The Nature of Cumulative Impacts on Biotic Diversity of Wetland Vertebrates

I.ARRu D. HARRIS about--makes using food chain support as a variable for Department of Wildlife and Range Sciences predicting environmental impacts very questionable. School of Forest Resources and Conservation Historical instances illustrate the effects of the accumula- University of Florida tion of impacts on vertebrates. At present it is nearly impos- Gainesville, Florida 32611, USA sible to predict the result of three or more different kinds of perturbations, although long-range effects can be observed. One case in point is waterfowl; while their ingestion of lead ABSTRACT/There is no longer any doubt that cumulative shot, harvesting by hunters during migration, and loss of impacts have important effects on wetland vertebrates. Inter- habitat have caused waterfowl populations to decline, the actions of species diversity and community structure produce proportional responsibility of these factors has not been de- a complex pattern in which environmental impacts can play termined. a highly significant role. Various examples show how wet- Further examples show muttiplicative effects of similar ac- lands maintain the biotic diversity within and among verte- tions, effects with long time lags, diffuse processes in the brate populations, and some of the ways that environmental landscape that may have concentrated effects on a compo- perturbations can interact to reduce this diversity. nent subsystem, and a variety of other interactions of in- The trophic and habitat pyramids are useful organizing creasing complexity. Not only is more information needed at concepts. Habitat fragmentation can have severe effects at all levels; impacts must be assessed on a landscape or re- all levels, reducing the usable range of the larger habitat gional scale to produce informed management decisions. I generalists while threatening the genetic integrity of small, conclude that a system of replicate wetland reserves that are isolated populations. The complexity of trophic interactions, allowed to interact naturally with the surrounding landscape and the propensity, or necessity, of vertebrates to switch will be more effective in preserving biotic diversity than iso- from one food source to another--something we know little lated sanctuaries.

The National Environmental Policy Act and similar landscape or regional scale of analysis. This means that state legislation dictate the performance of cumulative the wetland manager and program administrator must impact analyses prior to specific environmental manip- be increasingly sensitive to the presence and meaning ulations. Still other legislation requires the mainte- of within-species biotic diversity, not only to among- nance of biological diversity on certain Federal lands species diversity. What follows are descriptions of how (e.g., the National Forest Management Act), and "the wetlands maintain vertebrate biotic diversity and of protection and propagation of a balanced indigenous some ways that various impacts have accumulated to population of shellfish, fish, and wildlife in the body of reduce biodiversity. The majority of examples cited water to which the discharge is to be made ..." [1972 are from southern wetlands and related systems only amendments to the Federal Water Pollution Control because of my familiarity with them. Act, Section 315(a)]. Although logic and observations dictate that the impacts of repeated, similar actions, or a sequence of different actions, can accumulate to Hierarchy of Biotic Diversity cause severe or irreparable damage to biotic diversity, A great deal has been learned about genetics over some suggest that the concept of cumulative impacts is the last 80 years. It is now recognized that the lowest only "an interesting hypothesis." While it may not be levels of genetically encoded biotic diversity occur at possible to predict cumulative impacts a priori, they the level of the gene allele. Other important aspects of certainly exist, can be described, and are critically im- genetic diversity within a species or subspecies include portant. To be effective, assessment of cumulative im- the proportion of loci that are polymorphic, the pacts on vertebrate communities must be done at a number and types of alleles at these loci, and the average level of heterozygosity. Moreover, we now KEY WORDS: Cumulative impacts; Landscape ecology; Vertebrate know that the expression of a dominant, heterozygous, biodiversity;Wetland habitat; Wetland impacts or recessive allele at a single locus may determine

EnvironmentalManagement Vol. 12, No. 5, pp. 675-693 1988 Springer-VerlagNew York Inc. 676 L.D. Harris

adaptive characteristics such as the presence or ab- Presence, relative abundance, and activity levels of sence of a critical enzyme, resistance to malaria, or oc- some wetland vertebrates--beaver, alligator (Alligator currence of sickle-cell anemia. mississippens/s), and to a lesser extent muskrat, and Our former understanding of genetics led ornithol- large fish and turtles--may transform the physical ogists and wetland wildlife scientists to recognize the structure of a site, altering hydroperiod and water lesser snow goose (formerly Chen hyperborea) and blue flow characteristics to such an extent that they influ- goose (Chen caerulescens) as distinct species. The same ence the very presence/absence of wetlands through- situation was true for the great blue heron (Ardea he- out the landscape. rodias) and the great white heron (formerly Arclea occi- Unless laws and permitting and regulating policies dental~s). In both cases further investigation revealed begin to consider biotic diversity across the full hierar- that the groups were, in fact, only expressions of allele chical spectrum, they will not ensure conservation of variation within the same genome. Because of rapid the wetland biodiversity resource (Noss and Harris progress in the classification of vertebrates during this 1986). Only regulation and planning strategies that century, bird species have been reclassified such that maintain the full gamut of diversity (including pro- the more than 20,000 recognized species in 1920 have cesses) to the landscape level and above stand to be been reduced to about 9000 at present (Mayr 1982). effective at maintaining biodiversity at all lower levels. Thus, what formerly constituted diversity among However, even more may need to be done to maintain species is now considered to be genetic diversity within long-term diversity at the community, ecosystem, or a species. landscape levels. Both genetic (genotypic) and environmentally in- duced (phenotypic) variance may be manifested either as a gradual or clinal change, or as distinct ecotypes Wetland Wildlife Habitat and Food because of the separation of populations, with little or Chain Support no opportunity for interbreeding. Both situations rep- resent legitimate, and ecologically important, aspects Wetland impact analysis procedures frequently re- of biotic diversity that must be considered. volve around two specific biological functions: the The number of species (richness) and some mea- provision of habitat and of food chain support. These sure of their relative abundance (equitability) combine are neither simple nor independent concepts, and thus to form the concept of species diversity. Because this some explanation is warranted. concept was so heavily reviewed and debated by ecolo- gists in the 1960s and 1970s, it became the standard by' Wildlife Habitat which biotic diversity was monitored and evaluated. At least four different issues are integral to the as- Not until this decade have ecologists and regulators re- sessment of cumulative impacts on wetland habitat. alized that it is neither sufficiently broad nor sensitive These are (1) specialist vs generafist species; (2) alpha, to assess the complete spectrum of biotic diversity. beta, and gamma diversity; (3) component vs com- Interactive community processes, such as herbivory, pound ecosystems; and (4) animals' need to move. Nu- pollination, propagule dispersal, predation, parasitism, merous amphibian and reptile species are habitat spe- and competition, all function to influence the distribu- cialists: mammals such as bog lemmings (Synaptomys tion, abundance, form, and behavior of organisms. It spp) and round-tailed water rats (Neofiber allen,), and is the occurrence and magnitude of coevolved pro- breeding birds such as Swainson's warbler (Limnothlyp/s cesses such as these that distinguish functional biolog- swainsonii) and prothonotary warbler (Protonotaria ci- ical communities from random collections of species trea). Resident species that have specific habitat re- and determine biotic diversity at the community, level. quirements also usually have small home-range sizes, Thus, both the occurrence and relative abundance of and thus all their resource requirements may be pro- species, and the relative magnitude of key biological vided within a single wetland. In the classification processes, must be evaluated when assessing impacts scheme of Whittaker (1960), these within-habitat on wetland biotic diversity. Keystone wetland species species would constitute the alpha (et) component of such as muskrat (Ondatra zibethicus), nutria (Myocastor diversity. At the other extreme, large, frequently wide- coypus), beaver (Castor canadensis), manatee (Trichechus ranging species, such as bald eagles (Haliaeetus leucoce- manatus), and moose (Alces alces) can dramatically alter pha/us), mink (Mustela v/son), or Florida panther (Fel/s community vegetation structure, which in turn influ- concolor coryz), may be no less dependent on wetlands ences and controls the occurrence of yet other animal for specific life needs, but they cannot be restricted to, species. nor contained within, a small tract of wetland. Species Cumulative Impacts on Vertebrate Biodiversity 677

