Coho salmon (Oncorhynchus kisutch) are native south of San Francisco Bay

Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org 441

IntroduCtIon

Establishing the boundary of any spe-cies’ range is a difficult task. Since by definition species are the most vulner-able to extinction at the edge of their range, establishing this boundary often depends on historical information. Kaczynski and Alvarado (2006:374) present evidence that supports their position that it is improbable that coho salmon (Oncorhynchus kisutch) main-tained self-sustaining populations south of San Francisco Bay, California. Their publication presents the first challenge to the previously accepted southern bound-ary of coho salmon at the San Lorenzo River, Santa Cruz County (Sandercock 1991; Moyle 2002). This boundary has been widely accepted since 1912 when Snyder (1912) provided the first specific

Peter B. Adams, Louis w. Botsford, kenneth w. Gobalet, robert A. Leidy, dennis r. mcewan, Peter B. moyle, Jerry J. smith, John G. williams, and ronald m. yoshiyama

Adams is a research fishery biologist at the noAA national Marine Fisheries Service Southwest Fisheries Science Center in Santa Cruz, California. He can be contacted at [email protected]. Botsford and Moyle are professors and Yoshiyama is a research associate in Wildlife, Fish, and Conservation Biology, University of California, Davis. Gobalet is a professor, Department of Biology, California State University, Bakersfield. Leidy is an ecologist with the U.S. Environmental Protection Agency, San Francisco, California. McEwan is a staff environmental scientist, California Department of Water Resources, Sacramento. Smith is a professor, Department of Biological Science, San Jose State University, San Jose, California. Williams is a fisheries consultant in Davis, California.

Coho salmon Are native south of san Francisco Bay:

A reexamination of north American Coho salmon’s southern range Limit

ABstrACt: Kaczynski and Alvarado (2006) have challenged the established southern boundary of coho salmon (Oncorhynchus kisutch) at the San Lorenzo River. They conclude that it is improbable coho salmon maintained self-sustaining populations south of San Francisco Bay, based primarily on evidence from early museum collections and literature, the archaeological record, analyses of ocean conditions, and suitability of habitat. They suggest that hatchery plantings were the source of these coho salmon south of San Francisco Bay. Using the same and new information, we are able to counter these statements. our examination of existing records found no reason to discount the coho salmon collections made in 1895 from streams south of San Francisco. Early distributional records state that coho salmon were abundant from San Francisco northward, but did not indicate coho salmon were absent south of San Francisco. Recent archeological evidence documents the presence of coho salmon in middens south of San Francisco prior to European habitation of the region. Furthermore, we found no creditable climatic, oceanographic, or ecological evidence for habitat differences between areas immediately north and south of San Francisco Bay. in fact, we believe that there is a more reasonable habitat and faunal break south of the San Lorenzo River, encompassing the allegedly controversial southern range of the coho salmon.

el salmón coho es nativo del sur de la Bahía de san Francisco:

una reevaluación del límite sur de la distribución del salmón

coho de norte Américaresumen: Kaczynski y Alvarado (2006) establecieron que el Río San Lorenzo es la frontera sur de la distribución del salmón coho (Oncorhynchus kisutch). Sobre la base de evidencia museográfica y documentaria, registros arqueológicos, análisis de condiciones oceánicas y del hábitat, concluyen que es improbable que existan poblaciones viables de salmón coho al sur de la Bahía de San Francisco. Sugieren que el origen de estos salmones en la Bahía de San Francisco son granjas de engorda. Utilizando la misma y nueva información, en el presente trabajo se confrontan dichos argumentos. nuestro examen de los registros existentes muestra que no hay razón aparente como para descartar las colecciones de salmón coho realizadas en 1895 en los ríos del sur de San Francisco. Los primeros registros de la distribución geográfica de esta especie indican que era abundante desde San Francisco hacia el norte, pero no prueban que el salmón coho estuviese ausente del sur de San Francisco. La evidencia arqueológica reciente documenta la presencia del salmón coho en el sur de San Francisco incluso antes de los asentamientos europeos en la región. Más aún, no se encontró evidencia suficiente en cuanto al clima, oceanografía o ecología que corrobore diferencias de hábitat entre las áreas inmediatas al sur y norte de la Bahía de San Francisco. De hecho, creemos que existe una mayor y razonable diferenciación faunística y de hábitat al sur del Río San Lorenzo, lo que comprende buena parte de la controversia sobre el ámbito sureño de la distribución del salmón coho.

FeAture:HiSToRY

442 Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org

LIterAture revIew: ACCurACy And PreCIsIon oF AGes estImAtes And BACk-CALCuLAtIon oF LenGtH

Accuracy and precision of age estimates

description of the southern boundary of coho salmon.

A close look at the evidence pre-sented by Kaczynski and Alvarado (2006), combined with other evidence, leads us to the opposite conclusion, i.e., coho salmon are native to streams south of San Francisco Bay. Evidence provided by Kaczynski and Alvarado (2006) falls into the following general categories:

1. early distributional records including museum specimens,

2. archaeological evidence, 3. suitability of habitat, and 4. hatcheries as the source of

southern coho salmon.

