhuis 1997a, 1997b, Englund unpub!. data), all Heteroptera (Tummons 2003, Polhemus in litt.), carabid beetles (Liebherr and Zimmerman 2000; Liebherr 2005), nitidulid beetles (C. Ewing, pers. comm.), Rhyncogonus weevils (Samuelson 2003), Hylaeus bees (Daly and Magnacca 2003), and geometrids in the genus Scotorythra (Heddle 2003), totaling 997 species (Table 1). Recently well surveyed insect groups in Hawaii then account for 18.7% of the taxa that have had their conservation status examined. Of this total, 12.1 % of the included species are presumed extinct, with the presumption of extinction for this review based on the fact that a particular taxon has not been collected in the past 50 years (Liebherr and Zimmerman 2000). The calculated extinction rate of 12.1% is probably an overestimate for all native taxa as recent studies for several groups (e.g., damselflies, Rhyncogonus weevils) were funded because some of these taxa were believed to be threatened and at risk. However, even if this percentage was overly pessimistic it indicates that of the 5,368 described native insect species, at least 4,700 or more species are still likely extant. These estimates also do not factor in the many undescribed native insects, estimated to be as high as several thousand species (Howarth and Ramsay 1991). Thus, with an even more conservative 7-10% extinction rate estimate, there still remains a large number of native Hawaiian insect species of great cultural, scientific, and aesthetic value worthy ofconservation attention. The Hawaiian situation of some documented terrestrial and aquatic insect extinctions contrasts with that of continental North America where no extinctions of aquatic insect taxa have been documented, and only 204 out of 10,000, or 2% of estimated taxa were considered at risk for extinction in 1993 (Polhemus 1993). 148 Table 6.1. Extinction status of native insect taxa in the Hawaiian Islands that have recently had their conservation status examined, to lowest taxonomic resolution. No. Species Examined in Total Presumed % Presumed Taxa Study Extinct Extinct Odonata Zvgoptera 29 2 7 Anisoptera 4 0 0 Heteroptera 370 38 10 Coleoptera Carabidae 193 32 17 Nitidulidael 142 16 11 Curculionidae: Rhvncogonus 47 2 4 Diptera Drosophilidae: Drosophila 14 0 0 Dolichopodidae: Campsicnemus 26 0 0 Dolichopodidae: Emperoptera 5 4 80 Dolichopodidae: Sigmatineurum 11 0 0 Canacidae: Procanace2 7 0 0 Ephydridae: Scatell:r 16 1 6 Chironomidae: Telmatogeton2 7 0 0 Lepidoptera Crambidae: Omiodes 23 9 39 Geometridae: Scotorythra 43 4 9 Hymenoptera Colletidae: Hvlaeus 60 11 18 Totals 997 121 12.1 IC. Ewing, pers. comm, 2Englund, unpublished data. From: Foote and Carson (1995); Polhemus and Asquith (1996); Evenhuis and Polhemus (1994) Evenhuis (1997a, 1997b); Liebherr (2005); Liebherr and Zimmerman (2000); Nishida (2002); Evenhuis (2003); Daly and Magnacca (2003); Heddle (2003); Samuelson (2003); Tummons (2003), Haines (2004), Polhemus (in litt.); C. Ewing (pers. comm.); Englund (unpub!. data) Invasive species appear to be a leading cause ofinsect extinctions in Hawai'i and throughout the tropical insular Pacific. A nonindigenous or alien species is considered invasive if it spreads rapidly on its own and causes serious problems to human health, agriculture, commerce, or the environment (Simberloff 1997a; Staples and Cowie 2001). Invasive species have certain intrinsic features that allow them to outcompete native species including: a) being adaptable enough to thrive in different habitats; b) tolerance of a wide range of disturbances; c) being fast growing and capable of displacing other species; d) being highly dispersible to new locations; and e) a high reproductive rate (Simberloff 1997a; Staples and Cowie 2001). Invasive aquatic species are introduced to new habitats through a wide variety of pathways (Polhemus and 149 Englund 2003) including sportfish introductions, intentional food and aquaculture or aquarium introductions, association with aquarium plants, and ballast water (Englund et a1. 2000a; Englund 2002). Extrinsic factors such as habitat disturbance and transport by humans may also facilitate or increase the spread ofinvasive species. The majority of plant (Mueller-Dombois and Fosberg 1998) and animal species (Staples and Cowie 2001; Yamamoto and Tagawa 2000) encountered in and near the urbanized lowlands of Hawai'i are invasives, and likely are even more important in terms of native species declines than direct habitat losses caused by urban and agricultural development. Invasives are thus the primary factor in the demise of most of the lowland Hawaiian biota, and unlike anthropogenic perturbations that are amenable to regulation, control of an invasive species is quite difficult, though not impossible, once it becomes well established (Howarth and Ramsay 1991; Simberloff 1997b; Cox 1999). Because invasive ungulates such as feral pigs or goats directly modify large areas of native habitat, and other invasives prey upon or parasitize native species, their impacts on Hawaiian arthropods are likely much greater than those of direct anthropogenic habitat modification such as urban or agricultural developments. 0'ahu exemplifies the impacts of invasives, and even though large portions of the island are now devoted to housing, agriculture, or commercial developments, there still remain numerous undeveloped low-elevation areas that are dominated by invasive plants and insects. For example, surveys ofPearl Harbor wetland areas showed that 25% of identified taxa were native aquatic insect species (Englund et a1. 2000a), while terrestrial insect surveys in the undeveloped Wa'ahila Ridge area of O'ahu yielded 16% native species (Cowie et a1. 1999). Arthropods collected during similar surveys around the Kahului, Maui airport area identified only 11 % native species (Howarth and Preston 2002). On Kaua'i, only 24 of 283 (<10%) identified insect species from surveys of a 900 ha lowland mixed agricultural and undeveloped area were native species (Asquith and Messing 1993). These findings clearly indicate that invasive insect species are now dominant in the Hawaiian lowlands. 150 Overview of Invasive Ant Impacts on Hawaiian Arthropods Ants provide a particularly dramatic illustration of the impact of invasive species, and a thorough review of ant impacts in Hawai'i and the Pacific was recently provided by Nishida and Evenhuis (2000). The Hawaiian archipelago was one of the few large island groups in the world originally lacking native ant species (Wilson 1996), and 43 established invasive species of ants have been implicated in the demise of many native arthropod species (Howarth et a1. 2001). Without predatory social insects such as ants, a unique and disharmonic insect fauna developed in Hawai'i from the 350-400 colonizers that evolved into an estimated 10,000 endemic species (Howarth and Mull 1992). Additionally, without ants there have been large radiations of non-formicid predators and scavengers such as the predatory caterpillars (Eupithecia spp.), and certain groups such as carabid beetles are often extremely abundant (Wilson 1996; Liebherr and Zimmerman 2000). Characteristics such as flightlessness and naivete in island species have lead to the vulnerability of native insects to predation pressures from highly invasive and aggressive social insects such as ants. For example, the current distributional range in Hawai'i of native Tetragnatha spiders does not overlap with that of Pheidole megacephala (Fabricius), even in mostly native forests found at low elevations, and habitat disturbance seems to have lesser impacts on native Tetragnatha as long as P. megacephala is absent (Gillespie 1999). Gillespie and Reimer (1993) found that the crazy ant Anoplolepis gracilipes (F. Smith) [referred to as "longipes" by Gillespie and Reimer] and P. megacephala were the most problematic and that these ant species excluded native spiders from both native and non-native forests. They also found that introduced spiders have either strong exoskeletons or can lose their legs enabling them to escape when attacked by ants, or they can wrap the attacking ants in silk. Native Tetragnatha spiders have softer bodies, are smaller-sized, and never wrap their prey in silk for immobilization. Forest-inhabiting Tetragnatha were more at risk than riparian species, the latter having far fewer interactions with ants because of their occurrence in proximity to water barriers. The endemic Hawaiian Hylaeus bees also are quite susceptible to ant predation; and the absence of Hylaeus from otherwise suitable coastal habitats appears to be correlated with the presence of ants (Daly and 151 Magnacca 2003). In Haleakala National Park, Hylaeus have been severely impacted by the presence of the Argentine ant (Linepithema humile (Mayr», because the ants form large colonies and are attracted to sugar and nectar (Daly and Magnacca 2003) thus interfering with the feeding habitats of native bees. Currently, Argentine ants are found only in small areas of HaleakaIa, but if their spread remains unchecked then Hylaeus, which represents the largest extant group of native Hawaiian bees, will likely be decimated or at least significantly reduced in numbers. Other examples of taxa probably impacted by ants include the flightless fly Emperoptera mirabilis Grimshaw (Diptera: Dolichopodidae) that lived on the forest floor on Mt. Tantalus near Honolulu, but has not been collected since 1900 and is now considered extinct (Zimmerman 1970; Evenhuis 1997a). The extinction of this species occurred apparently quite rapidly, as Perkins was unable to recollect it on Mt. Tantalus between 1900 and 1907 (Perkins 1907) after finding it locally abundant during his original collections in 1900 (Evenhuis 1997a). Perkins remarked on the rapid changes that had occurred on Mt. Tantalus in the short period of time between his collections in 1900 and his subsequent visit in 1907, with Pheidole ants having become a problem in the lowlands of Hawai'i, including the Tantalus area, starting in the late 1890s (Perkins 1913). Similarly, Colpacaccus tantalus (Blackburn) a lowland generalist predator and scavenger that was formerly the most common O'ahu carabid beetle species, has also become extinct (Liebherr and Polhemus 1997) and was likely unable to compete with aggressive ant species such as P. megacephala (Liebherr and Zimmerman 2000). Areas of Haleakala National Park with Argentine ants contained significantly reduced densities of the carabid beetle Mecyc1othorax robustus (Blackburn) in contrast to areas of lacking Argentine ants (Cole et a1. 1992). A wide range ofother native arthropods in Haleakala National Park were found to have depressed populations, including Araneidae, Collembola, Hymenoptera, and Lepidoptera in areas where Argentine ants were present, as compared to areas where this ant species was absent (Cole et a1. 1992). 152 The Hawaiian Islands contain twice as many species of crickets as found in the entire continental United States (Otte 1994) and this large radiation is also threatened by ants. Most of the literature on the impacts of ants in Hawai'i has necessarily been anecdotal as many ants were introduced at about the same time Perkins was conducting the initial entomological surveys of the islands in 1892. However, Laupala cricket populations have been documented to have undergone catastrophic population declines when ants have been introduced. LaPolla et a1. (2000) found that cricket populations disappeared on eastern Kaua'i after the invasion of Pheidole megacephaIa. LaPolla et aI. (2000) demonstrated that LaupaIa crickets are now consistently absent or scarce in other areas that have been recently invaded by P. megacephala; these ants also invaded and quickly killed a laboratory colony ofthe native crickets being kept for research purposes at the University of Hawai'i during the same study. In contrast to native Trigonidium crickets that lay their eggs under bark in trees, LaPolla et a1. (2000) found that LaupaIa crickets lay their eggs in leaf litter and are more vulnerable to P. megacephala predation because this species is not generally arboreal. As a result, Trigonidium crickets appear to be unimpacted by ants in contrast to the large-scale disappearance of LaupaIa crickets from regions with these invasive ants. Ants also have likely played a major role in the extinction and rarity of some of the Hawaiian Odonata. Although not believed to be a direct cause of the severe range contraction of a remnant Megalagrion xanthomelas population at TripIer Medical Center on O'ahu, ant predation on emerging damselfly larvae of this species was documented at this site (Englund 2001a). Ants present a much more direct threat to the terrestrial naiads of Hawaiian damselfly species that appear to breed beneath banks of Dicranopteris linearis ferns. While one only species, Megalagrion oahuense (Blackburn, 1884), has been conclusively shown to have a terrestrial naiad life stage in fern litter (Williams 1936), phylogenetic evidence suggests that several closely related species also have a terrestrial immature stage (Jordan et a1. 