such as these must utilize two or more specific habitat systems, but the long-term viability of the subsystems types within a regional landscape in order to meet may well depend upon the presence of these land- their life requirements. Thus, they are more appropri- scape integrators. ately thought of as creatures of the landscape; in There seems to be a gradient in ability (presumably Whittaker's scheme they might constitute the gamma adaptive) and apparent need for the different classes (~t) component of diversity. of vertebrates to move. Some vertebrates that occur in As noted by Elton (1966), animals higher in the tro- frequently or constantly renewed environments (e.g., phic pyramid tend to be less habitat specific. Plants flowing water) reflect little need for movement. Am- and primary consumers are frequently habitat spe- phibians fit at this end of the gradient despite their cialists, but top carnivores rarely are. The principle of larval (aquatic) versus adult (mostly terrestrial) fife- the inverse pyramid of habitats suggests that while an- stage differences. Long-rived turdes that spend de- imals at the higher trophic levels are generally fewer in cades in an aquatic environment but still require a 1-h number, they also generally range over greater dis- nesting period in a safe terrestrial environment dram- Lances and derive their life requirements from a atize both the necessity of diverse component sub- greater number of community types or habitats (Fig- systems and the need to move between them. ure 1). With few exceptions, terrestrial mammals are The concept of component and compound ecolog- highly vulnerable to forces that create impassible bar- ical communities (Root 1973) suggests that regional or riers between component habitats in the landscape. landscape-level systems consist of component commu- This derives from the need to range over wide dis- nities articulated in space by the movement of nu- tances while being limited to movement on the trients, energy, propagules, and organisms. Within ground. Mammals that inhabit wetlands exhibit spe- any given component subsystem, ecological relations cific traits that make them highly vulnerable to isola- will be dominated by relatively intense, niche-specific tion and habitat fragmentation. Virtually every species and/or host-specific interactions "that tend to occur of North American mammal weighing more than 0.5 continuously. Presence or absence of a species or inter- kg that inhabits wedands is commerciaUy valuable as a action within the component will be largely deter- furbearer. Only four of the 30 commercially listed mined by conditions internal to the component, that is, furbeare~s (Deems and Pursley 1983) are herbivorous its content. (Table i). The remainder of the furbearers (26/30 = Wide-ranging vertebrates link various component 85%) are either carnivores in the strict sense or omniv- subsystems spatially and ecologically to help form orous members of the order Carnivora (e.g., the black compound landscape systems. Although transient bear). Because most are carnivores, they have much species that move from one component subsystem to larger home ranges than herbivores or omnivores of another are frequendy generalists, highly specialized equal size, and because they inhabit the water's edge, "mobile links" are not infrequent in temperate systems their home ranges tend to be long and narrow, aggra- and are common in tropical systems (e.g., Gilbert vating the probability of fatal encounters with humans 1980, Terborgh 1986). or human artifacts. Carnivorous mammals are prob- The need for vertebrates to move derives from ably the group most vulnerable to habitat fragmenta- many basic biological functions, ranging from the tion. need to access resources (food, water, or shelter), to No small population (fewer than several hundred) the need for sexual organisms to mate and outbreed, of vertebrates can maintain its demographic and ge- and the need to colonize new or disturbed environ- netic integrity indefinitely in the face of population merits (Chepko-Sade and Halpin 1987). Most am- fragmentation, genetic isolation, and inbreeding phibians require fresh water for their egg and larval (Frankel and Soule 1981). Thus, modifications to the stages, but many use dry land as adults. Reptiles re- environment that preclude movement between com- quire an aerated medium for their eggs, but many ponent subsystems may be as devastating to verte- then require an aquatic environment as adults. Large brates in the long run as are forces that actually de- resident species such as black bear (Ursus americanus) stroy the wetland. forage in upland environments during summer and This phenomenon can be dramatically illustrated fall but depend on lowlands in winter. Migrant species by citing the history of mammal extinction and endan- may interconnect subsystems thousands of kilometers germent in Florida, USA. A strong linear relation be- apart. Not only does the occurrence and abundance tween percent of size class now extinct and body size of the wide-ranging animals depend upon the exists for mammals dating from the Pleistocene to the asynchronous functioning of the component sub- present. While 100% of the very large (>10,000 kg) 678 L.D. Harris

" CARNIVORE i~ '~--"::-T Figure 1. Relationship between OMNIVORE the trophic pyramid of numbers ERBIVORE-GENERALIST \ ~.,..,.,, / of animals and the inverse pyr- amid of their habitats. The in- ERBiVORE-SPECIALIST verse pyramid suggests that I ~ PRODUCER \~ / while animals of the higher tro- phic levels are generally less / \so,, M'CROORG'SMSV abundant, they range over greater distances and derive their life requirements from a PYRAMID of NUMBERS INVERSE PYRAMID greater number of habitat types. of HABITATS (See Ehon 1966)

Table 1. North American furbearers classified by systematics, and trophic and habitat characteristics. Table 1. (continued) Tropic ! Tropic Species group a Characteristic habitat Species group a Characteristic habitat Marsupalia Fisher (Martes C Mature conifer and Opossum (Didelphis O' Riparian woodlands pennant,) mixed forests v/rg/n/ana) Short-tailed weasel C Cool conifer forests Rodentia (Mustela erminea) with water and openings Beaver (Castor H Slow-flowing forested Long-tailed weasel C Warm conifer forests canadensis) waterways with water and Muskrat (Ondatra H Fresh and saltwater (Mustelafrenata) zJbethicus) wetlands openings Mink (Mustela v/s0n) C Wetlands generalist Nutria (Myocastor H Coastal marshes and Wolverine C Boreal and montane coypus) swamps (Gulo gu/o) forests Carnivora Badger (Taxidea C Prairies and grasslands Coyote (Canis/atrans) C Grasslands, meadows, taxus) and open Striped skunk C Forest edges near environments (Mephitis mephitis) water Gray. wolf (Canis C Grasslands, tundra, Spotted skunk C Grasslands and lupus) and open woodlands (Spilogale putorius) meadow edges Arctic fox (Alopex C Alpine and Arctic River otter (Lutra C Riparian forests and /agopus) tundra canadensis) marshes Red fox (Vulpes C Woodlands and Cougar (Felis C C,eneralist vulpes) prairies CO*'bCOIoT) Gray fox (Urocyon O Woodlands and forests Lynx (Felis C Boreal forests and cinereoargenteus) canadensis) tundra Brown bear (Ursus C Generalist associated Bobcat (Felis rufus) C Broken forests arctos) with forested Pinnipedia landscapes Northern fur seal C Nearshore marine Black bear (Ursus O Principally hardwood (Callorhinus americanus ) forests Polar bear (Ursus C Arctic coast and pack ursinus) Harp seal (Phoca C Arctic marine mar/t/mus) ice groenlandica) Bassarisk (Bassariscu~ C Riparian forests in Hooded seal C Pelagic marine astutus) arid environments (Cystophora cristata) Raccoon (Protyon 0 Riparian forests and /otor) wetlands Marten (Mattes C Mature conifer and K) = omnivore, H = herbivore, C = carnivore. americana) mixed forests Prepared in consultation with J. Eisenberg. Cumulative Impacts on Vertebrate Biodiversity 679

mammals are extinct, there is no fossil evidence that species that are already common throughout the any of the very small (c.10 g) mammals are extinct human-dominated landscape. Therefore, activities (Harris and Eisenberg 1988). This same pattern still that may increase the number of species and biological pertains inasmuch as the large mammals have been diversity of individual component subsystems may in most dramatically impacted by development. Since fact cause the demise of some species and homogenize Florida's discovery, the bison (Bison bison), monk seal regional differences, thereby greatly reducing the bio- (Monachus monachus), and red wolf (Canis rufus) have diversity of the compound or regional system. been extirpated. The manatee (Trichechus manatus), black bear (Ursus americanus), Florida panther (Felis Food Chain Support concolor coryi), and Key deer (Odocoileus virginiana cla- While alteration and loss of habitat pose serious vium) are presently all listed as either threatened or threats to wetland fauna, alteration of food chain sup- endangered species. With the exception of legal bear port may be equally important. Yet, it is impossible to hunting in two designated areas, the number one use analyses based solely on biomass or available en- known source of mortality for all of these species de- ergy or simple trophic-dynamic models to predict im- rives from vehicular collision. With the exception of pacts unless we know what does and does not consti- mainland white-tailed deer and raccoon, all seven of tute food. For example, the muskrat, which is perhaps the native large mammal species in Florida are either the most studied wetland vertebrate in North America, extinct or threatened, and by mortality associated with is commonly thought of as an herbivorous habitat gen- their movement throughout the landscape is a primary eralist. Yet we know little of the conditions under Cause. which it can or would prefer to exist as a rhizovore Based on current knowledge, the direct conse- (roots, corms, and tubers), a granivore (seeds), a frugi- quences of habitat fragmentation on biodiversity may yore (nuts and fruits), a browser (shoots, twigs, and be assigned to one of the following four categories: cambium of woody plants), an omnivore, or even a I. Loss of large, wide-ranging species (~l species), espe- cannibalistic carnivore. cially top carnivores or otherwise threatening The literature on ecology of terrestrial vertebrates forms (e.g., bears). Cursorial forms, which are vul- is heavily biased toward the primary producer-pri- nerable to automobile collisions, and aquatic mi- mary consumer-carnivore energy pathways, while the gratory forms (e.g., fish, manatees), which are vul- role of detritus and the detritivore-omnivore-carni- nerable to obstacles to migration, are particularly vore food chains are much less well known. Macroin- sensitive. vertebrate detritivore food chains such as those asso- 2. Loss of area-sensitive or interior species (e~ species) that ciated with crayfish are probably more important than only reproduce in the interior of large tracts of most vertebrate ecologists acknowledge (e.g., Penn wetland and are therefore vulnerable to reduction 1950). Aside from the autochthonous detritus pro- in size of the individual component wetlands as duced annually within wetlands themselves, a variable well as reduction in total wetland acreage. but important amount of organic matter produced 3. Loss of genetic integrity from within species or popula- elsewhere is generally washed into and perhaps t/ons that inhabit areas too small for a viable popu- through many wetland systems. In southern wetland lation of individuals. This is especially important systems, such as the Mississippi River basin, these for large, wide-ranging carnivores or raptors that inputs probably exceed on-site production many times are territorial and require areas proportional to over. population number (i.e., are not amenable to pop- As a generality, food-habit specificity of terrestrial ulation packing). vertebrates tends to be inverse to the size and trophic 4. Increase in abundance of habitat gateralists character- level of the consumer. Specialists tend to be primary istic of disturbed environments (f3 species). Often these consumers with specialization reduced at the higher species serve as competitors (e.g., starlings), pred- trophic levels. With some exceptions (e.g., snail kite, ators (e.g., crows and raccoons),or parasites (e.g., Rostrhamus sociabilis), carnivorous terrestrial vertebrates brown-headed cowbirds) on native species and ac- are not specialists. Similarly, one of the adaptive ad- celerate their demise. vantages of large body size is that the body can serve as a food storage reserve and thus intake rate can be in- The ultimate result of these four classes of impacts creasingly variable as a correlate of size. Large carni- is that each region loses its unique and distinguishing vores (e.g., alligators, snapping turtles, snakes, rap- biological characteristics and acquires the generalist torial birds, Florida panthers, or red wolves) do not 680 L.D. Harris