Their article also deals briefly with genetic evidence, straying, and ocean conditions. Here we examine their analysis point-by-point and provide additional new evidence. our analy-ses demonstrate that the historically accepted assumption that coho salmon populations south of San Francisco Bay were self-sustaining is the only conclu-sion supported by the evidence. Finally, much of current source material used by Kaczynski and Alvarado (2006) comes from the authors of this article and we wish to correct Kaczynski and Alvarado’s misinterpretation of our work.

eArLy dIstrIButIonAL reCords

Museum Collections. Museum col-lection records from the California Academy of Science (CAS) document the collection of 11 coho salmon from Waddell Creek (SU 4667) and 4 coho salmon from Scott Creek (SU 4797) on 5 June 1895 by C. Rutter, n. Scofield, C. Pierson, and A. Seale of Stanford University. Collections of 2 coho salmon from San vicente Creek (SU 4685) and 1 coho salmon from Gazos Creek (SU 4868) were made by the same party but were undated. it can be reasonably assumed that they were made during the same collecting trip (D. Catania, CAS, per. comm.), partly because the 1895 Carmel River Expedition was the only one in which all four collectors were recorded as participants (Boehlke 1953). The fish from these 1895 collections originally were misidentified as Chinook salmon (quinnat; O. tshawytscha) and chum salmon (dog; O. keta) in the Stanford collections. This mistake is not

unusual given the documented difficulty at that time in distinguishing salmon species in California (CFGC 1913; Snyder 1931). The correct identification of these fishes as coho salmon was made at some later date by unknown museum staff, probably before the Stanford col-lections were transferred to the CAS (D. Catania, CAS. pers. comm.). in addi-tion, further confusion ensued when the original species identifications were entered in the CAS electronic data col-lection system around 1990 and were again corrected in 1999.

Kaczynski and Alvarado (2006) chal-lenged the reliability of these samples, due to the original misidentification of the fish and to the changes in recording ledgers and archiving systems described above. However, the jars containing coho salmon in the CAS collection include their original locality labels and metal identification tags and there is no question that these were the fish col-lected by the Rutter party (D. Catania, CAS, pers. comm.) or that they are coho salmon. Kaczynski and Alvarado (2006) speculate that the specimens or labels may have gotten mixed up as the result of the San Francisco earthquake of 1906, in which some of the bottles containing the Stanford collection were broken. The curators of the collection strongly believe that this not the case and point to the meticulous procedures by then curator J. o. Snyder (and others). Specifically, broken jars were recorded and accounted for, with their specimens bearing a unique labels stating “Bottle broken during earthquake.” (D. Catania, CAS, pers. comm.) Specimens were discarded if it could not be determined which broken bottle they belonged to and specimens for which there was some doubt were placed in jars with labels.

Kaczynski and Alvarado (2006:380) also reject these CAS collections as evi-dence of coho salmon presence because “the chain of custody has been broken and the reliability of the specimens is questionable.” Chain of custody is an incongruous concept to apply to this situation because it is a legal concept in which evidence used in court proceed-ings is valid only if each time that evi-dence changes hands the transaction is carefully recorded and usually certified. We are unaware of this concept ever being applied to museum collections prior to the Kaczynski and Alvarado

(2006) paper. it is unlikely that any museum specimens collected prior to the 1906 San Francisco earthquake, or even most modern reference collections, could withstand this legal standard.

Finally, Kaczynski and Alvarado (2006:381) state that “nevertheless, even if the dates, locations, and species identifications associated with the speci-mens were valid, these specimens are not by themselves evidence of a persisting native population of coho salmon south of San Francisco.” Alternatively, they suggest that these fish may have been the result of some “obscure fish plant-ing activities prior to 1895” or that they are strays. Both of these assertions are speculation.

Early Literature. Kaczynski and Alvarado’s (2006) primary historic source that coho salmon were not native south of San Francisco is David Starr Jordan, the preeminent American ich-thyologist of the late nineteenth cen-tury. Jordan and his colleague Charles H. Gilbert first visited the San Francisco area in 1880 while conducting a survey of Pacific Coast fisheries for the U.S. Fish Commission (Jordan 1922). Jordan later returned in 1891 to become the first president of Stanford University, where he established a major ichthyologi-cal program and fish collection (Jordan 1922; Boehlke 1953). This collection was intended to be global rather than local (Boehlke 1953), which at least par-tially explains the relatively low number of coho salmon samples collected from anywhere in California, including from south of San Francisco.

Kaczynski and Alvarado (2006) cite many of Jordan’s publications on California fishes, but do not quote him directly. in those cited papers, the most common statement by Jordan is that coho salmon “is abundant from San Francisco northward.” (Jordan et al. 1882:308; Jordan and Evermann 1896:481; 1902:154; 1904:154; 1905:154). The next most frequent statement is “of these species, the…Silver salmon….pre-dominates in Puget Sound and in most of the streams found along the coast” (Jordan 1892a:11; 1892b:50). “Silver” salmon is a synonym of coho salmon; “king” and “quinnat” salmon are syn-onyms of Chinook salmon (O. tshaw-ytscha); and “dog salmon” is a synonym of chum salmon (O. keta), although it

Hallprint

Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org 443

was also used for any hook jawed adult salmon. Here Jordan also states “only the King Salmon has been noticed south of San Francisco.” in Jordan (1894:131), he states, “the Silver Salmon is not com-mon south of the Columbia, but is some-times taken in California.” Again Jordan (1894:131) states “only the King Salmon has been noticed south of San Francisco.” in Jordan (1904:79; 1907:296), he states that “Silver salmon…predominate in most of the smaller streams along the coast. All of the species occur from the Columbia northward; but the blue-back is not found in the Sacramento.” Jordan (1904:297) also states that “only the quinnat and the dog salmon have been noticed south of San Francisco.” Quinnat and dog salmon are the original names used by the 1895 Rutter party for their collections, later correctly identi-fied as coho salmon.