2003). These include the Maui endemic MegaIagrion nesiotes (Perkins) and the Kaua'i endemic Megalagrion williamsoni (Perkins), species that appear to have been severely impacted by ant introductions and are now known from only one population each in lower mid-elevation sites (330-450 meters) still lacking ants along riparian areas. Other 153 presumed terrestrially breeding Megalagrion damselfly species probably impacted by ants and likely extinct include M. molokaiense (a Moloka'i endemic) last collected in 1927, Mjugorum (endemic to Maui and Uina'i) last found in 1896, and the Hawai'i Island population of M. nesiotes, last collected in 1906. Collectors have failed to rediscover these three species despite intense collection efforts since 1990 (Polhemus and Asquith 1996). Although the impacts of ants on endemic Hawaiian aquatic insects are generally difficult to assess because ants are now commonly found in most riparian areas below 800-900 meters elevation, Hardy (1979) reported that the recent invasion of the crazy ant, Anoplolepis gracilipes [referred to as "longipes" by Hardy] completely eliminated the aquatic insect fauna in certain sections ofPua'alu'u Stream, Maui. Crazy ants had completely invaded and "wiped out" most of Pua'alu'u Stream between 50 to 150 meters elevation (Hardy 1979). Hardy (1979) also noted that he had conducted surveys of nearby 'O'heo Stream many times prior to the invasion ofcrazy ants and he found it "a shocking contrast" that formerly large populations ofendemic aquatic flies such as Te1matogeton, Scatella ["Neoscatella"], and Procanace were gone, with the exception of a few individual Scatella that had escaped the ants by living under waterfalls. Feral ungulates and rats have also directly affected habitats and native plant species, and also likely preyed on larger invertebrate taxa such as some of the native beetle species. The large, flightless, endemic Rhyncogonus weevils of Hawai'i appear to be particularly susceptible to invasive species and disturbance. Although only 4% of these species are currently reported as possibly extinct, virtually all Hawaiian species, particularly those of the lowlands are now rare and the entire genus is listed as Species of Concern by the U.S. Fish & Wildlife Service (Samuelson 2003). Rhyncogonus bryani Perkins, a Laysan Island endemic, has not been collected since the introduction of the European rabbit (Oryctolagus cuniculus (Linnaeus, 1758)) in 1902 resulted in Laysan Island being denuded and also caused the extinction of three offive endemic island birds (Carlquist 1980) and at least three species of noctuid moths (Gagne and Howarth 1985). Rhyncogonus extraneus Perkins has not been collected since 1941 despite searches by numerous O'ahu 154 entomologists and this species is now believed extinct through a combination of rat and ant impacts coupled with pesticide applications (Samuelson 2003). Phytophagous insect species in Hawai'i are also at risk from the effects of feral ungulates. The first Hawaiian Heteroptera were collected 170 years ago and since then over 370 species have been described. Of these, 38 species have not been found during recent intensive collecting throughout the Hawaiian Islands by D.A. Polhemus (Tummons 2003). Because phytophagous Hawaiian Heteroptera are usually host-plant specific (Gagne 1997), native plant communities at greatest risk would be expected to also contain the most threatened insect communities. However, Polhemus (in litt.) has found that the greatest Heteroptera extinctions have occurred in the relatively abundant mid-elevation wet forest areas but, surprisingly, not in the more endangered lowland dry forest communities (Mueller-Dombois and Fosberg 1998). Even so, the floristically more diverse lowland dry forests are now found only in small remnant patches, and although they currently retain much of their Heteropteran diversity, they are critically threatened by fire and by a wide array of invasive plant and animal species (Wagner et aI. 1999). Invasive Species Impacts on Insects Outside Hawai'i Because early collections of insects in Hawai'i were made subsequent to Polynesian disturbances but prior to the more devastating urbanization and large-scale invasive species introductions that accompanied European contacts, there has been a thorough documentation of species impacts in the Hawaiian Archipelago. However, in many other parts of the tropics it is much more difficult to document extinctions or even invasive species impacts on native arthropods. This is because the biotas of most other tropical areas have not received the taxonomic treatment and study that has occurred in Hawai'i since the 1890s. The Hawaiian Islands occupy a relatively small and discrete area, so changes over time are easier to document than in continental areas or large islands like New Guinea. However, there are interesting parallels between the well-documented extinctions and negative impacts resulting from invasive species introductions in 155 Hawai'i and those on other Pacific islands and elsewhere such as South Africa in the Cape Fynbos, an island-like Mediterranean floral region. A suite of invasive plant and animal species often work in combination to eliminate native species by changing and eliminating native species habitat (Vitousek 1986). For example, feral pigs accelerate the spread of strawberry guava (Psidium cattleianum) into native wet Metrosideros forests above Hilo on the island Hawai'i, leading to a strawberry guava monoculture with virtually no native plant species (Englund et al. 2002). This same pattern of susceptible populations being displaced by invasive species has been repeatedly observed not only in Hawai'i, but also in other vulnerable areas of high endemism including isolated islands like New Zealand or continental areas such as Australia and South Africa. A clear example of an island insect group impacted by invasives would be the largest native insects on New Zealand, the large bodied (up to 40 g) orthopteran weta in the Families Anostostomatidae and Rhaphidophoridae, a group currently occupying a broad range of modified and unmodified habitats. As is the case with many island species, anthropogenic habitat disturbance does not appear to be a major cause of weta extirpation, as extinctions occurred long before the onset oflandscape changes (Gibbs 1998). Most of the very large weta species such as Deinacrida rugosa Buller and Deinacrida heteracantha White were extinct on the mainland North Island prior to 1900, where they formerly coexisted with native avian and reptilian predators, many of which are also now currently endangered or extinct (Gibbs 1998). Introduced in the past 200 years, the main threat to weta today are the Norway rat (Rattus norvegicus, Berkenhout) and the ship rat (Rattus rattus, (Linnaeus, 1758». Such land mammals are not native to New Zealand, and are more effective predators than the now rare and extinct native predators, especially because weta have a strong olfactory presence and are easily detected by mammals (Gibbs 1998). While some of the smaller, more aggressive weta species survive, albeit in lower densities, in areas with rats, the larger and more docile weta species have been completely eliminated except on island refugia lacking rats (Gibbs 1998). One undescribed species of weta survives today by feeding upon and living in hedges ofinvasive introduced gorse (Ulex 156 europaeus Linneaus) lining pasture areas, with the gorse thickets apparently providing protection from introduced predators because of the thick thorn layers. This weta species has never been found in the nearby indigenous forests that apparently do not provide sufficient refugia from rat predation (Gibbs 1998). Situated on the far southeastern tip of South Africa, the Cape Floristic Region with its Fynbos biome exhibits an island-like high endemicity for many plant and insect species (Wright and Samways 1998) that are vulnerable to invasive species. Although overall little research has been published on the impacts of invasive species on aquatic insect populations, the South African Odonata are one of the few exceptions (Samways 1995; 1996; 1999) and because of this are well worth examining. With 155 species of Odonata of which 6.5% are threatened, and a moderately high endemism rate of nearly 19%, South Africa contains a rich fauna, including species with narrow habitat ranges making them vulnerable to extinction (Samways 1999). Of the 29 endemic South African Odonata taxa, the most restricted species are mainly from the Eastern and Western Cape Provinces, and of the 10 highly threatened species, 6 are damselflies and 4 are dragonflies. Threats to the South African endemics include habitat disturbance and invasive species. Samways (1995; 1996; 1999) observed that a rare and threatened damselfly (Ecchlorolestes peringueyi Barnard) living in clear upland streams in the southwestern Cape area may have had its range restricted by introduced rainbow trout (Oncorhynchus mykiss Walbaum). This Gondwana relict damselfly was found only in stream reaches above waterfalls that rainbow trout could not ascend, and the larvae appear to be behaviorally susceptible to trout predation as they crawl on the surfaces of rocks and plants. A dragonfly species, Syncordulia gracilis Burmeister, also found in the Western Cape province, was similarly cited as being under pressure from introduced rainbow trout (Samways 1999). The policy of the conservation authorities for many years was to encourage trout in those South African streams for recreational fishing; this policy has recently been changed, and the trout are now considered to be invasive species, as are another sportfish introduction, smallmouth bass (Micropterus dolomieui Lacepede) (M. Wright, pers. comm.). 157 Overall, impacts of invasive species on the South African Odonata appear to be relatively small, with only 1.3% of the fauna known to be suffering negative effects (Samways 1999). The situation of introduced rainbow trout impacting native Odonata species in South Africa contrasts with that of Hawai'i, where neither extinctions nor range contractions were found in trout streams containing endemic damselflies (Englund and Polhemus 2001). This is hypothesized to result from the restricted range of trout in Hawai'i because of thermal limitations, and also possibly habitat segregation, given that trout and Hawaiian damselfly larvae occur in different habitats. Trout were mainly found in deeper pools while damselfly larvae inhabit waterfall faces and cascades and appear to rarely enter into the stream drift (Englund and Polhemus 2001). By contrast, other species of invasive fish in Hawai'i such as mosquitofish in the family Poeciliidae have been found to be much more serious threats to Hawaiian damselflies (Englund 1999). Other regions where the impacts of introduced fish on native Odonata have been examined include New South Wales, where a native Australian dragonfly, Hemicordulia tau Selys Longchamps, was found to be an important component in the diet of brown and rainbow trout in Lake Eucumbene (Faragher 1980). However, the long-term survival of the dragonfly population in Lake Eucumbene was not believed to be affected by trout predation because of seasonal changes in trout prey composition and density, as well as varying lake levels (Faragher 1980). Trout have also been extensively stocked throughout New Zealand, and reproduce naturally in both the North and South Islands, having been shown to negatively impact native stream fish populations (Townsend and Crowl 1991; Crowl et aI. 1992; Townsend 1996). Several New Zealand studies have shown that the presence of trout reduced native mayfly biomass and caused behavioral changes (McIntosh and Townsend 1994; 1995), forcing mayflies to graze algae less efficiently at night resulting in increased algal biomass in streams (Flecker and Townsend 1994). Although invertebrate biomass was reduced in streams containing trout as compared to areas containing only native galaxiid fish (McIntosh and Townsend 1996), extirpations or range reductions of aquatic insects were not shown, in contrast to well-documented range contractions caused by trout to the New Zealand native fish fauna. 158 SUCCESS AND FAILURE IN CONTROLLING HARMFUL AQUATIC INSECTS AND INVASIVE SPECIES IN HAWAI'I AND FRENCH POLYNESIA French Polynesia The potential difficulty ofeliminating established invasive species is well illustrated by programs aimed at controlling species affecting human health, particularly blackflies and mosquitoes. In French Polynesia, attempted control programs have been mostly unsuccessful in reducing native biting aquatic fly populations of such problematic species as nona noir (Simuliidae) and nono blanc (Ceratopogonidae). By far the most serious of these biting flies is Simulium buissoni Roubaud, a species endemic to Nuku Hiva and Eiao Islands in the Marquesas (Craig et a1. 