require daily energy input, as is the case with small flood water into generally terrestrial areas for re- carnivores such as shrews. This is especially true of the productive purposes. Most of these organisms are cold-blooded vertebrates. Thus, coupling the relations either detritivorous or carnivorous. of food habits and body size together virtually ensures . Primary production that occurs during spring and that large, wide-ranging carnivores are not only un- early summer is physically unavailable to land- predictable in terms of what they consume, but also based herbivores because of i.ts spatial location when and where they consume it. As observed by vis-h-vis the flood water. Craighead and Craighead (1956), the consumption of . As water levels recede, the site is reinvaded by prey species by predators seems not so dependent on herbivorous mammals. Seeds and fruits rapidly prey taxonomy or abundance as on its vulnerability. become available for granivorous and frugivorous In addition to these variables, a constant danger of mammals and for returning migrant birds, most utilizing a food-habits approach to impact analysis is of which were insectivorous during the breeding that organisms that do not have clearly identified food season. habits are neglected, while those that have specialized . As temperatures decline in fall and winter, the and well-known food habits tend to be emphasized. metabolic and activity levels of cold-blooded fish, Because of highly significant interactions among amphibians, and repdles also decline. This occurs spatially defined habitats, food chain type, the pre- simultaneously with reinvasion of the wetland by dominant consumer group present, and season, im- overwintering birds and mammals that are pre- pact analyses must be conducted at the landscape level. dominantly direct consumers of primary produc- Because hydroperiod and water level fluctuation are tion. In southern Florida, reproduction of wading of primary importance in wetlands, special attention birds such as wood storks (Mycteria americana) and must be given to these variables. Ironically, within roseate spoonbills (Ajaia ajaja), raptors such as the many wetlands of the deep South, the detritivore bald eagles (Haliaeetus leucocephalm) and osprey pathway may predominate during the growing season (Pandion haliaetus), and even the white-tailed deer when primary productivity is high, and the primary (Odocoileus virginianus) (Richter and Labisky 1985) consumer pathway predominates during winter when seems to be synchronized with the winter low primary production is low (e.g., Harris 1978). This water levels. pattern can be explained as follows: Only analyses that consider the habitat and wildlife 1. Mammal herbivores are largely driven from a site content of the component wetland against the tern-. by rising water levels in spring and early summer poral and spatial context within which the wetland (only one, or perhaps two, species of mammal occurs will be fruitful. Perturbations to the wetland it- such as muskrat, nutria, round4ailed water rat, or self will have direct and measurable effects on the ver- beaver remain). tebrates that live within it, while low-intensity actions 2. Bird abundance declines as overwintering and mi- that occur throughout the landscape will have only in- grant species emigrate to northern breeding direct impacts on this group. Conversely, organisms grounds. Resident birds shift their diet to emer- that occur high in the trophic hierarchy are likely to gent lifestages of principally detritivorous benthic range over a wide area and depend upon the invertebrates that fed on detritus derived from the asynchronous functioning of component subsystems previous year's production. Detritus therefore within the compound landscape. Low-intensity land- serves to store energy and introduces a seasonal scape manipulations that alter the ratio and configura- time lag into the production-consumption pro- tion of component subsystems will have direct impacts cess. on these species, while the manipulation, creation, or 3. Cold-blooded, detritivorous macroinvertebrates, elimination of a component wetland might have only such as the crayfish, and cold-blooded, carnivo- indirect effects. Therefore, whether the impacts re- rous fish, amphibians, and reptiles increase their suiting from any given wetland perturbation will have metabolic and ecological activities and their abun- direct or indirect effects depends largely upon the tro- dance in response to warming temperatures. The phic level and range requirements of the organism in- diversity and abundance of amphibians and rep- volved (Table 2). tiles exceeds that of birds and mammals at this season. 4. Macroinvertebrates, shellfish, finfish, and am- Assessing Accumulation of Impacts phibians follow the shallow, migratory edge of When approaching cumulative impact assessment, Cumulative Impacts on Vertebrate Biodiversity 681

Table 2. Interaction of environmental perturbation sponse curve is fundamentally different for semel- with vertebrate species, illustrating that a direct effect parous and iteroparous species. on one species might constitute an indirect effect on a different species type. Additive Responses of Dissimilar Actions The additive responses of dissimilar actions, such as Impact type continuing to trap furbearers after a highway has been Species type A B expanded to carry heavier traffic loads through a Stenotopic or habitat Direct Indirect marsh, are not easy to predict. It is possible to add specialists (a species) recorded road kill to recorded fur harvest, with the Eurytopic or habitat Indirect Direct sum representing the true number of animals re- generalists (13 species) moved, but it is not possible to add last year's fur catch A = sharp, focused, locallyintense, and within a wetland;B = atten- from a marsh without a road to last year's road kill uated, diffuse, regionallydistributed among wetlands. from a similar marsh not subject to trapping and accu- rately estimate last year's mortality in a marsh that has it is important to distinguish between the assessment of been trapped and has a road. Errington (1963) and consequences after the fact and the need to predict others have spent decades attempting to understand them beforehand. The knowledge and expertise to the intercompensatory nature of responses to such ac- conduct accurate post hoc assessments of cumulative tions, yet it seems beyond our ability to predict the cu- impacts resulting from major projects in the past exists mulative impacts of three or more simultaneous ac- (cf., Brinson 1988),--and increasingly common histor- tions. North American waterfowl populations are im- ical accounts of long-term ecological change (e.g., pacted by reduction of breeding range wetlands, Cronin 1983, Chase 1986, Crosby 1986) could be prof- harvesting during the course of migration, ingestion itably dissected and analyzed by ecologists. Multiple in- of lead shot and other toxicants, and loss and distur- dependent impact analyses 9 f several different com- bance of overwintering habitat in the South. Despite pleted projects might be performed to determine the evidence that impacts are accumulating (Figure 2), when such actions constitute a stress and when they it is impossible to pinpoint single-factor effects. are a benefit. The result will no doubt depend upon the intensity of the action as well as the time frame and Multiplicative Effects of Similar Actions scale of the landscape unit analyzed (Odum and others The multiplicative effects of similar actions, such as 1979, Barrett and Rosenberg 1981, Rykiel 1985). the serial addition of dams to a river, are perhaps most The following examples are chosen to represent the well known. If each of five dams were to exclude 50% fundamentally different ways that wetland vertebrate and allow 50% of a migrating fish population to pass, communities accumulate and/or manifest impacts of the reduction would not be 250%, but rather 1 - 0.55 different sorts. = 97%. Three percent would presumably survive the migration around all five dams. A similar geometric Simple Additive Responses response curve would result whenever a constant pro- Simple additive responses are likely to result when portion of the resource is removed or whenever a con- actions of similar type, such as furbearer or waterfowl stant fractional response occurs. harvest, are repeated. These impacts affect the biotic diversity directly with no intervening variables between Nonmonotonic or Threshold Responses the act (hunting or trapping) and its effect (population Nonmonotonic or threshold responses may occur reduction). The responses may accumulate in time under certain circumstances. For example, construc- (harvesting resident species) or in space (when e.ach ting a road through a large tract of bottomland hard- province and state along the migration route takes its wood forest might substantially increase local breeding allocated quota from a population of migratory water- bird species diversity (see Perkins 1973 for supporting fowl). data). If it were a single road, initially only a few forest Effects may be additive within a season but not nec- interior spedes would be negatively impacted, whereas essary between seasons or over longer periods of as many as 20 edge species might typically colonize the time. A large literature attests that all biological popu- disturbed environment along the road. An alien lations have some capacity to compensate for variable species might parasitize all the nests of a native warbler mortality rates. It is enough to say that the long-term within a distance of 30 m on each side of the road. population response will not be a simple linear func- This means thatroughly 10 ha (25 acres) of native tion of harvesting intensity, and the nature of the re- warbler habitat are lost per 1.6 km (1 mile) of road 682 L.D. Harris

13 12 E) II

I0 |174 Figure 2. Decline in the number of waterfowl in North America, 9 (3 @ 1955-1985. This is one manifesta- 0 (3 @ | | tion of cumulative impacts that de- a= 8 c3 | Q o- - o o rive from a combination of factors "5 7 spanning thousands of kilometers

o o . . ~ * . Mallard of space and decades of time. 0 6 ~. o -35% Slopes derived by least-squares ~E analysis are - 49.6 x for mallard 4 and -93.7x for pintail; percent "-,~ PINTAI L declines calculated from data are -50% 35% for mallard and 50% for pin- tail, Data from US Fish and Wild- life Service and Canadian Wildlife 1955 60 65 "70 "75 80 1985 Service 1985.

through the bottomland forest. As additional roads that temperature-dependent sex determination (TSD) are added, the percentage of habitat parasitized in- is a common phenomenon in reptiles and amphibians creases linearly to 100% when a density of 4 km/259 (e.g., Bull 1980, Morreale and others 1982). Since the ha (2.5 miles per square, mile) of habitat is reached gender of individuals is determined by only a few de- (Figure 3). Any further increase in road density will grees difference in incubation temperature, the sex of have little additional impact on warbler-nest para- entire cohorts of offspring depends on nest and site sitism, since all habitat has already been parasitized. conditions. If the species involved does not reach Temple (1986) and Wilcove (1985), among others, sexual maturity for many years (large turtles or alli- have documented the negative consequences of edge- gators), all individuals entering the breeding popula- induced predators, competitors, and nest parasites to tion in 1988 might be impacted by incubation temper- extend as far as 500 m from the forest edge. atures of the 1970s. It would be shocking to discover Although similar threshold responses can be de- that all breeding adults returning to a site 20 years scribed for other perturbations, this form of response after birth (e.g., marine turtles) were of the same curve is probably not sufficiently common to conclude gender. Under natural circumstances the population that there are definable or discrete critical levels to sex ratio would be counterbalanced by numerous in- which a system can be perturbed before the negative tersite, interyear, and intervening variables. But such consequences become unacceptable. An agency that opportunities are rapidly diminishing because of frag- hinges permitting or enforcement authority on "crit- mented and isolated habitats and greatly reduced pop- ical threshold levels" of acceptable impact will be frus- ulations. Increasing proportions of a population de- trated by the present inadequacy of data as well as the rive from the production of a single site or a single constant changes in the information base. year. A second example involves the recent extinction of Indirect Impacts the dusky seaside sparrow (Ammospiza maritima nigre- Indirect impacts occur when a perturbation directly scens). Originally classified as a distinct species, the impacts one wetland component and then proceeds to dusky was merged with other seaside sparrows as a impact other temporal or spatial components or pro- single species in 1973 (Kale 1977). With its natural cesses by chain reaction. Indirect effects can also result habitat limited between the 3- and 5-m contour lines in from perturbations that occur ou/side the wetland but a few hundred square kilometers of Brevard County, get carried into the wetland by diverse pathways. The Florida, the dusky's restricted distribution was unique impacts may not be manifest until a later time (time- among North American birds. The activities that en- lag effects) or in a different place (space-lag effects). dangered and ultimately exterminated the dusky began with DDT spraying for mosquito control in the Long Time-Lag Effects 1940s. During the 1950s marsh impoundment was im- Long time-lag effects can be illustrated by noting plemented as a mosquito-control and waterfowl- Cumulative Impacts on Vertebrate Biodiversity 683