From these statements, Kaczynski and Alvarado (2006:376) assert that, “The early coho salmon distribution literature stated that coho salmon were not found south of San Francisco or were only found north of San Francisco.” They also state (p. 381) that “The early scientific sur-veys and soft literature discussed above speak to the absence of coho salmon south of San Francisco prior to 1895 and in the early 1900s before the introduc-tory plants in 1906.” Jordan’s statements on coho salmon only say that they are “abundant” north of San Francisco. Jordan’s statements that only quinnat and dog salmon have been “noticed” in 1904 refer to the misidentified 1895 coho salmon collections discussed above.

it is important to note that in all of the literature preceding the Kaczynski and Alvarado (2006) publication, no one has ever suggested that coho salmon are not native south of San Francisco.

ArCHAeoLoGICAL evIdenCe

Using the archaeological record to help reconstruct past faunal assemblages in California has been accomplished with mixed success. on the one hand, Gobalet (1990, 1993) confirmed the pres-ence of thicktail chub (Gila crassicauda), Sacramento pikeminnow (Ptychocheilus grandis), and Sacramento perch (Archoplites interruptus) in the Pajaro and Salinas rivers through remains from archaeological sites on Elkhorn Slough in Monterey County. independently, Schulz (1995) reinforced Gobalet’s (1990, 1993) findings with the identifi-cation of thicktail chub and Sacramento perch at an inland archaeological site on the Pajaro River. on the other hand, these excavations at Elkhorn Slough have failed to document other native fish species expected from the drain-age, such as hardhead (Mylopharodon conocephalus) and splittail (Pogonichthys macrolepidotus). Thus, despite extensive anecdotal, historical, and ethnographic evidence that the indians of the Central valley of California were harvesting large quantities of Chinook salmon (Yoshiyama 1999), the archaeological record does not show this (Gobalet et al. 2004). Likewise, if archaeological evi-dence were used to establish the historic distributions of marine fishes, otherwise

well-known, documented, common, and easily harvested fishes such as Pacific tomcod (Microgadus proximus) and wolf eel (Anarrhichthys ocellatus) would be shown to be absent from most of their range.

Given that the absence of remains in archeological surveys is not reliable evi-dence, by itself, to establish the limit of a species’ historical distribution, Kaczynski and Alvarado’s (2006) citing of Gobalet’s work on the archaeological record of California (Gobalet 1990; Gobalet and Jones 1995; Gobalet et al. 2004) as nega-tive evidence for coho salmon in streams south of San Francisco is not appropri-ate. For the entire coast of California, there is only Follett’s (1966) report of coho salmon from a site in Del norte County (by the oregon border) and the Gobalet et al. (2004) report of either Chinook salmon or coho salmon from a site in Mendocino County on the north-ern California coast. To be accurate, Kaczynski and Alvarado (2006) should have reported that the archaeological record for coastal California showed no coho salmon south of Mendocino County and not just south of San Francisco.

in January 2006, new archeologi-cal evidence surfaced that showed coho salmon presence south of San Francisco. Diane Gifford-Gonzales of the University of California, Santa Cruz provided one of us (K. Gobalet) with fish remains recovered during excavations of Santa Cruz County archaeological site SMA-18 located in Año nuevo State Park sev-eral kilometers north of Waddell Creek, the nearest major stream. Remains

LIterAture revIew: ACCurACy And PreCIsIon oF AGes estImAtes And BACk-CALCuLAtIon oF LenGtH

Accuracy and precision of age estimates

444 Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org

were from a broad spectrum of coastal California species and including such small fishes as herrings (Clupeidae), sardines (Sardinops sagax caeruleus), northern anchovies (Engraulis mordax), smelt (osmeridae), northern clingfish (Gobiesox maeandricus), and Pacific sil-versides (Atherinidae). Also, the centra of two salmonids were recovered and independently evaluated by three fish faunal identification experts. All three were not informed of the determinations of the others. one vertebra was deter-mined to be from a coho salmon by all three experts and the second was identi-fied as coho salmon by two of the three. Thus, with these findings alone, the archaeological record for coho salmon south of San Francisco is now as exten-sive as the archaeological record north of San Francisco. Additionally, the ohlone tribelets living around Santa Cruz had distinct names for salmon and trout (Shipley 2002; M. Hylkema, Cal. Dept. of Parks and Recreation, per. comm.), so there are ethnographic suggestions of more than one salmonid in the local streams.

suItABILIty oF HABItAt

For an anadromous species such as coho salmon, the edge of its range varies naturally over time. in the Pleistocene, it is conceivable that the southern mar-gin of coho salmon distribution was in southern California during wet cool periods and in oregon or Washington during dry periods. obviously, the distribution of coho salmon would track the distribution of suitable habitat, in this case cool streams flowing through coastal rain-forests dominated by redwood (Sequoia sem-pervirens) and Douglas fir (Pseudotsuga menzie-sii). in recent times, coastal redwood forests reached their southern limits in Santa Cruz County, California, although there are smaller relic forests in the northern Big Sur area further south. The ending of these continuous forests marks Santa Cruz County as the southern boundary of the Environmental Protection Agency’s (EPA’s) Marine West Coast Forest Ecoregion (EPA 2006). This ecoregion

begins at the U.S. and Canadian bor-der and ends south of Santa Cruz at the northern edge of the Salinas valley. The southern limit of this ecoregion is an ecologically sound southern boundary for coho salmon.