2001). Simulium buissoni bites any endothermic animal and makes life virtually unbearable for humans living near stream and wetland areas. Attempts to control of disease vectoring species such as the mosquito Aedes polynesiensis Marks and other nuisance native aquatic insects found in French Polynesia began in 1970. The first recorded effort involved a native carnivorous fish, Kuhlia rupestris, in attempts to eliminate filariasis from the Marquesas Islands (Sechan et a1. 1998). Efforts were then made in Nuku Hiva to unsuccessfully replace the anthropophilic S. buissoni with non-biting simuliid species (Sechan et a1. 1998), although the species name and original location of the replacement black fly specie(s) was not mentioned. After these fruitless attempts at biological control, a shift to insecticides was made to eradicate S. buissoni in the Marquesas, and trials of different insecticides showed that temephos (Abate™) exhibited the greatest promise (Sechan et a1. 1998). In 1986, the entire Taiohae watershed of Nuku Hiva was treated with temephos with "excellent results" (Sechan et a1. 1998). By 1993 the government of French Polynesia sponsored a black fly eradication project for S. buissoni on the entire island of Nuku Hiva (Craig et a1. 1995; Sechan et a1. 1998). From January to April 1993, temephos was added to all flowing rivers on Nuku Hiva on a bi-weekly basis. Populations of biting female S. buissoni were reduced to 4% of previous levels after the first two applications of insecticides (Craig et a1. 1995). However, the eradication plan was cut short by heavy rains in March 1993. 159 By October 1993, S. buissoni populations had increased to pre-treatment levels (Craig et al. 1995). This failed effort cost nearly U.S. $200,000 and has not been repeated on Nuku Hiva or elsewhere. The Hawaiian Islands provide one ofthe few examples ofa successfully controlled and eradicated invasive aquatic insect species, the mosquito Aedes aegypti (Linnaeus). This species was responsible for the great dengue fever epidemics in Hawai'i during World War II and earlier, as documented by Hardy (1960). Aedes aegypti is a domestic species and preferentially breeds in artificial containers holding clean water, such as old tires, flower vases, beverage containers and other urban debris, and was formerly found in great numbers in urban Honolulu. However, the yellow fever outbreak of 1911 in the crew of a visiting vessel from Mexico led to drastic mosquito control measures around the city. Starting in 1911, banana plants within the city were eradicated as possible breeding sites, and other strict control measures by the State Board of Health were implemented that eventually led to the last A. aegyptibeing collected on O'ahu in 1957. It is probable the control efforts in the 1950s were successful while earlier efforts were not because of the diligent and continuous long term cleanup operations conducted by the State Board of Health. The last statewide collection of Aedes aegypti was in1971 based on BPBM collection data and subsequent surveys in the 1970s and 1980s (Evenhuis in litt.). Although Aedes albopictus (Skuse) was responsible for the 2002 dengue fever outbreak on Maui and O'ahu, A. aegypti is a far more efficient dengue vector than A. albopictus. It is quite likely the recent dengue fever outbreak would have been much worse if A. aegypti had not been eliminated from the Hawaiian Islands. Overall attempts to control and eliminate invasive species in Hawai'i and elsewhere in the Pacific islands have been successful in relatively few cases. Elimination offeral mammals such as rats and ungulates has been successful mostly on smaller, dry islands (Table 2). Although the results shown in Table 2 indicate only a few invasive animal species have been completely eradicated from the Hawaiian Islands, it provides further justification for intensive management of invasive species in small, discrete areas such as offshore 160 islets and the Northwest Hawaiian Islands. Perhaps the successes oferadicating vertebrates from these islands can also be applied to other isolated, non-overlapping habitats or remnant biological islands within Hawai'i. The same may apply to many freshwater streams, anchialine pools, and wetland areas that also represent disconnected insular habitats. Many streams on Hawaii's rocky Hfunakua coast include no estuary habitat and enter the ocean as waterfalls. These areas could be described as isolated islands of habitat similar to many anchialine pond areas on the islands of Hawai'i and Maui. Because of this disconnection, these areas would be suitable for removal of invasive fish species. In fact, anchialine pools on Hawai'i Island containing alien fish can be observed just meters away from ponds lacking alien fish (Englund, unpubl. data; Brock and Kam 1997). Table 6.2. Successfully eradicated invasive animal species in the Hawaiian Islands. Island!Area Year Year Eradicated or Extirpated Species Established (Reference) Mammals European Rabbit Laysan1 1902 1923 (Rauzon 2001) Pearls and Hermes <1916 1928 (Tomich 1986) HaleakaHi National Park, Maui 1989 1991 (Bishop Museum website) Manana 1890 2002 (D. Smith, pers. com.) Rats Rattus rattus Midwav/Sand Island 1944 1996 (Rauzon 2001) Rattus exulans Kure Atoll <1778? 1993 (Rauzon 2001) Goats, Sheep Kahoolawe early 1800s 1988 (HEAR website)* Axis Deer O'ahu > 1868 1950 (Tomich) Insects Aedes aegvvti O'ahu 1895 1957 (Hardv 1960) 1A rabbit colony at nearby Lisianski Atoll caused its own demise after destroying all vegetation on that island before 1923 (Tomich 1986). *[http://www.hear.org/naturalareas/kahoolawel] 161 DISCUSSION Lessons Learned from Successful Eradication Programs As stated earlier, successful eradication of invasive species has been decidedly erratic in the Pacific, with small islands being the primary area of notable success. Table 2 lists the invasive animal species successfully eradicated from the Hawaiian archipelago, but there have also been successful eradications of a variety of mammal and insect species from a number of other islands in the Pacific and elsewhere. Vertebrates such as black rats (R. rattus), Norway rats (R. norvegicus), European rabbits (Oryctolagus cuniculus) and the house mouse (Mus musculus Linneaus) have been completely removed from a number of New Zealand islands (Myers et a1. 2000). Harmful insects such as the tse-tse (Glossina spp.) were eradicated from the island of Principe in the Gulf of Guinea (Simberloff 2003), as was the screw-worm fly (Cochliomyia hominivorax (Coquerel)) from Curacao (Baurnhover 1955), and the Asian citrus blackfly (Aleurocanthus woglumi Ashby) from Key West, Florida (Hoelmer and Grace 1989). Only a short list exists for insects eradicated in the Pacific region. This includes the Oriental fruit fly (Bactrocera dorsalis (Hendel)) from Rota, Guam, Tinian, and Saipan (Steiner et al. 1965; Steiner et a1. 1970), and also from Okinawa (Tsubaki 1998). The tussock moth (Orgyia thyellina Butler) was eradicated from New Zealand (Hosking et a1. 2003); two species of ants, Pheidole megacephala and Solenopsis geminata (Fabricius) from 33 hectares of Kakadu National Park, Australia (Hoffman and O'Connor 2003); and the melon fly Bactrocera cucurbitae (Coquillett) from the Ryukyu Archipelago (Kuba et a1. 1996). Ongoing monitoring programs are critically important for successful eradication of incipient populations of invasives. This is because newly introduced species are found in localized areas and this limited distribution greatly increases the chances of eradication (Myer et a1. 2000). For example, in 1999 the black-striped mussel (Mytilopsis sallei Recluz) was eliminated from an enclosed marina in Darwin, Australia by a rapid response involving high doses of copper sulphate (Bax et a1. 2002). Christmas Island (Indian Ocean) has been severely impacted by invasive crazy ants (Anoplolepis gracilipes). An infestation of some 2500 hectares has resulted in an estimated 15-20 million land crabs being killed by these ants (Australian 162 Government Website 2004). This ant invasion led to a total forest ecosystem disruption as weed seedlings previously eaten by crabs were allowed to grow. The ants were also responsible for an explosive increase in the scale insect population feeding on native trees, which in turn further stressed the native forests. A control program in place on Christmas Island since 2000 has achieved some initial successes and significantly reduced the size of the crazy ant super colonies (Australian Government Website 2004), and it may be possible to ultimately eradicate them from the island. The ant eradication program at Kakadu National Park and ongoing ant eradication efforts on Christmas Island are relevant for managing and preserving native biodiversity at HaleakaHi National Park, other critical habitats for native arthropods in Hawai'i, and small island areas containing invasive ant species such as Palmyra Atoll (see page 150, Overview of Invasive Ant Impacts on Hawaiian Arthropods). It shows that ant eradication is feasible, and also points out the devastating ecosystem impacts of ants on native species if they are not controlled. Previous ant eradication success stories illustrate that coordinated efforts are needed to stop invasions at a relatively early stage, such as at Kakadu National Park. In that reserve, Pheidole megacephala colonies were first detected in mainly developed areas in June 2001. By October 2001 mapping of ant distributions throughout the park took place. Treatment with Amdro ant baits commenced at the end of October and into November 2001, prior to the beginning of the wet season. Post-treatment surveys indicated eight small remnant populations that remained within buildings, which were then treated with additional baits. After 17 months (April 2003) the ant colonies had been eliminated (Hoffman and O'Connor 2004). The Kakadu example of an immediate response followed by post-eradication monitoring has been a notable success. It is also important to note that when an invasion is past the point of eradication scarce resources should not continue to be poured into a hopeless control effort. Simberloff (2003) stated the failed U.S. $200 million program to eradicate the red imported fire ant (Solenopsis invicta Buren) in the southern U.S. is considered the "Vietnam" of eradication programs. This because it was a non-winnable situation and the widespread use of heptachlor for fire ant controlled to non-target impacts such as death of wildlife, cattle, and even predators offire ants (Simberloff 1997b). 163 The failed blackfly control effort on Nuku Hiva also provides additional perspective on eradication of undesirable species. For instance, prior to launching any eradication project it is important that the biology of the target species be well-known and that potential non-target impacts are evaluated. On Nuku Hiva, temephos was considered target-specific to black flies because Odonata and chironomid larvae were reportedly found alive both before and after treatments of streams. Sechan (1993) did not "notice mortality of other invertebrates associated with Simulium buissom", in spite of the fact that quantitative toxicity tests were conducted only on freshwater prawns; these studies found that native prawns (Macrobrachium spp.) were not harmed by concentrations of temephos used during field treatments (Sechan 1993). The authors of this black fly eradication project stated that "the taxa [in Nuku Hiva streams] have been identified which do not differ from that reported in Moorea, in the Society Archipelago" (Sechan et aI. 1998). However, the freshwater fauna found in Moorea, Society Islands, and in Nuku Hiva, Marquesas Islands are dissimilar (Keith et aI. 2002) as each French Polynesian island, similar to the Hawaiian Islands, has its own suite of endemic aquatic insects (Polhemus et a1. 2000, Englund 2003). Recent joint Smithsonian/Bishop Museum expeditions to Nuku Hiva in 1999 and 2001 have found many undescribed species of native aquatic insects, with descriptions of these new species just beginning (Evenhuis 2004). For example, a radiation of at least 5 undescribed, large, endemic damselfly species on Nuku Hiva was recently recorded (Polhemus et aI. 2000) along with two new species oflarge, endemic aquatic flies (Dolichopodidae) from that island (Evenhuis 1999). Evenhuis (2004) also described an additional three new species ofendemic aquatic Dolichopodidae from elsewhere in the Marquesas. Thus there was the potential for serious negative impacts to the endemic aquatic fauna in Nuku Hiva streams because of the lack of biodiversity surveys and toxicity data on aquatic organisms other than Macrobrachium prawns before the implementation of the control program for Simulium buissoni. Other efforts to control biting flies such as the introduced nona blanc (ceratopogonids) in the Marquesas have been more environmentally benign and involve the construction of seawalls in populated areas such as 164 Atuona on Hiva Oa and Taiohae on Nuku Hiva. These seawalls reduce the amount of available brackish water habitat favored as breeding sites for the nono blanc along beach areas by holding back the freshwater lens. Although there are no published accounts or studies on the efficacy of these freshwater retaining seawalls, many local residents claim they do work, and anecdotal observations of biting nono blanc in the areas where the seawalls have been installed indicate relatively low-levels of these ceratopogonids (Englund, pers. observ). Impacts on estuarine organisms from seawall construction are unknown. Threats and Opportunities in Conserving Native Aquatic Species in the Tropical Pacific Freshwater species and their habitats are currently suffering severe negative impacts worldwide (Saunders et a1. 