r~ I,g males. The consequences of manifold activities had ac- N IOO" cumulated throughout the previous 30 years but had not been adequately monitored. The cumulative effect 75" proved lethal; the last true dusky seaside sparrow died f_ in captivity June 17, 1987. Similar examples occur when the age structure of a Z population is significantly but unknowingly skewed W and a false security about population numbers is fol- lowed by a later catastrophic decline (e.g., the Puerto Rican parrot). These examples illustrate the impor- 0 IO 2.0 3.O 4:o tance of using more sensitive population parameters ROAD DENSITY (mi/mi =) (e.g., effective population size) that take into account Figure 3. Response curve showing percent of native warbler age and sex ratios, dominance, dispersal, and related nests parasitized as roads are constructed through a bottom- characteristics. land hardwood forest. Threshold response curves charac- terize a small class of environmental impacts, but this form of Indirect Effects Concentrated in a response does not seem widespread. Component Wetland Indirect effects may be concentrated in a compo- management technique. Impoundment during the nent wetland imbedded within the compound land- breeding season seriously limited nesting and feeding, scape where the actual environmental perturbation and the impounded areas of Merritt Island were the occurs. Three examples of increasing complexity are first to lose the dusky. Construction of the Beeline Ex- given. Kesterson Reservoir consists of a series of irriga- pressway (SR528) through one of the last remaining tion drainwater/evaporation ponds totaling about 500 patches of primary habitat in the early 1970s further ha in the much larger Grasslands area of Merced fragmented and jeopardized the dwindling population County, California, USA, on the west side of the San and restricted management of the remaining habitat. Joaquin Valley (Ohlendorf 1985, Tanji and others Drainage for domestic pasture and the development 1986, Ohlendorf and others 1987). As part of the US of residential subdivisions further fragmented and re- Bureau of Reclamation's Central Valley Project, Kes- duced habitat in the St. Johns River marshes. With terson Reservoir was completed and began to receive such a high concentration of roads, subdivisions, and surface irrigation tailwater and serve as the terminus related human facilities, the use of prescribed fire was for excess runoff in 1972. Although originally planned greatly limited. In the absence of prescribed fire, plant as a flow-regulation area, the reservoir and adjacent succession occurred, the fuel load of the habitat in- 1900-ha Kesterson National Wildlife Refuge now creased, and fires that were ignited on surrounding served as the terminus of the San Luis Drain. Histori- rangelands blew into the refuge and caused serious cally, the Grasslands area comprised the largest tract damage (and the loss of human lives as well as wildlife) of waterfowl habitat in the San Joaquin Valley and was in the mid-1970s (Kale 1977, Baker 1978, USFWS used by perhaps 65% of all migrating birds in the Pa- 1979a). cific Flyway. The average annual harvest of over The final, and fatal, mistake was not due to mis- 300,000 ducks (1966-1980) ranked Merced County as management, but rather to the inadequacy of survey the first or second most important waterfowl hunting methodology. Throughout North America and much county in the US (Ohlendorf and others 1987, Presser of the world, birds are censused by recording the and Ohlendorf 1988). abundance of territorial singing males. It is generally As early as the 1940s, US Department of Agricul- assumed that each territorial male is accompanied by a ture soil scientists had identified high selenium (Se) breeding female. After documenting a decline in rela- levels in western San Joaquin soils, and by 1960 a tive numbers through the 1960s, the first comprehen- water resources report warned that drainage from the sive census of duskies was conducted in 1968 (Sharp Panoche Fan area was "unusable for beneficial pur- 1970). The population decline was closely monitored poses" (Schuhz 1985). The excess drainage from irri- through the 1970s, and when only five territorial gated lands contains Se concentrations 400 times the males were recorded in 1980 the decision was made to safe drinking water standard (10 ppb) established by remove the birds from the wild and begin a captive the US Environmental Protection Agency (Tanji and breeding program. Unfortunately, the decision was others 1986). By 1981, subsurface water from the tile- too late; no female birds remained, and the only dusky drained agricultural lands, undiluted by surface water, seaside sparrows left in the world were the five captive was being emptied into the Kesterson. By 1983, the 684 L.D. Harris

US Fish and Wildlife Service detected toxic poisoning may accumulate and be manifest in high trophic level of wildlife. organisms that depend upon the compound landscape Bioconcentration of selenium averages less than a system. Organochlorine residues derive from manu- 1000-fold increase for algae and plants in the reser- factured chemicals used throughout the world, princi- voir, but more than 1000-fold for aquatic inverte- pally as agricultural pesticides. These residues can be brates. Mosquito fish (Gambusia affinis) from the reser- concentrated by simple water flow processes such as voirs contain selenium concentrations 100 times described above for the inert element selenium. But greater than from nearby unirrigated areas, and high the phenomenon of biomagnification in vertebrate rates of reproductive failure and deformities have de- wildfife populations is an additive process that delays, veloped in American coots (Fulica americana), black- accumulates, and then amplifies the impacts of the necked stilts (Himantopus mexicanus), and other species toxic substance. Because many of the hydrocarbon that nest in the reservoir ponds (Ohlendorf 1986, Oh- compounds are fat soluble, they can be consumed, me- lendorf and others 1986a,b, 1987). Hunters have now tabolized, and stored in fatty tissues of animals, espe- been warned not to consume waterfowl that over- dally warm-blooded vertebrates. It is commonly held winter in the area because they contain Se concentra- that this biomagnification process may result in a 10- tions of 3.0-9.5 ppm, perhaps 10 times the normal fold greater concentration of pesticide in each succes- background level. Waterfowl and wading bird mor- sive trophic level. tality became so critical that the Secretary of Interior Over 150 species of North American breeding birds announced closure of the Kesterson Refuge on March migrate to the lower latitudes annually, and over 90% 15, 1985, because of violation of the 1918 Interna- of the bird species of eastern North America are re- tional Migratory Bird Treaty (Tanji and others 1986). corded to use southeastern wedands for one reason or The fact that three different ecosystem components another (Sprunt 1967). Among the raptors, there are mobile and transient but are artificially concen- exists a close correlation between the magnitude and trated in a single wetland site explains this accumula- importance of migration and the nature of their pri- tion of impacts. Irrigati/)n water comes from the Sac- mary food source. Fish-eating, bird-eating, and espe- ramento and San Joaquin rivers and from deep well cially insect-eating species are highly migratory and sources. It is suffused over an agricultural landscape, dependent upon southern habitats during the winter and then disposed of in the Kesterson ponds. Se- months. Small-bodied birds are more heavily depen- lenium is leached from the agricultural soils by the dent on migration than are large-bodied birds. percolating irrigation waters. Because the Kesterson The American kestrel (Falco sparverius) is a small, serves as a collector and a concentrator, Se becomes predominantly insectivorous raptor that breeds readily available there; it is further bioconcentrated by throughout eastern North America but overwinters in plants and invertebrate consumers as it enters the wa- the southeastern US. Organochlorine toxicity in birds terfowl food chain. is most dramatically revealed during the reproductive Most North American waterfowl production de- process (or at other times of severe physiological rives from an extensive northern breeding ground, is stress) when fat reserves are mobilized and the toxin concentrated along migration corridors, and finally reenters the systemic system. By affecting calcium me- overwinters, in this case, at a man-made and ostensibly tabolism, it becomes manifest as eggshell thinning, human-enhanced wetland site. The loss of 95% of Cal- whether the adult females acquire the pesticides on ifornia's historic wetland (Gilmer and others 1982.) the breeding or overwintering grounds. (Eggshell greatly amplifies the importance of the few remaining thinning is caused by food-chain accumulation of sites, such as the Kesterson. Because waterfowl are dichloro-diphenyl-dichloroethylene, DDE, which is long lived relative to the organisms they consume, they formed from DDT by organisms lower in the food serve as bioconcentrators of micronutrients from the chain. DDE is both more persistent and more biologi- tissues of thousands of different seeds, plants, and cally active than the chlorinated organic pesticide benthic invertebrates. As a consequence of all three DDT.) Although a 20% thinning of the eggshells is anthropogenic concentrations (water, selenium, and sufficient to cause serious reproductive failures among waterfowl), the impact of a spatially removed and most bird species, there is perhaps a 10-fold difference perhaps low-intensity activity accumulates in the verte- in DDE levels necessary to cause different taxa of birds brates in and around the Kesterson. to reach this threshold level (Figure 4). After cessation of the use of organochlorine pesticides in New York Actions Occurring within a Component Subsystem State, American Kestrels continued to exhibit high Actions that occur within a component subsystem tissue concentrations of DDE residues: Cumulative Impacts on Vertebrate Biodiversity 685