The stream habitat in the south of San Francisco region was characterized by Kaczynski and Alvarado (2006) as marginal, harsh, and extreme for coho salmon due to extreme fluctuations in flow, severe drought conditions, high summer water temperatures, and exces-sive amount of fine sediment in the streams. Each of these claims is addressed below.

Extreme fluctuations in flow. Kaczynski and Alvarado (2006:383) show using graphs that the region south of San Francisco region is more likely to receive more than four inches of rain in a single day than the region immedi-ately north of San Francisco Bay (Marin County). While this is true for all months except october, the difference in rainfall pattern between Santa Cruz County and Marin County is small. in six months of the year, Santa Cruz County is less than one-quarter of one percent (0.22%) more likely to receive more than four inches of rainfall than Marin County on average; while in october, Marin County is only very slightly (0.1%) more likely to receive more than four inches than Santa Cruz County. in other words,

approxi- mately every three years, Santa Cruz County receives one additional day of rain more than four inches than Marin County. Although extreme flow events can devas-tate year classes of coho salmon in coastal California streams (Smith 1998a), it is very unlikely that such small differences

in extreme flow events were biologically significant under historical conditions.

Severe drought conditions. Periods of prolonged drought are also claimed to cause streams south of San Francisco Bay to be unsuitable for coho salmon (Kaczynski and Alvarado 2006:381-382), where only two streams retained a single coho salmon year class after the 1975-77 drought. in fact, coho salmon survived that extreme drought, and all year classes were apparently still present until 1991, when the first large storm providing adult access did not occur until March 8th, far later than the usual coho salmon spawning period (Smith 1994). The conditions in 1991 resulted in the weak-ening of the year class on Scott Creek. The same year class was also weakened in Redwood Creek in Marin County, but for both areas, water diversions during the drought were a major difficulty.

Warmer summer stream tempera-tures. Assertions by Kaczynski and Alvarado (2006:382) that the south of San Francisco region water temperatures were warmer than those of more north-ern waters are based on Weitkamp et al. (1995) who indicate that average annual sunshine along the Central California coast is greater than anywhere further north. However since the geographic region that Weitkamp et al. (1995) refers to extends from the San Lorenzo River in Santa Cruz County to Cape Mendocino

in Humboldt County, the relevancy of this supporting evidence

is unclear. Figure 7 from Weitkamp et al. (1995)

specifically shows that the maximum stream temperature range in the south of San Francisco region is

the same as that in several streams north of

San Francisco (in Marin and Mendocino counties),

while some inland coho salmon streams further north have higher maxi-mum temperatures. inland portions of rivers, like the Eel, that still support coho salmon, have warmer temperatures than streams south of San Francisco (Weitkamp et al. 1995). in addition, air temperature and the resulting water tem-perature are more likely to be affected by distance from the cooler, foggy coast

The most parsimonious explanation is that coho salmon were found in those

four streams at the same time because coho salmon populations persisted there and the

widespread freshwater distribution of coho salmon argues that coho salmon had persistent, self-sustaining populations

south of San Francisco.

o.s. systems, Inc.

Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org 445

and the size of the stream (and resultant shading) than by north-south distances. Coho salmon were first collected in the region in 1860 by Alexander Agassiz in San Mateo Creek, San Mateo County, which drains into South San Francisco Bay (Leidy et al. 2005). Given that coho salmon occurred in San Mateo Creek, and likely several other streams on the warmer inland side of the coastal ridge, the claim that the coastal streams were too warm for coho salmon is not justified.

Excessive amounts of fine sediment. Santa Cruz County streams do have an abundance of fine sediment and it is a problem for the fish (Smith 1996). However, the problem is neither new nor limited to streams south of the San Francisco Bay. All streams in the south-ern coho salmon region have received massive sediment impacts dating back to the 1850s from logging, sawmills, agri-culture, and development (Payne 1978). Thus, it is not surprising that coho salmon populations have declined or even disappeared from southern streams, as they have from many northern streams for the same reasons (Brown et al. 1994). in fact, the region south of San Francisco Bay already had been seriously degraded by 1895 (Payne 1978) when the Rutter and Scofield collections were made. it is reasonable to argue that if native coho salmon populations were not present in the region, it is because they were extir-pated by human abuse within the water-sheds, as Kaczynski and Alvarado (2006) indicate as an alternative hypothesis. Despite these detrimental land uses, natural reproduction of coho salmon in Scott and Waddell creeks has been tak-ing place at least since Shapovalov and Taft (1954) observed it in the 1930s. The present degraded stream substrate condi-tions reduce the resilience of southern coho salmon populations, because low spawning success can keep coho salmon populations that have been depressed from rapidly rebounding (Smith 1992, 1996, 1998b).

in short, there is no evidence that substantial habitat differences exist between coho salmon streams in Marin County, north of the San Francisco, and Santa Cruz County south of San Francisco. To the contrary, habitat con-ditions in Santa Cruz are fairly typi-cal of streams supporting coho salmon

along the coast of California, as noted by Shapovalov and Taft (1954:10) who described Waddell Creek as “…typical of the great majority of California streams of like size. Moreover, in miniature it is almost a replica of the larger stream sys-tems, such as the Klamath and the Eel.” Santa Cruz County also marks the south-ern boundary of the EPA’s Marine West Coast Forest Ecoregion (EPA 2006), and provides an ecologically logical southern boundary for coho salmon. Finally, Leidy et al. (2005) document collections of coho salmon in San Mateo Creek, San Mateo County, from 1860, pre-dating any known hatchery influence. if coho salmon could inhabit some southern San Francisco Bay streams on the landward side of the low-elevation coastal hills, it seems hard to make a credible argument that coho salmon are unable to survive in streams a short distance away on the cooler seaward side of those hills from these habitats. The undisputed declines of coho salmon populations in the south of San Francisco region are primarily the result of anthropogenic, not natural, factors.