2002) with invasive species and water diversions being the primary threats to such systems on tropical Pacific islands. Most water diversions in island groups other than Hawai'i, such as those in French Polynesia, are in the lower stream reaches and thus cause relatively minimal impacts. While little can be done in the long-term to reduce water diversions for municipal and agricultural uses on these islands, the threat of invasive species can still be addressed. An optimal conservation plan for freshwater organisms therefore needs to not only discourage the spread of invasive species, but also to target watershed protection and include freshwater protected areas to provide whole-catchment integrity (Saunders et a1. 2002). Although reserve designation does not automatically guarantee protection, it does at least provide legal and political recognition regarding the importance of watersheds and some regulatory mechanisms to deal with current and future threats. In some respects, the conservation of tropical Pacific island streams such as those found in Hawai'i and French Polynesia will be simpler than trying to protect and conserve much larger continental systems. Because of the relatively recent volcanic origin of these islands, these watersheds are much shorter in length, less integrated and have steeper topographic profiles (Craig 2003), allowing for an easier delineation of specific watersheds. Conversely, although these watersheds may be easier to protect because of their small 165 size, they are also more vulnerable to disturbance because of this same compact nature. In islands throughout Hawai'i and French Polynesia the entire watershed from the ocean to its mountain headwaters is often only a few kilometers long, and can be quickly comprised by invasives. Streams that are separated by steep topography often contain different suites ofspecialized and endemic aquatic insect species (Englund and Polhemus 2001), thus each stream should be viewed as a separate ecological and conservation unit. Potential Conservation Measnres for Rapa, French Polynesia Current opportunities in French Polynesia to protect both terrestrial and aquatic taxa include governmental officials from the Delegation 11 la Recherche working with indigenous people to protect native plants and habitats from feral ungulates through fencing. Rapa, in the Austral Islands of French Polynesia is a small 2 (40 krn ), 65O-m high island located at 27°S, with a temperate warm climate, and provides an example of the difficulties in achieving conservation goals in remote areas. Despite its small size, Rapa has a diverse flora and fauna, including many island endemics having large adaptive radiations, such as the Miocalles weevils (Paulay 1984), certain Lepidoptera (Clark 1971), a remarkably species-rich land snail fauna (Solem 1982; Fontaine and Gargominy 2003), and a highly diverse endemic flora (Florence 1997; Meyer 2003; Meyer et al. 2003). Working with the people of Rapa, the French Polynesian government is taking conservation steps to protect its most valuable biological and cultural assets: the cloud forests and dry forests. Recent biodiversity surveys on Rapa (Englund 2003; Fontaine and Gargominy 2003; Meyer 2003) funded by the French Polynesian Government have provided insights into potential conservation measures to help ensure that the unique biodiversity of this small, remote island is not lost. The primary invasive species problem on Rapa is the presence of feral cows, goats, and horses that have denuded and destroyed all but a small portion of the high elevation cloud forest. The ungulate problem has significantly worsened since the 1980s, with the horse population apparently increasing from one in 1980 (Paulay 1985) to a substantial herd of more than 50 that were observed in the lower Agairao Valley alone in 166 December 2003 (Englund 2003). Only a few high summit areas containing the original undisturbed cloud forest and middle elevation moist forest area survive (Meyer 2003). The central volcanic mass on Rapa is Mt. Perau, containing the last area of middle-elevation moist forest and cloud forest, and is important from a global biodiversity perspective (Paulay 1984, Clarke 1971). The native vegetation on Mt. Perau remains extant mainly because the slopes of Mt. Perau are generally steep, up to 800 (Paulay 1984), making sampling and conservation efforts in this area quite difficult, but also limiting access by ungulates to many areas around the summit. Current efforts are focused on saving this globally significant and biologically valuable high cloud forest. Rapa is a wet island and receives an average of 2500--3000 mm of rain at sea level per year but undoubtedly is much wetter in the cloud-covered mountains (Paulay 1985). It has many relatively large stream systems for an island ofits size, and these streams are still completely free of any introduced aquatic species (Englund 2003). Despite its small size, Rapa has an endemic aquatic fauna of worldwide interest. For example, the endemic damselfly Ischnura thelmae is the world's largest (up to 34 mm in length) in that cosmopolitan genus (Lieftinck 1966). Although common in the 1960s (Lieftinck 1966),1. thelmae now appears to be seriously threatened not by introduced alien fish species, such as what has occurred in Hawai'i (Englund 1999), but by riparian forest losses (Englund 2003). This species seems to be an obligate forest dweller; it was never observed during thorough surveys along riparian habitats on Rapa in overgrazed stream and open pasture areas, and was only found in heavily forested areas (Englund 2003). A significant observation was that Ischnura thelmae forages long distances away from streams in areas of native forest. Thus, there was a clear link between the condition of the native forest and the health of the native damselfly populations. Most of the native terrestrial insect biodiversity remaining on the island of Rapa is found in a narrow zone of native forest between 500--650 m at the summit of Mt. Perau (Englund 2003) that is estimated now to 167 be no more than 20 ha in size (J.Y. Meyer, pers. comm). The collection of many undescribed species from Mt. Perau in December 2002 by Englund (2003) illustrates that much remains to be discovered about the insect fauna from this mostly intact native forest area. The cattle grazing line at the summit of Mt. Perau starts at about 370-400 m elevation, and cattle were visibly trampling down Freycinetia sp. to gain further access up into the summit areas, with evidence of goats found near the very summit. In 2002, goat damage was observed as high as the 550 m elevation on the ridgelines of Mt. Perau. Habitat loss caused by feral ungulates clearly requires immediate actions to preserve the cloud forest of Mt. Perau. On Rapa, knives are still used to hunt cattle and goats. As authorities have strictly controlled firearms and ammunition, and hunting with knives is very inefficient, hunting will not be a short- or long term answer to control the large ungulate populations, which were estimated to number 500 cattle and 5,000 goats in 1984 (Paulay 1984). The permanent population of Rapa was 497 people in 1996 (Recensement General de la Population website) thus the amount of livestock present on the island far exceeds what could be consumed locally. The remaining option is to fence ungulates out of the Mt. Perau summit area and to work with and employ local residents to implement this plan. Dry forest areas are even more imperiled than cloud forests on Rapa and a small patch (1-2 ha) containing rare plant species at Pariati Bay will be fenced to exclude livestock as soon as funding becomes available. Plans are also currently being put in place to start a fencing project for the Mt. Perau area sometime after the Pariati Bay dry forest is fenced (J.Y. Meyer, pers. comm.), which will be a major step in protecting the high cloud forest region. Because relatively few invasive plant species are found in Rapa, fenced off areas should regenerate quickly with native species. The protection of this terrestrial ecosystem will have the added benefit of protecting streams flowing off Mt. Perau. Saving this unique terrestrial ecosystem will thus lead to the preservation of aquatic habitats and their associated aquatic insects, including Ischnura thelmae. 168 RECOMMENDATIONS Recommendations for the Conservation and Restoration of Island Insects This section briefly reviews previous recommendations regarding the conservation ofisland insects by limiting new species invasions, and provides new recommendations based on the research of this dissertation. Constant vigilance is needed to effectively keep invasive species from becoming established. Snake interceptions and captures in Hawai'i from 1990-2000 (Kraus and Cravalho 2001) are a good example of the concerted efforts required to keep out undesirable species entering as smuggled pets or via commerce. Even though in theory it may be easier to control incipient invasions of highly vocal animals, such as the coqui frog (Eleutherodactylus coquiThomas) in Hawai'i, because they are simpler to detect than the usually more cryptic invaders, a political and regulatory framework still needs to be in place at the time of invasion to allow eradication at an early stage. New freshwater fish invasions may be more difficult to detect than the more visible animals such as reptiles or amphibians because of the great number of established alien tropical aquarium fish throughout Hawai'i (Englund and Eldredge 2001), combined with relatively few competent observers monitoring the wide range of Hawaiian freshwater habitats. Terrestrial and aquatic insect invasions are even more difficult to detect and manage because of the small and inconspicuous nature oflarval and adult stages. Because of these factors, insects have the greatest rate of yearly establishment of all animal or plant groups in Hawai'i (Staples and Cowie 2001), with 2,782 established nonindigenous insects occurring in Hawai'i (Eldredge and Evenhuis 2003) and becoming established at an alarming yearly rate of 20-30 species (Howarth et a1. 2001). Eradication efforts for invasive insects are also problematic because after their initial detection they have usually already undergone a population explosion and often are found in high densities across a wide range of habitats. An ideal conservation strategy would ofcourse attempt to prevent invasions from ever occurring, but it is of interest to examine the recommendations of Howarth and Ramsay (1991) which still provide a valid set of solutions regarding the conservation of island insects and the habitats upon which they depend. The guidelines included a comprehensive and integrated program ofresearch and monitoring, education, reserve 169 management, legal and legislative actions, controlling the introduction of alien species, and pest control programs. While these suggestions provided a solid framework, many of the recommendations have unfortunately not been implemented, resulting mainly from a lack of funding attributable to a lack of political support. The recent publication of the State of Hawai'i Aquatic Invasive Species Management Plan (Shluker 2003) provides a much-needed comprehensive framework to prevent or at least reduce future problems caused by invasive aquatic plant species, such as the Salvinia molesta infestation in Hawai'i that cost U.S. $1.25 million to remove in 2003 (Gima 2003). Because this plan involved all interested stakeholders (aquarium, aquaculture, and shipping industries), and resource managers on each island from the very beginning (Shluker 2003), it has a reasonable chance of success. Other recent developments, such as the formation of the various Invasive Species Councils for each main Hawaiian Island (e.g., Big Island Invasive Species Council, etc.), have shown that grass-roots efforts at controlling invasive species can be successful if invasions are detected at an early phase. Examples of this include the nearly complete eradication of the invasive tree Miconia calvescens on O'ahu and the early elimination of many other incipient alien plant invasions on Maui Nui (F. Starr, pers. com.). Of course, interception and early detection of invaders is by far the most-cost effective manner to deal with invasive species, but even if enforcement and quarantine resources were unlimited there would still remain a need for the capacity to eradicate incipient invasions given the daily volume ofcommerce and visitor