(/3 some fifty nesting rookeries were visited, some of them containing ~.1 IO0- Z upwards of one hundred thousand nests.... The island, upwards of eight acres in extent, was nearly all occupied by the rookery. A count 0 ~~ of all the nests in a small section showed that there were many more I-- than one hundred thousand nests of birds of all kinds on the island -J 50" _J (Ward 1914). Lz.I "1" O3 kestrel pelican cormorant booby As late as 1934, one mixed-species nesting colony was estimated to contain as many as 250,000 pairs o tto zto (Allen 1958, in Ogden and others 1987). Roosting and PPM DDE in EGGS nesting generally occur in trees, with colonies ideally Figure 4. Relationship between the amount of dichloro-di- located on isolated islands or sites that provide protec- phenyl-dichloroethylene (DDE) in eggs and shell thickness, tion from predators. Diurnally, the birds disperse over for selected bird species. Variable concentrations of toxin the shallow marshes and Florida Bay to forage on may be tolerated before a specified level of impact, such as aquatic animals (Ogden 1978, Kushlan and others 20% loss of eggshell thickness, occurs. A 20% loss is sufficient 1985). to cause serious reproductive failures in most bird species. Using recent wading bird relative abundance data Modified from Keith and Gruchy 1970. (Kushlan and White 1977), a weighted average food consumption rate of 100 g/day (Schramm and others The sources of the high DDE residues in eggs of local wild kestrels 1987), and a mineral ash food composition of 2.2% and the degree of shell-thinning cannot be fully explained by the low (Rottiers and Tucker 1981), it follows that the average residue levels in prey items on the breeding grounds. The fact that breeding kestrels, feeding on a relatively uncontaminated food chain, wading bird withdraws 0.8 kg of minerals per year can produce highly contaminated and thin-shelled eggs points out the from the aquatic environment: importance of the quality of the wintering grounds to this and other migratory species. The disastrous role played by the far-removed, but (100 g food/d) x (365 d/yr) x (0.022 g ash/g food) inordinately contaminated, winter prey once again dramatically • (10 s g/kg) = 0.8 kg ash/bird/yr points out the global nature of the biocide problem (Lincer and Sher- burne 1974). A population of 2.0 million wading birds would withdraw 1.6 million kg of mineral ash from the As in other cases, the effect of the environmental aquatic environment per year. Assuming that 50% of toxicant was delayed in time, magnified in the trophic the excrement (i.e., half of daily intake) and 50% of all pyramid, and offset in space. Effects that appear as bird mortality occurred within the colony site, then direct and immediate within a component subsystem 25% of the total mineral ash withdrawal would be may well accumulate as indirect time- and space- translocated from the aquatic environment to the lagged impacts in the larger compound system. It can colony sites annually. By conservative estimates, a pop- reasonably be predicted that additional chain reaction ulation of 2.0 million wading birds would translocate effects will reverberate downward to component sub- 400 metric tons (400,000 kg) of mineral ash per year. systems in response to population reductions and/or In the 100 years since 1885, this translocation would elimination of landscape integrators (top carnivores). have amounted to 40,000 metric tons of mineral ash. Processes that are diffuse or only weakly operative Whereas numerous studies (Gillham 1956, Leent- among numerous landscape components may well ap- vaar 1967, Dobrowski 1973, Smith 1979) have ad- pear as, or engender, more concentrated and tangible dressed the consequences of avian guano transloca- effects within a component subsystem. tion, only a few have been directed at Florida environ- ments (Odum and Heald 1972, Onuf and others 1977, Diffuse Processes in Compound Landscape Watson 1986). Nonetheless, it has been demonstrated Diffuse processes in the compound landscape may that insect herbivory in bird-fertilized red mangroves have concentrated effects within component sub- (Rhizophora mangle) increased fourfold over that of un- systems. Throughout the 19th century perhaps 2.5 fertilized stands and that the difference disappeared million wading birds (principally Ciconiformes) inha- when the birds migrated from the area (Onuf and bited the Everglades area at the southern end of the others 1977). Watson (1986) demonstrated that taller Florida peninsula (Robertson and Kushlan 1974). A and sturdier black mangroves (Avicennia germinans) key characteristic of these birds is that most return were preferentially utilized as colony sites by brown nighty to colonial roots and also concentrate their pelicans (Pelicanusoccidentalis) and that these trees were nesting and reproduction in such colonies. One au- on average 1.7 m taller and grew significantly more in thoritative reviewer of 19th century conditions stated: circumference per season than did the mangroves not 686 L.D. Harris

utilized by pelicans. Thus, while notable accumulations NUMBER of WADING BIRDS of guano only occur in dry, cold, or sheltered environ- 2.6" ments, the effects of guano translocation between dif- 2.4 SR[[mN~ PAIRS O~ WOOOs,Or, XS ferent components of wetland landscapes may be no 2.2 =N r .AWO.AL e,mx ooo .~ less significant to ecosystem function. This is particu- 2.0 ", \ larly true in nutrient-poor areas such as the Ever- \ \ 1.8 ', \ glades. i t \ If the magnitude of the process involved only a few 1.6 ~ =ooo. X~ kilograms, the impacts might be negligible, but it 1.4, \ seems highly unlikely that a process involving the re- l \ -J 1.2 \ distribution of 400 metric tons of nutrient per year or \\ //~"~ \.\\ 40,000 tons during the last century could be trivial. 1.0 Even more dramatic (but not wetland) examples, such 0.8 as the spatial translocation of nutrients by hundreds of 0.6 millions of passenger pigeons, must surely be relevant 0.4 to issues such as forest patch dynamics and the long- 0.2 term reduction in forest productivity that has been re- 1 corded for several regions (e.g., Sheffield and others 1870 1900 1930 1960 1990 1985). Figure 5. Long-term decrease in the wading bird popula- Natural area reserves, such as Everglades National tions of south Florida and in the wood stork population (in- Park, rarely constitute more than a token vignette of sert) within Everglades National Park since its creation in the original functioning landscape system. In this case, 1947. Data from Robert.son and Kushlan 1974, Kushlan and Everglades Park constitutes only 6% of the original Ev- White 1977, Ogden and others 1987. erglades system. Wading.birds such as the wood stork (Mycteria americana) forage throughout the landscape as far as 80 km from the nest colony and are depen- Food Chain Support and Loss of Habitat dent upon the frequency, amplitude, and seasonal Food chain support may interact with loss of timing of water levels that concentrate the fish prey in nesting habitat either to cause or at least contribute to drying pools of surface water (Kushlan 1987, Ogden the demise of species and groups of species. The and others 1987). Virtually no degree of management swallow-tailed kite (Elanoidesforficatus) is one of many of the habitat within the park can compensate for poor species of North American neotropical migrant birds management of the landscape outside the park whose population levels show long-term decline (Kushlan 1979, Harris 1984, Schonewald-Cox and (Figure 6). Like several other species on the decline Bayless 1986). The one to threemillion wading birds (see Briggs and Criswell 1979, Robbins 1979, 1980, that inhabited south Florida 100 years ago have been Robbins and others 1986), the kite is classified as an reduced to perhaps 16,000 breeding birds (1%) by "area-sensitive" species because its occurrence and re- 1987 (Figure 5) (M. Collopy, personal communica- productive success is dependent on extensive tracts of tion). This drastic reduction reflects the cumulative habitat (Hamel and others 1982), in this case riparian impact of plume hunting at the turn of the century, forest interspersed with open wetland foraging areas. habitat loss and degradation in this century, and, most Not only is the total acreage of riparian forest de- importantly, water management decisions that are still clining rapidly (MacDonald and others 1979, Frayer debated today (Robertson and Kushlan 1974, Kushlan and others 1983, Abernethy and Turner 1987, Harris 1987, Ogden and others 1987). Park authorization in and Gosselink 1988), but fragmentation into smaller 1934 and dedication of Everglades National Park in and more isolated tracts (Figure 7) may lead to equal 1947 have not abated the decline, because the biota or even more serious indirect effects (e.g., Whitcomb within the park are critically linked to water manage- and others 1981, Harris 1984, Harris and Wallace ment in the landscape surrounding the park. 1984, Wilcove and others 1986). The destruction and In the words of one authority: "... while ecosystem fragmentation of riparian forest as well as selective management in reserves can provide habitats for colo- logging of large, emergent trees that serve as pre- nial water birds, this, alone may not protect popula- ferred nest sites are surely factors in the decline of tions. Such reserves maintain only a residuum of the swallow-tailed kites (Cely 1979). support system and population, usually for only part Like most other kites, the swallow-tailed kite is mor- of the year. A regional strategy is imperative" phologicaUy and behaviorally adapted to forage as an (Kushlan 1983). aerial predator, in this case on wetland macroinverte- Cumulative Impacts on Vertebrate Biodiversity 687

Interactions of Land Use, Habitat Fragmentation, Alien Species, and Native Vertebrates