tHe roLe oF HAtCHery FIsH

Kaczynski and Alvarado (2006) propose that hatch-ery plantings made between 1906 and 1910 were respon-sible for subse-quent observations of coho salmon in streams south of San Francisco. The fish numbers Kaczynski and Alvarado (2006:378) use to support their claim as “coho salmon …planted in streams south of San Francisco Bay” are in fact 500,000 fer-tilized coho salmon eggs brought into the Santa Cruz area from the state of Washington during the period from 1906–1910 and planted as fry. The practice of fry

plants has been discontinued in current times because of lack of success (iSAB 1998), but even so the fry plants would be roughly the production of 125 females. Coho salmon were commonly acknowl-edged to occur throughout Monterey Bay area streams by 1910 (Gilbert 1912; Snyder 1912; Scofield 1916) so for Kaczynski and Alvarado (2006) to be correct, the 250,000 coho salmon eggs planted as fry before 1908 would have had to been remarkably successful to account for their general acceptance, particularly in contrast to their current endangered state.

Recent genetic studies of coho salmon south of San Francisco (Bjorkstedt et al. 2005; Bucklin et al. 2007) show the cur-rent populations in Scott and Waddell creeks closest relatives are fish from Marin County, directly to the north of the region in question. While these data (Bjorkstedt et al. 2005) do not directly prove that coho salmon are native to this area, they do indicate that coho salmon presently in the south of San Francisco region experience extensive gene flow with populations in other parts of the California Coastal Cost (CCC) Coho ESU and are not the result of the

446 Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org

Washington state plantings. This genetic evidence also suggests that the Scott and Waddell creek populations are not descended from fish introduced from the noyo River, as suggested by Kaczynski and Alvarado (2006), because there is significant genetic distance between these populations in the bootstrap con-sensus tree (Figure 2.2, Bjorkstedt et al. 2005; Figure 3, Bucklin et al. 2007). The genetic evidence shows particu-larly strong relationships among south of San Francisco populations (Bucklin et al. 2007) and Bjorkstedt et al. (2005) suggests that this may be the result of interbasin transfers of Kingfisher Flat Hatchery coho salmon within the south of San Francisco region. in addi-tion, the statement in Kaczynski and Alvarado (2006:379) that “the latest genetic data for the stocks south of San Francisco do not support concordance between genetic and geographic popula-tion structure” is based upon examina-tion of unpublished data provided to the authors. Two formal analyses of genetic data (Bjorkstedt et al. 2005; Bucklin et al. 2007) strongly contradict their state-ment and indicate a similar concordance between genetic and geographic dis-tance in both the newer dataset and that described by Bjorkstedt et al. (2005).

Finally, Kaczynski and Alvarado (2006) incorrectly state that Smith (2005) concluded that Kingfisher Flat hatchery operations are responsible for the continued presence of coho salmon south of San Francisco and that as of 2003, only a single year class persisted south of San Francisco (Smith 2004). Smith (2005) emphasized that a single year class had remained strong during his 1988-2005 studies, but that an acceler-ated growth program at the Kingfisher Flat restoration hatchery in the Scott Creek watershed probably was neces-sary for the restoration of coho salmon populations south of San Francisco Bay. Precocious returns from some acceler-ated growth fish reared at the restoration hatchery did strengthen or restore year classes on Scott and Waddell creeks in 1995 and 1997 and again in 2004 and 2006 (Smith 1995, 1998b, 2005, 2006).

strAyInG

Alternatively, Kaczynski and Alvarado (2006) argue that coho salmon popu-lations south of San Francisco are the

result of straying. As noted above, coho salmon were found by Rutter and Scofield in 1895 in four geographically sequen-tial streams south of San Francisco. in 1909 (only 14 years after the Rutter and Scofield collections), coho salmon were found in the San Lorenzo River (Snyder 1912) and in 1916 were considered com-monly occurring there (Scofield 1916). The presence of coho salmon in these neighboring streams and the later pres-ence in the San Lorenzo River argue strongly against the concept that these coho salmon are simply strays from more northern populations, especially since all four streams contain coho salmon today. The possibility seems remote that coho salmon strayed simultaneously and regu-larly into four geographically sequential streams from more northern streams in large enough numbers to be collected. The most parsimonious explanation is that coho salmon were found in those four streams at the same time because coho salmon populations persisted there and the widespread freshwater distribu-tion of coho salmon argues that coho salmon had persistent, self-sustaining populations south of San Francisco.

information from the commercial ocean salmon fishery also helps in assess-ing whether natural coho salmon in the streams south of San Francisco Bay were strays that barely persisted in a mar-ginal environment or constituted viable populations. The Monterey Bay fishery harvested substantial numbers of coho salmon even before the first coho salmon egg imports to the Brookdale Hatchery. Wilcox (1907:53) reported:

Salmon in any considerable amount have been taken in Monterey Bay only since 1900, during which period the catch has increased. In 1904 the fishery began on May 27 and lasted until August 6....The catch in 1904 comprised 132,790 pounds of silver and 531,110 pounds of chinook salmon....Silver salmon weigh from 4 to 10 pounds each, the average being 6 pounds.