arrivals to Hawai'i. Because so many indigenous aquatic insect taxa, particularly in Hawai'i but also elsewhere in the Pacific, now have seriously reduced ranges (Englund 1999,2001,2003), the ultimate goal of biodiversity preservation should be to restore populations to a level robust enough to allow species to withstand major environmental disturbances such as hurricanes or droughts. Population restoration should at first involve small, discrete habitat units that can be permanently cleared of the invasive species identified to have caused declines. Native species should then be translocated and reintroduced to areas having either natural or constructed dispersal barriers to prevent re-invasion by the problematic alien species, but this should only be done within the same island or nearby islands that have identical or very similar taxa. In some cases 170 active management to protect aquatic insects may be required, such as ant monitoring and control for the protection of terrestrially breeding damselfly species such as Megalagrion nesiotes or M williamsoni. Recommendations for Successful Invasive Species Removal Attempts to remove invasive aquatic species need to be well planned, and a thorough understanding of their ecology and life history is necessary to ensure success. For example, treatment of lower Kane'ohe Bay streams in O'ahu, Hawai'i, with a ichthyocide such as rotenone to eliminate invasive fishes would ultimately be unsuccessful because low salinity estuarine regions provide refugia for several species of salinity-tolerant poeciliids and tilapia that are currently harming native aquatic life. The large size of Kane'ohe Bay and potential public outcry over treating the entire bay would certainly preclude the effective use of rotenone in this region. Isolated and unconnected stream habitats are common on geologically younger islands in volcanic archipelagoes such as Hawai'i or the Society Islands. Because many aquatic habitats such as high gradient stream areas entering the ocean as terminal waterfalls include little to no estuarine habitat, there would be a higher probability for aquatic ecosystem restoration once detrimental alien species such as poeciliid fish were removed, similar to what has occurred in the successful eradications listed in Table 2. Anchialine ponds provide another example of a discrete habitat where full ecological restoration is not only feasible, but has actually taken place with a few small-scale experiments on Hawai'i Island. Brock and Kam (1997) conducted pilot studies in several anchialine ponds where they found the only viable alternative to remove alien fish was through the use of an ichthyocide like rotenone. They found native biota recolonized treated ancialine pools shortly after the alien fish were removed, and that full ecological restoration took at least one year. Their findings have significant implications in restoration of aquatic ecosystems, and they emphasized that their greatest restoration success was in isolated, smaller pools where invasive fish had no chance to recolonize or obtain refugia during the application phase of the ichthyocide (Brock and Kam 1997). Anchialine ponds are one of Hawaii's most endangered ecosystems, supporting native damselflies 171 such as Megalagrion xanthomelas and other aquatic insects. Even in the highly protected Kaloko Honokohau National Historic Park on Hawai'i Island, 84% of these habitats were infested with invasive fish such as guppies in 1997 (Brock and Kam 1997). Recommendations for Regular Monitoring to Detect New Invasions For Hawaiian aquatic systems there is currently no regular monitoring program in place to detect new invasions of aquatic biota. Regular monitoring is necessary for managers and researchers to detect and find new invasions at an early stage. Although staff of both the Hawaii Division of Aquatic Resources and the Hawaii Biological Survey of the Bishop Museum conduct numerous surveys throughout the year, these are mainly on a contract, project, or as-needed basis. The monitoring program to detect early invasions of aquatic species initiated in the year 2000 at Pelekunu Stream could serve as a statewide model. The Pelekunu Nature Preserve, Moloka'i contains large, free-flowing streams that are a refuge for some of the rarest aquatic animals in Hawai'i and the world, and is one of the last areas in Hawai'i lacking alien aquatic vertebrates of any kind. The Nature Conservancy Hawai'i, Moloka'i office, along with cooperating scientists from Bishop Museum's Hawaii Biological Survey and Hawaii Division of Aquatic Resources, conducted surveys from 2000-2002 to monitor the status of certain native aquatic species in this system. Aquatic invertebrate monitoring was conducted in conjunction with endemic freshwater fish monitoring to provide information to help effectively manage and preserve native aquatic biodiversity. The initial monitoring of Pelekunu Stream provided extremely valuable information (Englund 2000, 2001b; Englund and Arakaki 2003), but because of lack of funding has not been conducted in either 2003 or 2004. Because alien fish and other invasive aquatic species continue to rapidly spread throughout the Hawaiian Islands, this monitoring program should be reinstated as a matter of priority. In addition to regular monitoring, a contingency or rapid-response plan should be drawn up that would immediately eliminate any alien aquatic vertebrate species accidentally or intentionally introduced into Pelekunu Stream. For example, immediate chemical treatment of the stream with rotenone should occur if 172 any introduced fish species were detected during monitoring. Although this would also eliminate most native aquatic invertebrates in the treated areas, recolonization from nearby areas should be immediate, and harm would be short-term and inconsequential compared to the much longer-term threat of alien fish. By contrast, lack of action when fish or other major alien species introductions occur into the Pelekunu watershed would deal a severe blow to the preservation of native Hawaiian aquatic fauna biodiversity. Currently, high mountain ridges keep introduced amphibians out ofPelekunu Stream, even though bullfrogs (Rana catesbeiana Shaw) are found in the adjacent Wailau Stream watershed. Bullfrog control at Pelekunu Stream would be much more difficult than eliminating invasive fish species, and research into controlling and reducing frog numbers in neighboring Wailau Stream should be undertaken to alleviate this threat. The necessity of early detection of invasive aquatic species underscores the importance of periodic aquatic monitoring in the Pelekunu watershed. Similarly, regular monitoring of a select number of waterbodies on each main Hawaiian island could have been accomplished for a small fraction of the cost required to remove invasive Salvinia molesta from Lake Wilson on 0'ahu in 2003. As the above example of Pelekunu Valley demonstrates, in tropical insular Polynesian streams even small areas devoid ofalien species currently provide a last refuge for native aquatic species, and are of great conservation value. The most dramatic illustration of this is the 95-meter section of an unnamed tributary at TripIer Army Medical Center containing the last O'ahu population of Megalagrion xanthomelas. The continued precarious existence of M. xanthomelas there provides the impetus for translocating to a suitable alien-free habitat and restoring this formerly common damselfly species (Englund 200la). The case of M. xanthomelas at TripIer indicates endemic aquatic biota can survive in extremely restricted habitats for up to 90+ years after the introduction of poeciliid fish to an area (Englund 1999). Although this is not an optimal situation for the long-term survival of this species, it does provide an indication of the resilience of island species when even small amounts of habitat lacking invasive species are available. The TripIer habitat is relatively secure because of its location inside a U.S. military facility, but other similarly restricted habitats for endangered aquatic organisms are in need of similar protections, such the last known Megalagrion 173 nesiotes population (Englund, unpub. data) occurring along a small reach of East Wailua Iki Stream on Maui next to a major highway. Preservation of Aquatic Biodiversity through Invasive Species Prevention and Watershed Preservation A prudent recommendation would be that at a minimum, one relatively intact watershed from each island in an archipelago should be selected as a biodiversity reservoir for aquatic species. Specialized habitats in island areas are of particular concern as they may be considered islands within islands, and many have restricted populations ofendemic biota. It is important to have one reserve on each island within a given archipelago, because numerous single island endemics have been found during surveys of the few island groups so far examined in detail, such as Hawai'i (Polhemus and Asquith 1996), the Marquesas (Evenhuis 1999,2003; Polhemus et al. 2000), and the Australs (Englund 2003, 2004). Such reserves would also be beneficial by providing baseline sites as control areas for assessing changes in other watersheds within an island or island group. Realistically, funding for surveys and taxonomic expertise will remain limited for the foreseeable future, therefore efforts to preserve native species by providing a reserve protection framework may end up being implemented in the absence of such baseline information. Measures such as providing at least one watershed preserve per island could be instituted throughout tropical Polynesia in areas such as Hawai'i, the Marquesas, and the Austral Islands where there are currently sizable uninhabited valleys containing significant streams. These valleys historically supported large Polynesian populations cultivating taro (Kirch 1985), but today are nearly or completely uninhabited. Notable examples in Hawai'i would include virtually all north shore Moloka'i streams; Hanakoa, Nualolo, and Kalalau Valleys on Kaua'i; the Hakaui Valley on Nuku Hiva in the Marquesas; and virtually the entire island of Rapa in the Austral chain. The native aquatic fauna of these areas remains largely intact because it has not sustained measurable long-term impacts from historical Polynesian taro cultivation and settlement and these areas were subsequently depopulated and neglected after European colonization. For example, the 174 north shore Moloka'i streams, used for taro cultivation by ancient Polynesians, now contain some of the rarest aquatic species in Hawai'i because they lack invasive species (Englund and Arakaki 2003) that spread elsewhere in Hawai'i after European contact. The stream areas mentioned above present opportunities for the creation of freshwater reserves in lightly populated regions with little corresponding human conflict. The Nature Conservancy's Pelekunu Preserve on Moloka'i provides the best example in tropical Polynesia of an intact, pristine watershed formerly heavily cultivated for taro that has been abandoned and is now uninhabited, yet highly protected. In order to preserve the rare native aquatic species found there, the Pelekunu Preserve is now intensively protected from feral ungulates with the support and assistance of local Moloka'i hunters, and as previously mentioned was monitored for invasive aquatic species. Because of the severity of the invasive species threats, it is urgent that conservation actions are implemented without delay even if comprehensive taxonomic treatments of the fauna found on each archipelago or individual island within an island group are unavailable. Limited knowledge ofPacific island biotas is the norm, and there are presently few taxonomists available to describe the many new species found within each archipelago. Because ever-increasing global trade is leading to the rapid spread of invasive species, and human population increases are leading to greater resource demands, it is necessary to protect the relatively intact watersheds in the tropical Pacific as soon as possible. Hawai'i clearly demonstrates that even in a relatively well-studied tropical island system, new taxa are constantly being discovered. For example, eight undescribed native aquatic insect species were found during recent helicopter-accessed surveys of 14 remote stream areas on three Hawaiian islands (Englund et a1. 2003). These findings were noteworthy as a literature review from the period 1990-2003 revealed an average ofonly 0.9 new aquatic insect species described per year from the Hawaiian Islands (Englund et a1. 