Interactions of land use, habitat fragmentation, alien species, and native vertebrates frequently accu- mulate to levels that are catastrophic for native species. Although intensity of competition, parasitism, and predation have frequently been analyzed at the popu- lation and community levels, only rarely have the con- cepts been applied at the level of the ecosystem (Esch and others 1975). Yet, Odum (1985) proposes that in- creased levels of parasitism and species loss might be anticipated as an ecosystem response to stress. Early explorers of the Great Plains described a flocking bird (lVlolothrus ater) in close association with American bison (Bison bison). The advantages of re- maining close to the nomadic bisOn apparently pro- vided sufficient selective advantage to cause the buf- falo bird to become North America's only obligate nest parasite (Friedmann 1929). Simultaneous with the / Figure 6. Distribution of swallow-tailed: kite in the United near extirpation of the bison, cattle were introduced, States, 1880-1985. The swallow-tailed kite is one of many and the greatly opened and fragmented forest of the North American neotropical migrant bird species that have eastern states provided the bird with a substitute large been greatly reduced in number and breeding range during ungulate in a greatly modified environment (the bird's recent decades. The reduction no doubt represents the cu- name was conveniently changed to brown-headed mulative effect of several land use and wetland management cowbird). Brittingham and Temple (1983) have factors. Modified from Cely 1979. charted the frequency of occurrence of cowbirds on eastern bird counts from 0% late in the last century to 100% by the 1980s. Because of the cowbird's evolution brates, such as dragonflies, and cold-blooded verte- on the Great Plains and prairies of the Midwest, the brates, such as amphibians and reptiles (e.g., Snyder females are largely restricted to open lands associated and Wiley 1976). Concurrent with the reduction of ri- with fields, homesteads, powerlines, and forest parian forest nesting habitat, large proportions of clearings. Whereas nest parasitism by female cowbirds open wetland that produced the food base and served is essentially nonexistent in the interior of large forest as kite foraging areas were also lost. The effects of stands, nearly all open nests near clearings or forest stream, river, and wetland pollution have probably re- edges may be parasitized (Brittingham and Temple duced production of macroinvertebrates such as drag- 1983). Effects of cowbird parasitism on species such as onflies (Franz 1982) to the detriment of swallow-tailed Kirtland's warbler (Dendroica kirtlandii) are so severe kites. Perhaps pesticide residues are also implicated. that, in addition to habitat management, an active This example illustrates not only the interrelation cowbird-trapping program has been implemented and of habitat and food chain support necessary to main- the specific details of a captive-breeding program for tain area-sensitive or wide-ranging species, but also Kirtland's warbler are being explored. how dramatic the portrayal of cumulative impacts, can Throughout the 19th century Bachman's warbler be in the face of almost total ignorance of causality. (Vermivora bachmaniz) nested in gaps in the pristine While experimental research directed toward isolating forest of the Southeastern Coastal Plain (Hamel 1986). and verifying causal, proximal, single-factor mecha- Because the wetter of these forested openings were nisms is no doubt called for, the level of ongoing re- commonly invaded and dominated by canegrass (Ar- search is so low that it does not even document the undinaria spp.), they are frequently referred to as bird's declining status (this species is not listed in the canebrakes. Cane becomes very dense after several breeding bird status report). Thus, while the cumula- years without burning, and because it is intolerant of tive impact seems dear, and the ultimate causes (wet- its own shade, its vigor, reproduction, and perpetua- land fragmentation and area loss and prey base reduc- tion are dependent upon the natural occurrence of tion) seem clear, establishment of specific causal mech- fire (Shepherd and others 1951, Hughes 1966). Cane anisms seems unlikely in this century. is also vulnerable to heavy, sustained grazing, and re- 688 L_. D. Harris

1962 1979 NORTH CAROLIN/ S POCOSlNS Figure 7. Examples of habitat fragmentation over time. Fragmentation of formerly expansive tracts of wetland habitat into isolated patches causes change in mahy aspects of habitat structure and ecological function. Upper panel: Jackson County, Arkansas. From US Fish and Wildlife Service, n.d. Lower panel: Pocosins, North Carolina. Modified from Richardson 1981.

quires periodic respite from intense grazing in order bird and increased densities of other potential nest to maintain vigor (Hughes and others 1960). Thus, for predators and competitors such that Bachman's war- somewhat different reasons, both cane and Bachman's bler was apparendy forced to extinction. warbler occurred as gap-phase species on wet sites in No single factor, or combination of factors, is de- the pre-Columbian coastal plain. monstrably responsible for this extinction. Yet the Bachman's warbler has not been observed in 10 combination of direct and indirect factors definitely years and is now presumed extinct (P. Hamel, per- accumulated to levels beyond the ability either of sci- sonal communication). The combination of relentless entists to predict extinction or of conservationists to and/or intense grazing since the introduction of cattle, prevent it. overly protective fire exclusion programs throughout Loss of genetic integrity may result from accumu- the first half of this century, and active forestation pro- lated impacts that affect population distribution and grams seems to have greatly reduced the occurrence demography. Concurrent with increased human den- of canebrakes and their suitability as habitat for species sity and land use intensity, both of the large mammal such as Bachman's warbler. The combination of a gen- carnivores native to the southeastern United States-- erally open landscape with ubiquitous cattle facilitated the red wolf and the Florida panther--became in- range expansion of the parasitic brown-headed cow- creasingly limited to isolated wetlands and seem to Cumulative Impacts on Vertebrate Biodiversity 689

have accumulated severe demographic and genetic ef- are more prevalent and pathogenic in warm moist en- fects as a consequence. The following have been sug- vironments with wet soil than they are in more arid gested as probable causes for decline: (1) loss of hab- upland environments (D. Forrester, University of itat and increasing fragmentation and isolation of re- Florida Veterinary College, personal communication ) . maining wetland areas, (2) amplified man-induced Thus, as the red wolf became more restricted to wet- mortality beginning with active predator control and lands, the incidence and pathogenicity of these para- ultimately involving accidental death from hunters sites may well have increased. The introduction of and automobile traffic, (3) depreciated prey base free-ranging domestic dogs and the invasion of the within the wetlands and decreased access to prey out- landscape by coyotes provided a much higher canid side the wetland, (4) invasion of the habitat by alien population for heartworms to parasitize, and it is pos- species, and (5) amplified disease and parasite impacts. sible that this combination of factors accumulated to The red wolf (Canis rufus) formerly inhabited the critical levels. The few remaining red wolves thought Mississippi River Valley and the Gulf coastal plain to represent the original genetic stock were removed westward to east Texas and east through Florida and from the wild beginning in 1973 and were used to ini- Georgia. Consistent with Gloger's rule, the red wolf is tiate a captive breeding and reintroduction program characterized by dark pelage, and totally black spec- (USFWS 1979b). imens were not uncommon. Because the isolated The environmental circumstances described for the swamps and marshes in the Mississippi Valley and red wolf apply equally to the Florida panther (Felis along the Gulf coast constituted the last wilderness concolor co~yO. In this case the release of western areas of the South, remnant populations of red wolves cougars into Florida panther range allowed some in- remained in these wetlands until the 1960s (Nowak terbreeding and perhaps some loss of genetic integ- 1972, Lowery 1974). Under conditions of widespread rity, but this is not the principal issue. The last rem- distribution and viable population levels, wolf nant Florida panther population became isolated in breeding and demography were governed by strong the Everglades and Big Cypress areas of extreme behavioral mechanisms. The pack is the social unit, so- south Florida. Because of Florida's rapid population cial dominance governs breeding, females mate for growth throughout the state (4% per year for 100 life, and population level relative to available food re- years) and the long and increasingly narrow form of sources influences survivorship and productivity. Re- the peninsula (because of sea-level rise), the south lated species or subspecies maintain genetic distinctive- Florida cats have probably been reproductively iso- ness by a combination of geographical and ecological lated for a considerable length of time. Distinct genetic separation and species-specific behavioral patterns. maladies, possibly the result of inbreeding, are now When excessive levels of hunting, trapping, and other being documented. For example, all adult males from human-induced mortality impinge, populations are which data have been obtained manifest nearly 95% reduced and fragmented, and the social system is seri- infertile spermatozoa (Roelke 1986, USFWS 1987). ously disrupted or may collapse altogether. The extent to which this malady affects the demog- With increased land clearing, agricultural develop- raphy and population dynamics of panthers in the ment, and a general "opening-up" of the Eastern wild is yet unknown. Nonetheless, because animals of forest landscape, the coyote (Canis latrans) expanded the higher trophic levels are rare under the best of its range eastward and gradually increased in abun- circumstances, they seem particularly vulnerable to dance throughout the southeastern US. Because it is isolation, demographic variations, and inbreeding. more of a habitat and food-habits generalist (Gipson Thus the loss of genetic diversity from within the 1974), the same forces that militated against the red species of wide-ranging top carnivores may translate wolf served to benefit the smaller, more omnivorous directly to loss and/or diminished function (e.g., pre- coyote. The combination of reduced and disrupted dation and competition) of those carnivores in the wolf populations that were behaviorally and demo- landscape-level system and a subsequent loss of species graphically stressed, and increasingly abundant diversity at the community level. coyotes that were no longer geographically or ecologi- The home-range size of Florida panthers is very cally separated from the wolf, led to interbreeding be- large (adult males, 500 kin2), and the males can be tween wolves and coyotes. The consequent genetic lethally territorial. Thus, little ecological compression "swamping" of the wolf led to a virtual elimination of can occur, and the relation between viable population the native red wolf genome (Shaw and Jordan 1977, and viable area may be deterministic. No single refuge USFWS 1979b, Truett and Lay 1984). of even a single tract of a few million km 2 will maintain Both hookworms and mosquito-borne heartworms a viable population in perpetuity. Only by maintaining 690 L.D. Harris