Based on the average weight given in the quote, the coho salmon catch in 1904 would have amounted to about 22,130 fish. Published notes of the California Department of Fish and Game further indicate that substantial

numbers of coho salmon frequented the inshore waters south of San Francisco, particularly Monterey Bay (viz., Scofield 1919:198, 1920:175). Although the ori-gins of those ocean-caught coho salmon are unknown, their significant pres-ence in Monterey Bay in numbers much higher than occur today makes it less likely that only a few occasional strays entered the spawning streams south of San Francisco.

oCeAn CondItIons

Kaczynski and Alvarado’s (2006) assertions that the combination of periodic, decadal-scale linked warm unproductive California Current con-ditions and warm and dry inland cli-mate create such stressful conditions in the area south of San Francisco that coho salmon persistence there would be extremely improbable are also question-able. Reasonably good information on the population dynamics of the fresh-water phase of coho salmon in oregon, Washington, and British Columbia is available including numbers of spawn-ers and smolts produced in 14 streams (Bradford et al. 2000; Barrowman et al. 2003). Beverton-Holt models have been fit to these data to obtain the critical parameter regarding population persis-tence, α, the slope of the spawner/smolt relationship at the origin (i.e., at low abundance; Barrowman et al. 2003). Estimated values of α range from near 20 to 100 smolts/spawner.

information on survival through the ocean stage is also available, based primarily on coded wire tag data from oregon and Washington, the oregon Production index (oPi) survival esti-mates (e.g., Botsford et al. 2005). There has been an increasing appreciation of the prevalence of long-term (i.e., decadal scale) variability in these data. A recent example is the decadal-scale decline in ocean survival of coho salmon from values near 10% in the early 1970s to values less than 1% in the 1990s (e.g., Botsford et al. 2005). More recent data show a population increase to 4% in 2001, which may have been associated with a decadal-scale shift to more favor-able ocean conditions (Peterson and Schwing 2003). As a consequence of the long-term decline in survival, catch declined in the mid-1970s synchronously along the coasts of Washington, oregon,

floy Tag and Mfg., Inc.

Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org 447

and California (Botsford and Lawrence 2002). identification of the prevalence of variability on long time scales pres-ents the problem that sampling over longer times is necessary to establish the distribution of possible survivals.

To explore the implications of these freshwater and marine data, Botsford et al. (2005) simulated coho salmon popu-lations based on oPi survival rates and values of the slope of the spawner/smolt relationship at the high end and the low end of the estimated distribution of val-ues (i.e., α = 20 smolts/spawner and α = 100 smolts/spawner, from Barrowman et al. 2003). Simulated abundance declined for both values of α, with populations for which α = 20 smolts/spawner dropping to extremely low values. The conclu-sion that can be drawn from this analy-sis is that populations of coho salmon with low values of α along the coast of oregon, Washington, and British Columbia were indeed in great jeopardy during the 1980s and 1990s.

Kaczynski and Alvarado (2006) took a different approach to persistence. They calculated static replacement rates as an indicator of likely persistence of popula-

tions at various latitudes, asking whether (fecundity) x (freshwater survival) x (marine survival) is less than one. in doing so, the authors offer a simplistic constant survival population equation to support their case. Their results depend directly on their choice of different values of freshwater survival at differ-ent latitudes. They used a value of 3% for the oregon Production index (oPi) area, based on five studies in oregon and Washington (oDFW 1982). Their cal-culated replacement rates dropped from 8.4, a value adequate for replacement, in 1970 to a marginal value of 1.01 in 1980, then further to values inadequate for replacement in 1996. These follow the well-known trend in coho salmon survival. Kaczynski and Alvarado (2006) then chose a freshwater survival of 1%, and slightly lower female fecundities, for coho salmon south of San Francisco, to demonstrate that an ocean survival of 8.6% would be required for population persistence, a ocean survival value not seen anywhere since the mid-1970s. Kaczynski and Alvarado (2006:384) fur-ther emphasized that if they had chosen 0.5% as a freshwater survival rate, the

required value of ocean survival would be 17.1%, "a value never seen." While these statements appear to cast doubt that there were persistent populations south of San Francisco, they are not likely to be true for two reasons: (1) the survival estimates from oregon are biased low, and (2), we do not know what the freshwater survival was in the pristine streams south of San Francisco. The problem with the simple replace-ment approach is that unless the surviv-als were calculated in years in which the populations were on the increasing por-tion of the Beverton-Holt relationship, they will be biased low because they are constrained by the nonlinear nature of the relationship to be less than α, the true replacement capacity at low abun-dance. it is unlikely that these popula-tions were on the lower abundance part of the curve because populations were not reported to be at low abundanace. in fact, the basis for the survivals given by Kaczynski and Alvarado (2006:384) was not given; rather they merely appeared in a table (Table ii.B-1) with references to two other documents. The one docu-ment readily available (Moring and Lantz

448 Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org

1975) was a study of 14 years of sampling adults and the smolts they produced, in 3 streams. The authors remark,

Despite the fact that numbers of spawners varied greatly during the years, numbers of outgoing smolts remained relatively steady from year to year,...

which supports the presence of strong density-dependence and the fact that the "survivals" calculated did not rep-resent replacement capacity, but rather the effects of density-dependent limita-tions on smolt abundance.