2003). This, despite the fact that aquatic insect collection efforts had been unusually high in Hawai'i throughout the 1990-2003 period. Additionally, this example illustrates how little basic information is known regarding the numbers and types of aquatic insect species in Hawaiian inland waters, let alone their basic ecological, evolutionary, or life history parameters; this lack of knowledge is even more acute in the remainder of the Pacific. In the best-case scenario of 175 planning ecological reserves to preserve native biodiversity, a systematic inventory of the aquatic insect fauna for all major Pacific high islands would occur ftrst, with conservation priorities and candidate reserve areas subsequently delineated for highly diverse, sensitive, and intact watersheds. Realistically, however, the preservation of aquatic ecosystems prior to the full or even partial delineation of all their component taxa is necessary to avoid future large-scale biodiversity losses. Ifonly limited funds are available then rapid biodiversity surveys targeting a select number of indicator species from each proposed area (Howarth 1990) would have great value in prioritizing conservation agendas. In the tropical insular Paciftc, damselflies (Zygoptera) (Polhemus and Asquith 1996) and certain aquatic Heteroptera such as Veliidae (Polhemus and Polhemus 2004) or Saldidae (Cobben 1980), could serve as indicator species sensitive to disturbance in a wide of range aquatic habitats. Damselflies are also striking in appearance and charismatic enough to be easily observed and collected by amateurs, and are often known and appreciated by indigenous peoples, making these insects good "flagship species" (Englund et al. in press). Such an approach could be appropriate for the Marquesas archipelago, which is exceptional in Polynesia because they lack introduced fish (Polhemus et a1. 2000; Keith and Marquet 2002), therefore providing a unique opportunity to preserve native aquatic biodiversity. The easiest and most cost-effective way to preserve the Marquesan aquatic biota would be to maintain a strict prohibition on the importation of any non-native freshwater or estuarine ftsh species. Because freshwater is limited and the Marquesas Islands undergo frequent severe droughts, it is doubtful that aquaculture (a worldwide primary source ofinvasives) would ever be promoted, however, the aquarium trade might. Direct jet service from Papeete, Tahiti provides a two-hour link between the island groups, and has the potential to facilitate importation of aquarium ftsh, which are sold in Papeete. To enhance awareness among the public in French Polynesia, educational materials on the potential impacts of releasing aquarium fish into streams should be prepared and distributed to pet stores, governmental agencies, and particularly 176 schools. This material would explain the nature of the threat, and advocate the preservation and protection of the native aquatic fauna, since most island residents are not aware of the unique biota found in Marquesan streams and elsewhere in French Polynesia. Educational information in the form of color posters, brochures, and especially lesson plans and materials for school teachers could also inform the public and children as to why freshwater stream animals are an important part of their cultural heritage, and explain the steps that they can take to preserve this patrimony. Educational fact sheets stressing the importance of native aquatic life can be prepared and handed out with every aquarium-related purchase. These fact sheets can discuss how to properly dispose offreshwater aquarium plants and animals, and the implications of improper disposal, including potential cleanup costs to stakeholders (e.g., the Salvinia molesta in Hawai'i). SUMMARY OF CONSERVATION EFFORTS Conservation of aquatic ecosystems in Hawai'i and French Polynesia has come a long way in the past century, from advocacy of large-scale introductions of invasive species by governments to a current recognition that many, if not most, species of the aquatic invasives introduced into insular environments have had long-term, deleterious consequences. Introductions of species in the mosquitofish family (Poeciliidae) started in Hawai'i in 1905 (Van Dine 1907) and in Tahiti in 1920 (Keith et a1. 2002), and proved to have deadly consequences for native aquatic biota. A pattern of state-sponsored early biological control introductions of fish shifted after World War II to sportfish introductions (Maciolek 1984; Polhemus and Englund 2003). Some of these sportfish species, such as smallmouth bass, have had negative consequences for stream biota, while others like trout have either not become established in the case of Tahiti (Keith et a1. 2002) or are so thermally restricted (as in Hawai'i) that they exhibit few measurable impacts (Englund and Polhemus 2002). Within the past 20 or so years there has been a shift away from the state-sponsored fish and other aquatic species introductions that occurred in Hawai'i and Tahiti, to introductions by individuals. Most introductions of aquatic biota into insular Pacific environments since the 1980s have been from aquarium 177 releases, intentional food releases by migrant populations, or the intentional spread of sportfish by individual anglers. For example, several species of armored catfish common in the aquarium trade have been introduced into Hawai'i (Sabaj and Englund 1999) and appear to be anthropogenically spreading to new watersheds; in Tahiti, green swordtails, Xiphophorus helleri (Heckel), were first observed in 2003 in the Papeno'o River (Englund, unpubl. data) and are also probably spreading. Smallmouth bass were stocked intentionally in Waikele Stream, O'ahu since 2000 by individual anglers, and likely will be spread further on that island. Intentional food releases are a continuing problem and are exemplified by the apple snail, Pomacea canaliculata (Lamarck). Apple snails were first introduced to Hawai'i in 1989 and greatly harm the culturally important wetland taro crop and are now widely distributed throughout the archipelago (Lach and Cowie 1999). Drawing on relevant examples of invasive species introductions in island areas and also certain vulnerable terrestrial systems, general trends on the impacts of invasive species can be discerned. Endemic island biotas are particularly susceptible to invasive species, with disruptive species such as ants causing major problems in both terrestrial and aquatic ecosystems on a world-wide basis. Some terrestrial invasives such as feral ungulates have also had clearly discernible negative impacts on aquatic species through their destruction of watersheds by eliminating riparian vegetation thereby severely reducing water quality. Because of their insular nature and limited size in island environments, aquatic habitats are particularly at risk from invasive species. Where introduced, fish and other harmful aquatic species introductions have eliminated key elements of the native aquatic insect fauna had other unintended side effects. These negative impacts include predation on native fish, spreading new parasites to which the native aquatic biota has not been previously exposed, and competition for food resources. Proposed measures such as regular monitoring for new invasives, stricter quarantine laws, and aquatic biodiversity reserves on maintained on each island would provide a measure of stability for endemic insular biotas in the tropical Pacific. Research is now beginning to reveal why certain native aquatic insect taxa have declined in tropical Pacific stream areas such as Hawai'i and French Polynesia, but arresting this decline and beginning the process ofrestoration will require the 178 concerted efforts of a variety of governmental agencies, nongovernmental organizations, and private citizens. Community involvement, particularly "grassroots" participation in the conservation of aquatic biotas will provide the greatest protection to these systems in the long run. 179 REFERENCES Adamson, A. M. 1936. Marquesan insects: environment. Bemice P. Bishop Museum Bulletin 98 139: 1 73. Adamson, A. M. 1939. Review of the fauna of the Marquesas Islands and discussion of its origin. Bemice P. Bishop Museum Bulletin 159: 1-93. Alexander, C. P. 1932. New and little known Tipulidae from the Marquesas. Bemice P. Bishop Museum Bulletin 98: 87-92. Alexander, C. P. 1933. New Tipulidae from the Society Islands. Bemice P. Bishop Museum Bulletin 113: 53-56. Alexander, C. P. 1935. A new species of Tipulidae from the Marquesas. Bemice P. Bishop Museum Bulletin 142: 123-24. Allison, A. and R. A. Englund. in press. Chapter 4: Methods for Terrestrial and Aquatic Invertebrate Faunal Surveys. In: Biodiversity assessment of tropical island ecosystems, PABITRA manual for interactive ecology and management (eds. D. Mueller-Dombois, KW. Bridges, and c.c. Daehler). Andrews, S. 1985. Aquatic species introduced to Fiji. Domodomo 3: 67-82. Asquith, A. and R. H. Messing. 1993. Contemporary Hawaiian insect fauna of a lowland agricultural area on Kauai: implications for local and island-wide fruit fly eradication programs. Pacific Science 47: 1-16. 180 Australian Government Department of the Environment and Heritage Website. 2004. Yellow crazy ants. [http://www.deh.gov.au/parks/christmas/fauna/crazy.html] [Version 20 Jun 2004] [Last accessed 13 August 2004]. Baumhover, A. H., A. J. Graham, B. A. Bitter, D. E. Hopkins, W. D. New, F. H. Dudley, and R. C. Bushland. 1955. Journal ofEconomic Entomology 48:462-466. Bax, N., K. Hayes, A. Marshall, D. Parry, and R. Thresher. 2002. Man-made marinas as sheltered islands for alien marine organisms: establishment and eradication of an alien invasive marine species. Turning the tide: the eradication of invasive species. Proceedings of the International Conference on Eradication of Island Invasives (eds. C. R. Veitch and M. N. Clout), p. 26-39. Occasional Paper ofthe IUCN Species Survival Commission 27: 414 pp. Bickel, D. J. 1994. Diptera: Dolichopodidae Part I. Sciapodinae, Medeterinae and Sympycninae (part). Micronesica 27: 73-118. Bille, S. 1851. Report on the Corvette Galathea's circumnavigation 1845, 1846 and 1847, parts II and III. Unpublished English translation in Bishop Museum Library, Carter Collection, 11-F-1. Bishop Museum Backyard Aliens Website. 1999. Rabbits at Haleakala [http://www.bishopmuseum.org/exhibits/pastExhibits/1999lbackyard_alienslb-results/x30 rabbitsAtHaleakala.html] [Last accessed 25 July 2004]. Blackburn, T. 1884. Notes on Hawaiian Neuroptera, with descriptions of new species. Annals and Magazine ofNatural History 14(5): 412-421. 181 Brock, R. E. and A. H. Kam. 1997. Biological and water quality characteristics of anchialine resources in Kaloko-Honokohau National Historical Park. Technical Report 112, Cooperative National Park Resources Studies Unit, Honolulu. Cheesman, E. L. 1926. Two new species of Veliidae from the S. E. Pacific. The Annals and Magazine of Natural History 106: 364-367. Cheesman, E. L. 1927. The insect fauna of French Oceania. Transactions ofthe Entomological Society of London 75: 147-161. Christiansen, K. and P. Bellinger. 1992. Insects ofHawaii. Volume 15. Collembola. University of Hawaii Press. 445 pp. Clarke, I.F.G., 1971. The Lepidoptera of Rapa Island. Smithsonian Contributions to Zoology, No. 56, Smithsonian Institution Press, City of Washington. Cobben, R. H. 1980. The Saldidae of the Hawaiian Archipelago (Hemiptera: Heteroptera). Pacific Insects 22: 1-34. Cole, F. R., A. C. Medeiros, L. L. Loope, and W. W. Zuehlke. 1992. Effects of the Argentine ant on arthropod fauna of Hawaiian high-elevation shrubland. Ecology 73: 1313-1322. Cowie, R. H., G. M. Nishida, and R. A. Englund. 1999. Kamoku-Piikele 138-kV transmission line project, Wa'ahila Ridge, O'ahu: assessment of potential impacts on invertebrates (snails, insects and other arthropods). Final report prepared for CH2MHILL, Inc. Honolulu, Hawai'i. 44 pp. + 14 pp. addendum. 182 Cox, G. W. 1999. Alien species in North America and Hawaii: impacts on natural ecosystems. Island Press, Washington DC. 387 pp. Craig, D. A. 1997. A taxonomic revision of the Pacific black fly subgenus Inseliellum (Diptera: Simuliidae). Canadian Journal ofZoology 75: 855-904. Craig, D. A. 2001. Report on collecting trip to Polynesia to obtain blackflies (Simuliidae). Prepared for Mission d'aide financiere et de cooperation regionale. Papeete, Tahiti. 14 pp. Craig, D. A. 2003. Geomorphology, development of running water habitats and evolution of black flies on Polynesian Islands. BioScience 53: 1-15. Craig, D. A. and D. C. Currie. 1999. Phylogeny of the central western Pacific subgenus Inseliellum (Diptera: Simuliidae). Canadian Journal ofZoology 77: 610-623. Craig, D.A., O. Fossati and Y. Sechan. 1995. Blackflies (Diptera: Simuliidae) of the Marquesas Islands, French Polynesia: redescriptions and new species. Canadian Journal ofZoology 73: 775-800. Craig, D.A. and Joy, D.A. 2000. New species and redescriptions in the central-western Pacific subgenus Inseliellum (Diptera: Simuliidae). Annals ofthe Entomological Society ofAmerica 93: 1236 1262. Craig, D. A., D. C. Currie, and D. A. Joy. 2001. Geographical history of the central-western Pacific black fly subgenus Inseliellum (Diptera: Simuliidae: Simulium) based on a reconstructed phylogeny of 183 the species, hot-spot archipelagoes, and hydrological considerations. Journal ofBiogeography 28: 1101-1127. Crowl, T. A, A R. Townsend, and A R. McIntosh. 