core populations within large tracts of public land and Briggs, S., and J. Criswell. 1979. Gradual silencing of spring facilitating movement throughout the intervening in Washington. Atlantic Naturalist 32:19-26. landscape matrix will we be able to maintain these im- Brinson, M. M. 1988. Strategies for Assessing the Cumulative portant elements of native biotic diversity (MacClin- Effects of Wetland Alteration on Water Quality. Environ- mental Management 12:xx-xx. tock and others 1977, Harris 1985, Branan 1986). Brittingham, M., and S. Temple. 1983. "Have cowbirds caused forest songbirds to decline? BioSdence 33:31-35. Conclusion Bull, J. 1980. Sex determination in reptiles. ~uarterly Rev/ew of Biology 55:3-21. The above examples of cumulative impacts on ver- Cain, S.A. 1966. Biotope and habitat. Pages 38-54 in F. F. tebrate biotic diversity demonstrate the breadth of the Darling and J. P. Milton (eds.), Man's role in changing the issue. Loss of genetic diversity within species may, in face of the earth. University of Chicago Press, Chicago. the long run, be as serious as loss of the species them- Cely, J. 1979. Status of the swallow-tailed kite and factors af- selves. Similarly, loss of diversity at the landscape level fecting its distribution. Pages 144-150 in D. Forsythe and W. Ezell Jr. (eds.), Proceedings of the First South Carolina may well lead directly to loss of species and genetic Endangered Species Symposium. South Carolina Wildlife diversity from within species. Land-use managers and and Marine Resources Department, Columbia, South Car- decision makers, faced with the need to consider main- olina. tenance of species diversity among their goals; are in a Chase, A. 1986. Playing God in Yellowstone. Atlantic difficult position. At present, because of the com- Monthly Press, Boston, 446 pp. plexity of trophic interactions, we do not have the Chepko-Sade, B., and Z. Halpin (eds.). 1987. Mammalian dis- ability to predict the results of three or more different persal patterns. University of Chicago Press, Chicago, 342 perturbations, although we can ~ecognize those results pp. retrospectively. Realistically, until much more infor- Craighead, F., and J. Craighead. 1956. Hawks, owls, and mation and understanding are developed, the best wildlife. Stackpole, Harrisburg, Pennsylvania, 443 pp. managers and decision makel-s can do may simply be Cronin, W. 1983. Changes in the land. Hill and Wang, New York, 241 pp. to be aware of the histor~ of impacts of varied land-use actions and alert to parallels with them in their.own Crosby, S. 1986. Ecological imperalism, the biological expan- sion of Europe, 900-1900. Cambridge University Press, activities. At the same time, establishing biotic reserves Cambridge, 368 pp. must be accomplished. However, maintenance of Deems, E., and D. Pursley (eds). 1983. North American fur- biotic diversity cannot be achieved by simply setting bearers, a contemporary reference. Worldwide Furbearer aside reserves and disregarding the landscape-level Conference, Washington, DC, 217 pp. processes occurring around them. Only by establishing Dobrowski, K. 1973. Role of birds in Polish wetland eco- a system of replicate wetland reserves and allowing systems. Polish Archives of Hydrobiology 20:217-221. each to function in an interactive landscape context Elton, C. 1966. The pattern of animal communities. can the maintenance of wetland biodiversity be as- Methuen, New York, 432 pp. sured. Errington, P. L. 1963. Muskrat populations. Iowa State Uni- versity Press, Ames, Iowa, 665 pp. Esch, G.,J. Gibbons, andJ. Bourque. 1975. An analysis of the Literature Cited relationship between stress and parasitism. American Mid- land Naturalist 93:339-353. Abernethy, Y., and R. Turner. 1987. U.S. forested wetlands: Forman, R. T. T., and M. Godron. 1986. Landscape ecology. Status and changes, 1940-1980. BioScience 37(10):721- John Wiley & Son s, New York, 619 pp. 727. Frankel, O., and M. Soule. 1981. Conservation and evolution. Baker, J. 1978. Dusky Seaside sparrow. Pages 16-19 in H. Cambridge University Press, Cambridge. 327 pp. Kale II (ed.), Rare and endangered biota of Florida, vol. 2. Birds. University Presses of Florida, Gainesville, Florida. Franz, K. (ed.). 1982. Invertebrates, vol. 6. Rare and endan- gered biota of Florida. University Presses of Florida, Barrett, G., and R. Rosenberg (eds.). 1981. Stress effects on Galnesville, Florida, 131 pp. natural ecosystems. John Wiley & Sons, New York, 305 pp. Frayer, W., T. Monahan, D. Bowden, and F. Graybill. 1983. Bormann, F. H., and G. E. Likens. 1979. Pattern and process Status and trends of wetlands'and deepwater habitats in in a forested ecosystem. Springer-Verlag, New York, the conterminous United States, 1950s to 1970s. Published 253 pp. for US Fish and Wildlife Service by the Department of Branan W. (ed.). 1986. Survival of the Florida panther, a dis- Forest and Wood Sciences, Colorado State University, Fort cussion of issuesand accomplishments. Florida Defenders Collins, Colorado, 31 pp. of the Environment, Tallahassee, Florida, 67 pp. Friedmann, H. 1929. The cowbirds, a study in the biology of Bray, J. R. 1956. Gap phase replacement in a maple-bass- social parasitism. Charles Thomas, Baltimore, Maryland, wood forest. Ecology 37:598-600. 421 pp. Cumulative Impacts on Vertebrate Biodiversity 691

Gilbert, L. 1980. Food web organization and the conservation Kushlan, J. 1979. Design and management of continental of neotropical diversity. Pages 11-33 in M. Soule and B. wildlife reserves: Lessons from the Everglades. Biological Wilcox (eds.), Conservation biology. Sinauer, Sunderland, Conservation 15:281-290. Massachusetts. Kushlan, J. 1983. Special species and ecosystem preserves: GiUham, M. 1956. Ecology of the Pembrokeshire Islands. V. Colonial waterbirds in US national parks. Environmental Manuring by the colonial seabirds and mammals, with a Management 7:201-207. note on seed distribution by gulls. Journal of Ecology Kushlan, J. 1987. External threats and internal management: 44:429-454. The hydrologic regulation of the Everglades, Florida, Gilmer, D., M. Miller, R. Bauer, andJ. LeDonne. 1982. Cali- USA. Environmental Management 11:109-119. fornia's central valley wintering waterfowl: Concerns and Kushlan, J., and D. White. 1977. Nesting wading bird popu- challenges. Transactions of the North American Wildlife Natural lations in southern Florida. Flm4d~ Scientist 40:65-72. Resources Conference 47:441-452. Kushlan, J., G. Morales, and P. Frohring. 1985. Foraging Gipson, P. 1974. Food habits of coyotes in Arkansas. Journal niche relations of wading birds in tropical wet savannas. of Wildlife Management 38:848-853. Ornithological Monographs 36:663-682. Hamel, P. 1986. Bachman's warbler: A species in peril. Leentvaar, P. 1967. Observations in guanotrophic environ- Smithsonian Institution Press, Washington, DC, 110 pp. ments. Hydrobiologia 29/30:441-489. Hamel, P., H. LeGrand Jr., M. Lennartz, and S. Gauthreaux, Lincer, J. L., and J. A. Sherburne. 1974. Organochlorines in Jr. 1982. Bird-habitat relationships On southeastern forest kestrel prey: A North-South dichotomy. Journal of Wildlife lands. US Forest Service General Technical Report SE 22, Management 38(3):427-434. U.S. Forest Service, Washington, DC, 417 pp. Lowery, G. 1974. The mammals of Louisiana and its adjacent Harris, L. 1978. El pantano de los espanoles: A job comple- waters. Louisiana State University Press, Baton Rouge, tion report on the interpretation and management of Louisiana, 565 pp. Spanish Pond, Fort Caroline National Memorial. School of MacClintock, L., R. Whitcomb, and B. Whitcomb. 1977. Evi- Forest Resources and Conservation, University of Florida, Gainesville, Florida, 197 pp + appendix. dence for the value of corridors and minimization of isola- tion in preservation of biotic diversity. American Birds Harris, L. 1984. The fragmented forest: Application of is- 31:6-16. land bi0geography principles.to pi'eservation of biotic di- MacDonald, P., W. Frayer, andJ. Clauser. 1979. Documenta- versity. University of Chicago Press, Chicago, 211 pp. tion, chronology, and future predictions of bottomland Harris, L. 1985. Conservation corridors: A highway system hardwood habitat losses in the lower Mississippi alluvial for wildlife. ENFO Report of the Florida Conservation plain, vols. 1 and 2. US Fish and Wildlife Service, Jackson, Foundation, Winter Park, Florida, 10 pp. Mississippi, 133 pp. and 259 pp. Harris, L., and J. Eisenberg. 1988. Enhanced linkages: Nec- Mayr, E. 1982. The growth of biological thought. Harvard essary steps for success in conservation of faunal diversity. University Press, Cambridge, Massachusetts, 974 pp. In M. Pearl and D. Western (eds.), Conservation 2100. Ox- ford University Press, Oxford. Morreale, S., G. Ruiz, J. Spotila, and E. Standora. 1982. Tem- perature-dependent sex determination: Current practices Harris, L., and L. Gosselink. 1988. Cumulative impacts of threaten conservation of sea turtles. Science 216:1245- bottomland hardwood conversion on hydrology, water 1247. quality, and terrestrial wildlife. Chapter in J. Gosselink, L. Lee, and T. Muir (eds.), Bottomland hardwood wetlands: Noss, R., and L.D. Harris. 1986. Nodes, networks, and ecological and regulatory perspective, in press. MUMs: Preserving biotic diversity at all scales. Environ- Harris, L., and T. O'Meara. 1988. Changes in southeastern mental Management 10:299-309. bottomland forests and impacts on vertebrate fauna. In R. Nowak, R. 1972, The mysterious wolf of the south. Na'tural Sharitz and J. Gibbons (eds.), Freshwater wetlands and History 81:50-53, 74-77. wildlife. Oak Ridge Press, Oak Ridge, Tennessee, in press. Odum, E.P. 1985. Trends expected in stressed ecosystems. Harris, L., and R. Wallace. 1984. Breeding bird species in B ioScience 35:419-422. Florida forest fragments. Proceedings of the Annual Con- ference of the Southeastern Association of Fish and Wild- Odum, E. P., J. T. Finn, and E. H. Franz. 1979. Perturbation life Agencies 38:87-96. theory and the subsidy-stress gradient. BioScience 29:349- Hughes, R. 1966. Fire ecology of canebrakes. Proceedi~gs of 352. the Tall Timbers Fire Ecology Conference 5:148-158. Odum, W., and E. Heald. 1972. Trophic analysis of an es- Hughes, R., E. DiUard, andJ. Hilmon. 1960. Vegetation and tuarine ecosystem. Bulletin of Marine Science 22:671-738. cattle response under two systems of grazing cane range in North Carolina. North Carolina Agricultural Experiment Ogden, J. 1978. Freshwater marshlands and wading birds in Station Bulletin 412, 27 pp. south Florida. Page xvi/n H. Kale (ed.), Rare and endan- Kale, H. 1977. Dusky seaside sparrow. Florida Naturalist gered biota of Florida, vol. 2. Birds. University Presses of 50:I6-21. Florida, Gainesville, Florida. Keith, J., and I. Gruchy. 1970. Residue levels of chemical pol- Ogden, J., w. Loftus, and W. Robertson, Jr. 1987. Wood lutants in North American birdlife. In Proceedings of the storks, wading birds, freshwater fishes. Unpublished report 15th International Ornithological Congress, E.J. Brill, to US Army Corps of Engineers in preparation of a Gen- Leiden, The Netherlands. eral Design Memorandum for south Florida water man- 692 L.D. Harris