Available information on survival of coho salmon south of San Francisco is limited to a study of four cohorts in a single spawning stream, Waddell Creek in the mid-1930s (Shapovalov and Taft 1954). Freshwater survivals ranged from 1.16 to 1.56. We can assume that values under pristine habitat conditions would have been greater. Four data points would not be sufficient to estimate α with rea-sonable precision or to establish distri-bution of ocean survival at this location. Combining the highest value of survival with fecundity of 1,168 female eggs per female indicates the value of α was greater than 18. Marine survivals ranged from 1% to 8%.

The simulations in Botsford et al. (2005) address another issue raised by Kaczynski and Alvarado (2006), the effect on persistence of differences in spawning age structure of the popu-lations due to precocious spawning. Kaczynski and Alvarado (2006) draw attention to the obligate semelparity (i.e., all spawning at age 3) of south-ern coho salmon, implying it will have a negative effect on their persistence. Two aspects of their argument are con-tradicted by existing information. The first is that the mechanism leading to precocious spawning involves better conditions for growth, not worse con-ditions as implied by Kaczynski and Alvarado (2006). The second is that, as shown by both these simulations and other analytical methods (Hill et al. 2002), while obligate semelparous pop-ulations do have a greater probability of extinction, this slight difference is so sensitive to small amounts of precocious spawning that in practical terms, it is inconsequential. A very low amount of precocious spawning will remove any

differences in the likelihood of extinc-tion (Hill et al. 2002; Botsford et al. 2005).

Kaczynski and Alvarado (2006) also state that ocean conditions and the rainfall driving stream conditions covary in such a way that when the California current is warm and unpro-ductive (El niño conditions), rainfall is less on land and the streams are dry. They refer to Kaczynski (1998) who drew this conclusion by noting sev-eral occurrences of drought conditions from the late 1970s to the 1990s. other evaluations of this covariability from the primary literature do not support this conclusion. Warm, unproductive conditions in the California Current are associated with low coho salmon survival (e.g., Botsford and Lawrence 2002). of the 11 El niño-Southern oscillation events between 1950 and 1982, 6 were unusually wet years, 4 were normal years and 1 was a drought (Schonher and nicholson 1989). There is a correlation between lower rainfall and El niño conditions at higher lati-tudes (Redmond and Koch 1991), but it declines from north to south, being much weaker in California than in Washington state.

in summary, we have a reasonably comprehensive picture of the effects of the physical environment on coho salmon along the Pacific coast from oregon to British Columbia over the past several decades. Unfortunately, we do not have the information nec-essary to perform the same analysis on populations south of San Francisco. Simple replacement rate calculations by Kaczynski and Alvarado (2006) pres-ent a biased view of persistence because they do not represent the true spawner/smolt relationship. Comprehensive evaluation of persistence through simulations shows that some popula-tions of coho salmon along the coast between oregon and British Columbia could have declined to precariously low numbers, but they also indicate that an apparent lack of precocious spawning is not a reliable indicator of lower persis-tence. Lastly, the proposed reinforcing effect of drought conditions co-occuring with warm, unproductive ocean (i.e., El niño conditions) in central California is not supported by relevant data.

FALse CItAtIons

Kaczynski and Alvarado (2006:386) claim that “false citations and citations of erroneous information [are]…used to substantiate the hypothesis that coho salmon are native south of San Francisco.” in fact, their “false cita-tions” are either careless errors in non-peer reviewed papers or simply assumptions of the obvious. We believe that the authors of most papers on these fish simply did not feel a need to justify coho salmon being native to the region. Even Shapovalov and Taft (1954), the authors of what is still the most com-plete coho salmon study in California, never questioned the idea that the fish were native there, stating, “The only introduced species in Waddell Creek is the Striped Bass.”

ConCLusIons

Existing evidence strongly supports coho salmon as being native to streams south of San Francisco, contrary to the arguments of Kaczynski and Alvarado (2006). The key evidence we have pre-sented is summarized below:

1. Coho salmon were collected from four streams in the south of San Francisco region in 1895 by Rutter and Scofield. These collections were made long after a great deal of envi-ronmental degradation had already occurred. There is little reason to doubt that these collections, now in the CAS, are authentic coho salmon records, even though they were origi-nally misidentified. The likelihoods of coho salmon randomly straying and spawning in these four streams are very small.

2. Early literature records only identify coho salmon as abundant from San Francisco northward, not as being absent south of San Francisco. The misidentified coho salmon from the 1895 Rutter and Scofield collections are included in the early literature records.

3. Coho salmon vertebrae have been found in a native American midden at Año nuevo in 2006 between Gazos and Waddell creeks. The archaeo-logical evidence that Kaczynski and Alvarado (2006) cited has changed

Halltech

Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org 449

and now there is positive archaeological evidence for the

presence of coho salmon in streams south of San Francisco.

4. While habitat at the southern edge of any species’ range

is likely to be of less-than-optimal quality, there is no evi-

dence that a distinct habitat or faunal change occurs at

San Francisco Bay, while south of the San Lorenzo River

has been identified as the southern boundary of the EPA’s

Marine West Coast Forest Ecoregion (EPA 2006).