1992. The impact of introduced brown and rainbow trout on native fish: the case of Australasia. Reviews ofFish Biology and Fisheries 2: 217-241. Daly, H. V. and K. N. Magnacca. Insects of Hawaii, Volume 17, Hawaiian Hylaeus (Nesoprosopis) bees (Hymenoptera: Apoidea). University of Hawaii Press. 234 pp. Edwards, F. W. 1932. Marquesan Simuliidae. Bernice P. Bishop Museum Bulletin 98: 103-109. Edwards, F. W. 1933a. Tahitian Simuliidae. Bernice P. Bishop Museum Bulletin 113: 35-38. Edwards, F. W. 1933b. Mycetophilidae, Culicidae and Chironomidae and additional records of Simuliidae, from the Marquesas Islands. Bernice P. Bishop Museum Bulletin 114: 85-92. Eldredge, L. G. 1994. Perspectives in aquatic exotic species management in the Pacific Islands: Volume 1, introductions of commercially significant aquatic organisms to the Pacific Islands. South Pacific Commission, Noumea, New Caledonia. 127 pp. Eldredge, L. G. and N. L. Evenhuis. 2003. Hawaii's biodiversity: a detailed assessment of the number of species in the Hawaiian Islands. Bishop Museum Occasional Papers 76: 1-28. Elton, C. S. 1958, The ecology of invasions by animals and plants. University of Chicago Press. 181 pp. 184 Englund, R. A., D. J. Preston, R. Wolff, S. L. Coles, L. G. Eldredge and K. Arakaki. 2000a. Biodiversity of freshwater and estuarine communities in lower Pearl Harbor, O'ahu, Hawai'i with observations on introduced species. Bishop Museum Technical Report 16: 1-166. Englund, R. A., K. Arakaki, D. J. Preston, S. L. Coles and L. G. Eldredge. 2000b. Nonindigenous freshwater and estuarine species introductions and their potential to affect sportfishing in the lower stream and estuarine regions of the south and west shores of O'ahu, Hawai'i. Bishop Museum Technical Report 17: 1-121. Englund, R.A., K. Arakaki, DJ. Preston, N.L. Evenhuis and M.K.K. McShane. 2003. Systematic Inventory of Rare and Alien Aquatic Species in Selected O'ahu, Maui, and Hawai'i Island Streams. Final report prepared for Hawaii Department of Land and Natural Resources, Division of Aquatic Resources, Honolulu. 14 pp. Englund, R. A. and K. Arakaki. 2003. Report on long-term aquatic insect monitoring by Hawaii Biological Survey, Bishop Museum in Pelekunu Valley, Molokai, Hawaii. Report prepared for TNCH Molokai Office. 10 pp. Englund, R. A., M. G. Wright, and D. A. Polhemus. in press. Aquatic insect taxa as indicators of aquatic species richness, habitat disturbance, and invasive species impacts in Hawaiian streams. Bishop Museum Bulletin ofZoology. Englund, R. A. 1999. The impacts of introduced poeciliid fish and Odonata on endemic Megalagrion (Odonata) damselflies on Oahu Island, Hawaii. Journal ofInsect Conservation 3: 225-243. 185 Englund, R. A. 2000. Report on Aquatic Insect Monitoring of May 2000 in Pelekunu Valley, Moloka'i, Hawai'i. Submitted to TNCH Moloka'i Office, Kualapu'u, Hawai'i. 4 pp. Englund, R A. 200la. Long-term monitoring of one of the most restricted insect populations in the United States, Megalagrion xanthomelas (Selys-Longchamps), at TripIer Army Medical Center, Oahu, Hawaii (Zygoptera: Coenagrionidae). Odonatologica 30: 255-263. Englund, R A. 2001b. Report on long-term aquatic insect monitoring by Hawaii Biological Survey, Bishop Museum in Pelekunu Valley, Moloka'i, Hawai'i. Report prepared for TNCH Moloka'i Office. 7 pp Englund, RA. 2002. The loss of native biodiversity and continuing nonindigenous species introductions in freshwater, estuarine, and wetland communities of Pearl Harbor, Oahu, Hawaiian Islands. Estuaries 25: 418-430. Englund, R A. 2003. Report for the 2002 Pacific Biological Survey, Bishop Museum Austral Islands, Expedition to Raivavae and Rapa Iti. Prepared for Delegation ala Recherche (Ministere de la Culture, de l'Enseignement Superieur et de la Recherche, Gouvernement de la Polynesie fran9aise). 30 pp. Englund, R. A. 2004. Report for the 2003 Pacific Biological Survey, Bishop Museum Austral Islands, French Polynesia Expedition to Tubuai and Rurutu. Prepared for Delegation ala Recherche (Ministere de la Promotion des Ressources Naturelles, Gouvernement de la Polynesie fran9aise). 29 pp. 186 Englund, R. A, D. A. Polhemus & D. J. Preston. 2000. Assessment of the impacts of rainbow trout predation on native aquatic invertebrate species within Kokee State Park streams, Kaua'i, Hawai'i. Bishop Museum Technical Report No. 18: 1-125 pp. Englund, R.A., DJ. Preston, G.A. Samuelson, K. Arakaki and N.L. Evenhuis. 2002. Aquatic organisms study for the Koa Timber Commercial Forestry Operation, South Hilo District, County of Hawaii. Report prepared for Wilson Okamoto & Associates, Honolulu, Hawaii. 30 pp. Englund, R. A, D. 1. Preston, N. L. Evenhuis, R. H. Cowie, C. Imada, C. Puttock, and K. Arakaki. 2001. Native and exotic organism study. Lower Wailoa River, Waipi'0, County of Hawaii. Report prepared for USDA Natural Resources Conservation Service, Honolulu, Hawaii. 57 p. Englund, R.A and L.G. Eldredge. 2001. Fishes. In: Hawaii's invasive species. A guide to invasive plants and animals in the Hawaiian Islands (eds. G. W. Staples and R. H. Cowie), p. 32-40. Mutual Publishing & Bishop Museum Press, Honolulu. Englund, R. A & D. A Polhemus. 2001. Evaluating the effects of introduced rainbow trout (Oncorhynchus mykiss) on native stream insects on Kauai Island, Hawaii. Joumal ofInsect Conservation 5: 265-281. Evenhuis, N. L. 1997a. Review of flightless Dolichopodidae (Diptera) in the Hawaiian Islands. Bishop Museum Occasional Papers 53: 1-29. Evenhuis, N. L. 1997b. The genus Sigmatineurum Parent in Hawaii (Diptera: Dolichopodidae), with a revised key to species. Bishop Museum Occasional Papers 47: 66-73. 187 Evenhuis, N. L. 1999. Water-skating Campsicnemus of the Marquesas Islands, including two new species (Diptera: Dolichopodidae). Journal ofthe New York Entomological Society 107: 289-296. Evenhuis, N. L. 2000. A new species ofSigmatineurum (Diptera: Dolichopodidae) from seeps in Waimanu Valley on the Big Island ofHawai'i. Bishop Museum Occasional Papers 63: 31-33. Evenhuis, N. L. 2003. Review of the Hawaiian Campsicnemus species from Kaua'i (Diptera: Dolichopodidae), with key and descriptions of new species. Bishop Museum Occasional Papers 75: 1-34. Evenhuis, N. L. 2004. Humongochela, a new genus of waterfall-loving flies from the Marquesas (Diptera: Dolichopodidae). Bishop Museum Bulletin in Entomology 12: 35-42. Evenhuis, N. L. and Polhemus, D. A. 1994. Review of the endemic Hawaiian genus Sigmatineurum Parent (Diptera: Dolichopodidae). Bishop Museum Occasional Papers 37: 1-19. Faragher, R. A. 1980. Life cycle of Hemicordulia tau Selys (Odonata: Corduliidae) in Lake Eucumbene, N.S.W., with notes on predation on it by two trout species. Journal of the Australian Entomological Society 19: 269-276. Flecker, A. S. and C. R. Townsend. 1994. Community-wide consequences of trout introduction in New Zealand streams. Ecological Applications 4: 798-807. Flint, O. S., Jr., R. A. Englund, and B. Kumashiro. 2003. A reassessment and new State records of Trichoptera occurring in Hawai'i with discussion on origins and potential ecological impacts. Bishop Museum Occasional Papers 73: 31-40. 188 Florence, J. 1997. Flore de la Polynesie fran~aise. Volume 1. Collection Faune et Flore tropicales 34. Editions de 1'0RSTOM, Paris. 393 p. Fontaine, B. and O. Gargominy. 2003. Inventaire des mollusques continentaux de l'archipel des Australes (Raivavae, Rapa). Museum national d'Histoire naturelle Paris, Comite franc;aise pour l'UICN. 25 pp + appendices. Foote, D. and H. Carson. 1995. Drosophila as monitors of change in Hawaiian ecosystems. In: Our living resources: a report to the nation on the distribution, abundance, and health ofU.S. Plants, Animals, and Ecosystems, (ed. Edward T. LaRoe), p. 368-372. U.S. Department ofInterior National Biological Service. Gagne, W. C. 1997. Insular evolution, speciation, and revision of the Hawaiian genus Nesiomiris (Hemiptera: Miridae). Bishop Museum Bulletin in Entomology 7: 1-126. Gagne, W. C. and F. G. Howarth. 1985. Conservation status of endemic Hawaiian Lepidoptera. Proceedings ofthe Third Congress ofEuropean Lepidoptera, Cambridge 1982, pp. 74-84. Gargominy, 0., P. Bouchet, M. Pascal, T. Jaffre, and J.-c. Tourneur. 1996. Consequences des introductions d'especes animales et vegetales sur la biodiversite en Nouvell-Caledonie. Revue D'ecologie (La Terre EtLa Vie) 51: 375-401. Gima, C. 2003. Green monster sucks down cash: saving Lake Wilson will cost at least 1.25 million [online]. Honolulu Star Bulletin, February 27, 2003. [cited 25 July 2004]. Available from: 189 Gibbs, G. W. 1998. Why are some weta (Orthoptera: Stenopelmatidae) vulnerable yet others are common? Journal ofInsect Conservation 2: 161-166. Gillespie, R. G. 1999. Naivete and novel perturbations: conservation of native spiders on an oceanic island system. Journal ofInsect Conservation 3: 263-272. Gillespie, R. G. and N. Reimer. 1993. The effect of alien predatory ants (Hymenoptera: Formicidae) on Hawaiian Endemic Spiders (Aranae: Tetragnathidae). Pacific Science 47: 21-33. Hagen, H. 1867. Notizen beim studium von Brauer's novara-Neuroptera. Verhandlungen der kaiserlich koniglichen Zoologish-botanischen Gesellschaft in Wien 17: 31-62. Haines, W. P. 2004. Reevaluating arthropod extinctions: rediscoveries of Hawaiian leaf-roller moths in the genus Omiodes (Crambidae). Hawai'i Conservation Conference, Honolulu, Hawaii June, 2004, p. 31 in abstracts. Hawaiian Ecosystems at Risk Project (HEAR) 2003. Kahoolawe. [http://www.hear.org/naturalareaslkahoolawe/l [Version: June 22, 2003] [Last accessed 25 July 2004]. Hardy, D. E. 1960. Insects ofHawaii. Volume 10. Diptera: Nematocera-Brachycera. University of Hawaii Press. 368 pp. Hardy, D. E. 1979. Report of preliminary entomological survey of Pua'alu'u Stream, Maui. Part III. Technical Report 27, pp. 34-40. Cooperative National Park Resources Studies Unit, Honolulu. 190 Heddle, M. L. 2003. Systematics and phylogenetics of the endemic genus Scotoryhtra (Lepidoptera: Geometridae) in the Hawaiian Islands. PhD dissertation, University of California, Berkeley. 310 pp. Hoelmer, K. A. and J. K. Grace. Citrus blackfly. In: Eradication of exotic pests (eds. D. L. Dahlsten and R. Garcia), p. 147-165. Yale University Press. Hoffman, B. D. and S. O'Connor. 2004. Eradication of two exotic ants from Kakadu National Park. Ecological Management and Restoration 5: 98-105. Hosking, G., J. Clearwater, J. Handiside, M. Kay, J. Ray, and N. Simmons. 2003. Tussock moth eradication-a success story from New Zealand. Intemational Joumal ofPest Management 49: 17- 24. Howarth, F. G., G. M. Nishida, and N. L. Evenhuis. 2001. Insects and other terrestrial arthropods. In: Hawaii's invasive species. A guide to invasive plants and animals in the Hawaiian Islands (eds. G. W. Staples and R. H. Cowie), p. 41-62. Mutual Publishing & Bishop Museum Press, Honolulu. Howarth, F. G. and D. J. Preston. 2002. Baseline survey of arthropods (insects and relatives) of Kahului Airport environs. Final report prepared for Edward K. Noda and Associates, Inc. 92 pp. Howarth, F. G. and G. W. Ramsay. 1991. The conservation of island insects and their habitats. in: The conservation of insects and their habitats. Eds. N.M. Collins and J.A. Thomas, pp. 71-107. Academic Press, London. Howarth, F. G. and W. P. Mull. 1992. Hawaiian insects and their kin. University of Hawaii Press, 160 pp. 191 Howarth, F. G. 1990. Hawaiian terrestrial arthropods: an overview. Bishop Museum Occasional Papers 30: 4-26. Jordan, S., C. Simon, and D. Polhemus. 2003. Molecular systematics and adaptive radiation of Hawaii's endemic damselfly genus Megalagrion (Odonata: Coenagrionidae). Systematic Biology 52: 89-109. Jordan, S., C. Simon, D. Foote and R. A. Englund. submitted to Molecular Ecology. Phylogeographic patterns of Hawaiian damselflies (Odonata: Coenagrionidae) correlate with Pleistocene island boundaries. Juvik, J. O. 2001. Elwood Curtin Zimmerman: an uncommon dedication. Insects ofHawaii. Reissue of Volume 1. Introduction, pp. xvii-xxvii. Keith, P., E. Vigneux, and G. Marquet. 2002. Atlas des poissons et des crustaces d'eau douce de Polynesie franc,:aise. Patrimoines Naturels 55: 175 pp. Kirch, P. V. 1985. Feathered gods and fishhooks. University of Hawaii Press. 349 pp. Kraus, F. and D. Cravalho. 2001. The risk to Hawai'i from snakes. Pacific Science 55: 409-417. Kuba, H., T. Kohama, H. Kakinohana, M. Yamagishi, K. Kinjo, Y. Sokei, T. Nakasone, and Y. Nakamoto. 1996. The successful eradication programs of the melon fly in Okinawa. In: Fruit fly pests: a World assessment oftheir biology and management (eds. B. A. McPheron and G. J. Steck), p. 534-550. St. Luicie Press, Delray Beach, Florida. 192 Lamb, C. G. 1933. Dolichopodids from the Society Islands. Bemice P. Bishop Museum Bulletin 113: 71 73. Lach, L. and R. H. Cowie. 1999. The spread of the introduced freshwater apple snail Pomacea canaliculata (Lamarck) (Gastropoda: Ampullariidae) on O'ahu, Hawai'i. Bishop Museum Occasional Papers 58:66-17. LaPolla, J. S., D. Otte, and L. A. Spearman. 