agement. US Army Corps of Engineers, Jacksonville, Roelke, M. 1986. Medical management, biomedical finding s, Florida. and research techniques. Pages 7-14 in Survival of the Ohlendorf, H. 1985. Aquatic birds and selenium in the San Florida panther: A discussion of issues and accomplish- Joaquin Valley. Pages 15-24/n Symposium Proceedings: ments. Conference Proceedings, Florida Defenders Of the Selenium and Agricultural Drainage, vol. 2. Bay Institute Environment, Tallahassee, Florida. of San Francisco, US Fish and Wildlife Service, and Uni- Root, R. B. 1973. Organization of a plant-anthropod associa- versity of California, Berkeley, California. tion in simple and diverse habitats: The fauna of collards Ohlendorf, H. 1986. Bioaccumulation and effects of se- (Broasiva oleracea). Ecological Monographs 43:95-124. lenium in wildlife. Symposium Proceedings: Selenium in Rottiers, D., and R. Tucker. 1981. Proximate composition Irrigated Agriculture. American Society of Agronomy, in and caloric content of eight Lake Michigan fishes. US Fish press. and Wildlife Service Technical Paper 108, US Fish and Ohlendorf, H., R. Hothem, C. Bunck, T. Aldrich, and J. Wildlife Service, Washington, DC, 8 pp. Moore. 1986a. Relationships between selenium concentra- Rykiel, E.J., Jr. 1985. Towards a definition of ecological dis- tions and avian reproduction. Transactions of the North Amer- turbance. Australian Journal of Ecology 10:361-365. ican Wildlife and Natural Resources Conference 51:330- 342. Schonewald-Cox, C., and J. Bayless. 1986. The boundary Ohlendorf, H., D. Hoffman, M. Saiki, and T. Aldrich. 1986b. model: A geographical analysis of design and conservation Embryonic mortality and abnormalities of aquatic birds: of nature reserves. Biological Conservation 38:305-322. Apparent impacts of selenium from irrigation drainwater. Schramm, H. Jr., M. Collopy, and E. Okrah. 1987. Potential Science in the Total Environment 52:49-63. problems of bird prediation for fish culture in Florida. Pro- Ohlendorf, H., R. Hothem, T. Aldrich, and A. Krynitsky. gressive in Fish Culturist 49:44-49. 1987. Selenium contamination of the grasslands, a major Schultz, A. 1985. Background and recent history. Sympo- California waterfowl area. Science and the Total Environment sium Proceedings: Selenium and Agricultural Drainage, 53, in press. vol. 2. Pages 3-9 in Bay Institute of San Francisco, US Fish Onuf, C. J. Teal, and I. Valiela. 1977. Interactions of nu- and Wildlife Service, and University of California, trients, plant growth, and herbivory in a mangrove eco- Berkeley, California. system. Ecology 58:514-526. Sharp, B. 1970. A population estimate of the dusky seaside Penn, G. H. 1950. Utilization of crawfishes by cold-blooded sparrow. Wilson Bulletin 8:158-166. vertebrates in the eastern United States. American Midland Shaw, J., and P. Jordan. 1977. The wolf that lost its genes. Naturalist 44:643-658. Natural History 86:80-88. Perkins, C. 1973. Effects of clearcutting and site preparation Sheffield, R., N. Cost, W. Bechfold, J. McClure. 1985. Pine on the vegetation and wildlife in the flatwoods of Kemper growth reductions in the Southeast. US Department of county, Mississippi. Dissertation. Mississippi State Univer- Agriculture Resource Bulletin SE-83, Asheville, North sity, Mississippi State, 236 pp. Carolina, 112 pp. Presser, T., and H. Ohlendorf. 1988. Biogeochemical cycling Shepherd, W, E. Dillard, and H. Lucas. 1951. Grazing and of selenium in the San Joaquin Valley, California, USA. fire influences in pond pine forests. North Carolina Agri- Environmental Management 11:805-821. cultural Experiment Station Technical Bulletin 97, Chapel Richardson, C. 1981. Pocosin wetlands: An integrated anal- Hill, North Carolina, 56 pp. ysis of coastal plain freshwater bogs in North Carolina. Shugart, H. H. 1984. A theory of forest dynamics. Springer- Hutchinson Ross Publishing Company, Stroudsburg, Verlag, New York, 278 pp. Pennsylvania, 364 pp. Richter, A., and R. Labisky. 1985. Reproductive dynamics Smith, V. 1979. The influence of seabird manuring on the among disjunct white-tailed deer herds in Florida. Journal phosphorus status of Marion Island (subantarctic) soils. Oe- of Wildlife Management 49:964-971. cologia (Berlin) 41:123-126. Robbins, C. 1979. Effect of forest fragmentation on bird pop- Snyder, N., and J. Wiley. 1976. Sexual size dimorphism in ulations. Pages 198-212 in R. DeGraaf and K. Evans (eds.), hawks and owls of North America. Ornithological Mono- Proceedings of the Workshop on Management of North graphs 20:96 pp. Central and Northeastern Forests for Nongame Birds. US Sprunt, A. 1967. Values of the South Atlantic and Gulf Coast Forest Service GTR NC51, US Forest Service, Washington marshes and estuaries to birds other than waterfowl. Pro- DC. ceedings of the Marsh and Estuary Management Sympo- Robbins, C. 1980. Effect of forest fragmentation on breeding sium, Louisiana State University, Baton Rouge, Louisiana. bird populations in the Piedmont of the mid-Atlantic re- gion. Atlantic Naturalist 33:31-36. Tanji, K., A. Lauchli, and J. Meyer. 1986. Selenium in the San Joaquin Valley. Environment 28:38-49. Robbins, C., D. Bystrak, and P. Geissler. 1986. The breeding bird survey: Its first fifteen years, 1965-1979. US Fish and Temple, S. 1986. Predicting impacts of habitat fragmentation Wildlife Service Resource Publication !57, US Fish and on forest birds: A comparison of two models. Pages Wildlife Service, Washington, DC, 196 pp. 301-304 in J. Verner, M. Morrison, and C. Ralph (eds.), Wildlife 2000. University of Wisconsin Press, Madison, Robertson, W., and J. Kushlan. 1974. The southern Florida Wisconsin. avifauna. Pages 414-452/n P. Gleason (ed.), Environments of south Florida: Past and present. Miami Geological So- Terborgh, J. 1986. Keystone plant resources in the tropical dety Memoir 2, Miami, Florida. forest. Pages 330-344 in M. Soule (ed.), Conservation bi- Cumulative Impacts on Vertebrate Biodiversity 693

ology: The science of scarcity and diversity. Sinauer Asso- Ward, C. 1914. Florida's wealth of bird life. Chapter 12 in G. ciates, Sunderland, Massachusetts. Chapin (ed.), Florida, 1513-1913, four hundred years of Truett, J., and D. Lay. 1984. Land of bears and honey. Uni- wars and peace and industrial development. S.J. Clarke, versity of Texas Press, Austin, Texas, 176 pp. Chicago, illinois. USFWS (Fish and Wildlife Service). 1979a. Dusky seaside Watson, A. 1986. Nutrient-production relations at Seahorse sparrow. Species account in Endangered and threatened Key Lagoon, Florida: Some consequences of shorebird ac- species of the southeastern United States. US Fish and cumulations. M.S. thesis. University of Florida, GainesviUe, Wildlife Service, Adanta, Georgia. Florida, 87 pp. USFWS (Fish and Wildlife Service). 1979b. Red wolf. Species Whitcomb, R., C. Robbins, J. Lynch, B. Whitcomb, M. Klim- account in Endangered and threatened species of the kiewicz, and D. Bystrak. 1981. Effects of forest fragmenta- southeastern United States. US Fish and Wildlife Service, tion on avifauna of the eastern deciduous forest. Pages Adanta, Georgia. 125-206 in R. Burgess and D. Sharpe (eds.), Forest island dynamics in man-dominated landscapes. Springer-Verlag, USFWS (Fish and Wildlife Service). 1987. Florida panther New York. (Felis concolor coryi). Recovery plan. Prepared by the Florida Whittaker, R. 1960. Vegetation of the Siskuyu Mountains, Panther Interagency Committee for the US Fish and Wild- Oregon and California. Ecological Monographs 30:279-338. life Service, Adanta, Georgia, 92 pp. Wilcove, D. 1985. Forest fragmentation and the decline of USFWS (Fish and Wildlife Service). n.d. Bottomland hard- migratory songbirds. Ph.D. Dissertation. Yale University, woods along the Mississippi: A disappearing resource. New Haven, Connecticut. Brochure. US Fish and Wildlife Service, Jackson, Missis- Wilcove, D., C. McLeUan, and A. Dobson. 1986. Habitat sippi. fragmentation in the temperate zone. Pages 237-256 in M. USFWS (Fish and Wildlife Service) and Canadian Wildlife Soule (ed.), Conservation biology: The science of scarcity Service. 1985. Status of waterfowl and fall flight forecasts. and diversity. Sinauer Associates, Sunderland, Massachu- Washington, DC, 32 pp. setts.