2000. Assessment of the effects of ants on Hawaiian crickets. Journal ofOrthopteran Research 9: 139-148. Liebherr, J. K. 2005. Patterns of endangerment or pathways to enlightenment? Reconstructing the Fauna Hawaiiensis. Systematics and Biodiversity 2: 175-189. Liebherr, J. K. and D. A. Polhemus. 1997. Comparisons to the century before: the legacy of R. C. L. Perkins and Fauna Hawaiiensis as the basis for a long-term ecological monitoring program. Pacific Science 51: 490-504. Liebherr,1. K. and E. C. Zimmerman. 2000. Insects ofHawaii. Volume 16. Hawaiian Carabidae (Coleoptera), Part 1: Introduction and Tribe Platynini. 494 pp. Lieftinck, M. A. 1966. Some Odonata of Rapa Island, with descriptions of three Polynesian species of Ischnura Charpentier. Tijdschrift voor Entomologie 109: 89-102. Loope, L. L., F. G. Howarth, F. H. Kraus, and T. K. Pratt. 2001. Newly emergent and future threats of alien species to Pacific birds and ecosystems. Studies in Avian Biology 22: 291-304. 193 Maciolek, J.A. 1984. Exotic fishes in Hawaii and other islands of Oceania. In: Distribution, Biology, and Management of Exotic Fishes, Eds. W.R. Courtenay, Jr. and lR. Stauffer, Jr., John Hopkins University Press, Baltimore and London, pp. 131-161. Macfie, l W. S. 1935. Ceratopogonidae from the Marquesas Islands. Bernice P. Bishop Museum Bulletin 114: 93-103. McIntosh, A. R. and C. R. Townsend. 1994. Interpopulation variation in mayfly antipredator tactics: differential effects of contrasting predatory fish. Ecology75: 2078-2090. McIntosh, A. R. and C. R. Townsend. 1995. Impacts of an introduced predatory fish on mayfly grazing in New Zealand streams. Limnology and Oceanography 40: 1508-1512. McIntosh, A. R. and C. R. Townsend. 1996. Interactions between fish, grazing invertebrates and algae in a New Zealand stream: a trophic cascade mediated by a fish-induced changes to grazer behaviour. Oecologia 108: 174-181. Malloch, J. R. 1935. Some acalyptrate Diptera from the Marquesas Islands. Bernice P. Bishop Museum Bulletin 114: 3-31. Marquet, G., P. Keith, and E. Vigneux. 2003. Atlas des poissons et des crustaces d'eau douce de Nouvelle CalMonie. Patrimoines Naturels 58: 282 pp. McLachlan, R. 1883. Neuroptera of the Hawaiian Islands- Part I. Pseudo Neuroptera. Annals and Magazine ofNatural History (6) 10: 176-179. 194 Meyer, J.-Y. in press. Biological invasions in French Polynesia: an overview of the main invasive alien species, and strategic, legal, and management efforts to combat them. In C. Shine, J. K. Reaser & A. T. Gutierrez (eds.). Invasive Alien Species in the Austral Pacific Region: National Reports & Directory of Resources. Global Invasive Species Programme, Cape Town. Meyer, J.-Y. 2003. Rapport de mission de l'expedition scientifique 11 Raivavae et Rapa (Australes) du 18 novembre au 20 decembre 2002. Delegation 11 la Recherche, Papeete. Meyer, J.-Y., J. Florence and T. Motley. 2003. Studies ofthe endangered flora and conservation of the last remnants of native vegetation in the Austral Islands (French Polynesia, South Pacific Ocean). 20th Pacific Science Congress, Bangkok, Thailand, 2003, Symposium 1.6 paper -10, in abstracts. Meyer, J.-Y., and J. Florence. 1997. Tahiti's native flora endangered by the invasion of Miconia calvescens DC. (Melastomataceae). Journal ofBiogeography 23:775-781. Mueller-Dombois, D. and F. R. Fosberg. 1998. Vegetation ofthe tropical Pacific islands. Springer-Verlag. 733 pp. Mumford, E. P. 1935. Taxonomic notes on insects of the Marquesas Islands. Sixth Congreso International Entomology pp. 263-274. Mumford, E. P. 1936. Terrestrial and fresh-water fauna of the Marquesas Islands. Ecology 17: 143-157. Mumford, E. P. 1942. A new genus Bedfordia, and other records of dragonflies from the Marquesas Islands. Annals and Magazine of Natural History (11)9: 824-837. 195 Myers,1. H., D. Simberloff, A. M. Kuris and 1. R. Carey. 2000. Eradication revisited: dealing with exotic species. Trends in Ecology and Evolution 15: 316-320. Needham, 1. G. 1932. Coenagrion interruptum, new species, from the Marquesas, and nymph of Hemicordu1ia assimilis Hagen. Bernice P. Bishop Museum Bulletin 98: 111-114. Needham, J. G. 1933. Pseudagrion demorsum, new species, from the Marquesas. BerniceP. Bishop Museum Bulletin 114: 71-72. Needham, J. G. 1935. A new species of dragonfly and other notes from the Marquesas Islands. Bernice P. Bishop Museum Bulletin 114: 167-169. Needham, 1. G. 1935. Odonata from Tahiti. Society Islands Insects, Bernice P. Bishop Museum Bulletin 113: 21-23. Nishida, G. M. and N. L. Evenhuis. 2000. Arthropod pests of conservation significance in the Pacific: A preliminary assessment of selected groups. In: Invasive species in the Pacific: a technical review and draft regional strategy, p. 115-142. South Pacific Regional Environment Programme (SPREP). Otte, D. 1994. The crickets ofHawaii. The Orthopterists Society, Philadelphia. 396 pp. Parent, O. 1934. Dipteres dolichopodides exotiques. Memoires de 1a Societe Nationale des Sciences Naturelles etMath6matiques de Cherbourg41: 257-308. Paulay, G. 1982. Rapa Island and its weevils. Discovery 16: 25-31. 196 Paulay, G. 1985. Adaptive radiation on an isolated oceanic island: the Cryptorhynchinae (Curculionidae) of Rapa revisited. Biological Journal ofthe Linnean Society 1985 26: 95-187. Paulay, G. 1994. Biodiversity on oceanic islands: its origin and extinction. American Zoologist 34: 134 144. Paulian, R. 1998. Les insectes de Tahiti. Paris, Boubee. 331 pp. Perkins, R. C. L. 1907. Insects of Tantalus. Proceedings ofthe Hawaiian Entomological Society (1906) 1: 38-51. Perkins, R. C. L. 1913. Introduction. Pp. clxxv-clxxx. In: D. Sharp (editor), Fauna Hawaiiensis, Vol. 1, Cambridge University Press, Cambridge. Polhemus, D. A. 1993. Conservation of aquatic insects: worldwide crisis or localized threats? American Zoologist 33: 588-598. Polhemus, D. A. 1996. The orangeblack Hawaiian damselfly, Megalagrion xanthomelas (Odonata: Coenagrionidae): clarifying the current range of a threatened species. Bishop Museum Occasional Papers 45: 30-53. Polhemus, D. A. and A. Asquith. 1996. Hawaiian damselflies: a field identification guide. Bishop Museum Press, Honolulu. 122 pp. 197 Polhemus, D. A. and R. A. Englund. 2003. Predictably sequential pathways for invasions on tropical Pacific islands. In: The ecological and socio-economic impacts ofinvasive alien species on inland water ecosystems: report ofan experts' consultation (eds. L. A. Meyerson, K. A. Cirona, A. Gutierrez, and E. Watson), p. 58-60. Global Invasive Species Programme, Washington, D.C. Polhemus, D. A., R. A. Englund, S. Jordan and 1. T. Polhemus. 2000. Trip report for 1999 entomological and freshwater surveys of the Smithsonian Institution and Bishop Museum Expedition to the Marquesas Islands and Tahiti. Report for the Ministere de la Sante et de la Recherche, Gouvernement de Polynesie fran~aise, Papeete, Tahiti. 18 pp. Polhemus, D. A. and J. T. Polhemus. 2004. Two new genera and thirty new species of Microveliinae (Heteroptera: Veliidae) from the East Papua composite terrane, far eastern New Guinea. Tijdschrift voor Entomologie 147:113-189. Rauzon, M. J. 2001. Isles ofrefuge: wildlife and history ofthe Northwestern Hawaiian Islands. University of Hawai'i Press. 206 pp. Recensement General de la Population Website. 2002. Evolutions de la population des communes de Polynesie fran~aise entre 1996 et 2002. [http://www.ispf.pf/(5dnOuc45ncsoqwz4qelsv4fk) /statidemo/fr/rgp/RGP2002.htrn] [Last accessed 11 Nov 2004]. Sabaj, M.H. and R.A. Englund. 1999. Preliminary identification and current distributions of two suckennouth armored catfishes (Loricariidae) introduced to O'ahu streams. Bishop Museum Occasional Papers 59:50-55. 198 Samuelson, G. A. 2003. Review of Rhyncogonus of the Hawaiian Islands (Coleoptera: Curculionidae). Bishop Museum Bulletin in Entomology 11: 107 pp. Samways, M. 1. 1995. Conservation of the threatened, endemic dragonflies of South Africa. Proceedings of the International Symposium on the Conservation of Dragonflies and Their Habitats. Japanese Society for Preservation of Birds, Kushiro, pp. 8-15. Samways, M. J. 1996. Workability of the new 'IVCN Categories of Threat' with Odonata in South Africa. Odonatalogica 25: 347-354. Samways, M. 1. 1999. Diversity and conservation status of South African dragonflies (Odonata). Odonatalogica 28: 13-62. Saunders, D. L., J. J. Meeuwig, and C. J. Vincent. 2002. Freshwater protected areas: strategies for conservation. Conservation Biology 16: 30-41. Sechan, Y., O. Fossati, P. Guillet, J. Roux, and P. Martin. 1993. Projet d'eradication de Simulium buissoni aux Marquises. Rapport ORSTOM/ITRMLM: ITRM/DOC.ENT./6/93. Sechan, Y., O. Fossati, AH. Danigo, S. Loncke, and P. Guillet. 1998. Control of Simulium buissoni in Nuku Hiva (French Polynesia). Verhandlungen Intemationale Vereinigung fijr Limnologie 26: 2145-2147. Shluker, A D. 2003. State of Hawai'i aquatic invasive species management plan. Hawaii Department of Land and Natural Resources, Division of Aquatic Resources, Honolulu. 140 pp. 199 Simberloff, D. 1986. Introduced insects: a biogeographic and systematic perspective. In: Ecology of biological invasions ofNorth America and Hawaii. Eds. H. A. Mooney and J. A. Drake, pp. 3-26. Springer-Verlag, New York. Simberloff, D. 1995. Why do introduced species appear to devastate islands more than mainland areas? Pacific Science 49: 87-97. Simberloff, D. 1997a. The biology ofinvasions. In: Strangers in paradise: impact and management of nonindigenous species in Florida, D. Simberloff, D. C. Schmitz, and T. C. Brown, eds, pp. 3-17. Island Press, Washington DC. Simberloff, D. 1997b. Eradication. In: Strangers in paradise: impact and management of nonindigenous species in Florida, D. Simberloff, D. C. Schmitz, and T. C. Brown, eds, pp. 221-228. Island Press, Washington DC. Simberloff, D. 2003. Eradication-preventing invasions at the outset. Weed Science 51: 247-253. Solem, A. 1982. Endodontoid land snails from Pacific islands (Mollusca: Pulmonata: Sigmurethra) Part II. Families Punctidae and Charopidae, Zoogeography. Chicago Field Museum of Natural History. Staples, G. W. and R. H. Cowie. 2001. Hawaii's invasive species. a guide to invasive plants and animals in the Hawaiian Islands. Mutual Publishing & Bishop Museum Press, Honolulu. 115 pp. Steiner, L. F., W. C. Mitchell, E. J. Harris, T. T. Kozuma, and M. S. Fujimoto. 1965. Oriental fruit fly eradication by male annihilation. Journal of Economic Entomology 58: 961-964. 200 Steiner, L. F., W. G. Hart, E. J. Harris, R. T. Cunnigham, K. Ohinata and D. C. Kamakahi. 1970. Eradication of the Oriental fruit fly from the Mariana Islands by the methods of male annihilation and sterile insect release. Journal ofEconomic Entomology 63: 131-135. Tomich, P. Q. 1986. Mammals in Hawaii. Bishop Museum Press. 375 pp. Townsend, C. R. 1996. Invasion biology and ecological impacts of brown trout (Salmo trutta) in New Zealand. Biological Conservation 78: 13-22. Townsend, C. R. and T. A. Crowl. 1991. Fragmented population structure in a native New Zealand fish: an effect of introduced brown trout? Oikos 61: 347-354. Tsubaki, Y. 1998. Fluctuating asymmetry of the Oriental fruit fly (Dacus dorsalis) during the process ofits extinction from the Okinawa Islands. Conservation Biology 12: 926-929. Tummons, P. 2003. Maui beetle 'lost' for 107 years is rediscovered near Kaupo Gap. Environment Hawai'i 14: 5. Van Dine, D. L. 1907. The introduction of top-minnows (natural enemies of mosquitoes) into the Hawaiian Islands. Hawaiian Agricultural Experiment Station Press Bulletin 20: 1-10. Van Riper, C. III, S. G. Van Riper, M. L. Goff, and M. Laird. 1986. The epizootiology and ecological significance of malaria in Hawaiian land birds. Ecological Monographs 56: 327-344. Van Riper, C. and J. M. Scott. 2001. Limiting factors affecting Hawaiian native birds. Studies in Avian Biology 22: 221-233. 201 Vitousek, P. M. 1988. Diversity and biological invasions of oceanic islands. In: Biodiversity (ed. E. O. Wilson, pp. 181-189). National Academy Press, Washington, D.C. Wagner, W. L., D. R. Herbst, and S. H. Sohmer. 1999. Manual ofthe flowering plants ofHawai'i. 2 vols. Rev. ed. Univ. Hawai'i Press & Bishop Museum Press, Honolulu. Williams, F. X. 1936. Biological Studies in Hawaiian water-loving insects. Part 1. Coleoptera or beetles. Part 2. Odonata or dragonflies. Proceedings ofthe Hawaiian Entomological Society 9: 235-345. Wilson, E. O. 1996. Hawaii: a world without social insects. Bishop Museum Occasional Papers 45: 3-7. Wright, M. G. and M. J. Samways. 1998. Insect species richness tracking plant species richness in a diverse flora: gall-insects in the Cape Floristic Region, South Africa. Oecologia 115:427-433. Yamamoto, M. and A. Tagawa. 2000. Hawaii native and exotic freshwater animals. Mutual Publishing, Honolulu, 200 pp. Zimmerman, E. C. 1935. Report of Elwood C. Zimmerman to the Director. Bernice P. Bishop Museum Bulletin 133: 68-71. Zimmerman, E. C. 1948. Insects ofHawaii. Volume 2. Apterygota to Thysanoptera inclusive. University of Hawaii Press, Honolulu. 475 pages. Zimmerman, E. C. 1970. Adaptive radiation in Hawaii with special reference to insects. Biotropica 2: 32- 38. 202