5. The probability that 500,000 salmon eggs brought to hatch-

eries between 1906 and 1910 could have established these

coho salmon populations in the early twentieth century is

extremely low, based on the success of hatchery plants else-

where, and the hatchery origin hypothesis is inconsistent

with the current available genetic evidence.

in short, the case made by Kaczynski and Alvarado (2006)

that coho salmon are not native south of San Francisco Bay is

based on inadequate analysis of the existing information and

an excessive willingness to accept information that favors

their hypothesis, while discounting information that does

not. Although in trying to establish historical range limits

there will always be some uncertainty, a careful and thorough

review of the historical and current information demonstrates

that coho salmon are native south of San Francisco Bay.

ACknowLedGments

We wish to thank David Catania for his help on the California Academy of Science collections, P. Fiedler for edi-torial assistance, and Jessica White, Catherine Johnston, and Allison Collins for literature research.

reFereenCes

Barrowman, n. J., r. A. myers, r. Hilborn, d. G. kehler, and C. A. Field. 2003. The variability among populations of coho salmon in the maximum reproductive rate and depensation. Ecological Applications 13(3):784-793.

Bjorkstedt, e. P., B. C. spence, J. C. Garza, d. G. Hankin, d. Fuller, w. e. Jones, J. J. smith, and r. macedo. 2005. An analysis of historical population structure for Evolutionarily Significant Units of Chinook salmon, coho salmon, and steel-head in the north-Central California Coast Recovery Domain. noAA Technical Memorandum nMFS-SWFSC-382.

Bœhlke, J. 1953. A catalogue of the type specimens of recent fishes in the natural History Museum of Stanford University. Stanford ichthyological Bulletin 5:1-167

Botsford, L. w., and C. A. Lawrence.2002. Patterns of co-variability among California Current, chinook salmon, coho salmon, Dungeness crab, and physical oceanographic condi-tions. Progress in oceanography 53(2-4):283-305.

Botsford, L. w., C. A. Lawrence, and m. F. Hill. 2005. Differences in dynamic response of California Current salmon species to

450 Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org

changes in ocean conditions. Deep Sea Research (ii Topical Studies in oceanography) 52(1-2):331-345.

Bradford, m. J., r. A. myers, and J. r. Irvine. 2000. Reference points for coho salmon (Oncorhynchus kisutch) harvest rates and escapement goals based on freshwater production. Canadian Journal of Fisheries and Aquatic Science 57(4):677-686

Brown, L. r., P. B. moyle, and r. m. yoshiyama. 1994. Historical decline and current status of coho salmon in California. north American Journal of Fisheries Management 14(2):237–261.

Bucklin, k. A., m. A. Banks, and d. Hedgecock. 2007. Assessing genetic diversity of protected salmon (Oncorhynchus kisutch) populations in California. Canadian Journal of Fisheries and Aquatic Science 64:30-42.

CFGC (California Fish and Game Commission). 1913. Twenty-second Biennial Report for the years 1910-1912. CFGC, Sacramento, CA.

ePA (environmental Protection Agency). 2006. Level i Ecoregions. EPA Western Ecology Division, Corvallis, oregon. Available at www.epa.gov/wed/pages/ecoregions/na_ecoregions.htm.

Follett, w. I. 1966. Fish species. Pages 85–86 in R. A. Gould, ed. Archaeology of the Point St. George site and Tolowa prehis-tory. University of California, Publications in Anthropology, volume 4, Berkeley.

Gilbert, G. H. 1912. Age at maturity of the Pacific Coast salmon of the genus Oncorhynchus. Bulletin of the United States Bureau of Fisheries 32:1-32.

Gobalet, k. w. 1990. Prehistoric status of freshwater fishes of the Pajaro-Salinas River system of California. Copeia 1990:690–685.

_____. 1993. Additional archaeological evidence for endemic fishes of California’s Central valley in the coastal Pajaro-Salinas basin. Southwestern naturalist 38(3):218–223.

Gobalet, k. w. and t. L. Jones. 1995. Prehistoric native American fisheries of the central California coast. Transactions of the American Fisheries Society 124:813–823.

Gobalet, k. w, P. d. schulz, t. A. wake, and n. siefkin. 2004. Archaeological perspectives on native American fish-eries of California, with emphasis on steelhead and salmon. Transactions of the American Fisheries Society 133:801-833.

Hill, m. F., A. Hastings and L. w. Botsford. 2002. The effects of small dispersal rates on extinction times in structured meta-population models. American naturalist 160:389-402.

IsAB (Independent scientific Advisory Board, northwest Power and Conservation Council). 1998. Review of artificial produc-tion of anadromous and resident fish in the Columbia River Basin. Document 98-33. Portland, oR. Available at: http://www.nwcouncil.org/library/1998/98-33/98-33.htm#contents.

Jordan, d. s. 1892a. Salmon and trout of the Pacific coast. Biennial Report of the State Board of Fish Commissioners of the State of California. 1891-1892 (12):44-58.

_____. 1892b. Salmon and trout of the Pacific coast. California Fish and Game Commission Bulletin 4. State office A.J. Johnston Superintendent State Printing, Sacramento.

_____. 1894. Salmon and trout of the Pacific coast. Biennial Report of the State Board of Fish Commissioners of the State of California. 1893-1894 (13):125-141.

_____. 1904. Pacific species of salmon and trout. Biennial Report of the State Board of Fish Commissioners of the State of California. 1903-1904 (18):75-97.

_____. 1907. Fishes. H. Holt and Company, new York._____. 1922. The days of a man. volume one, 1851-1899. World

Book Company, new York.Jordan, d. s., and B. w. evermann. 1896. The fishes of north

and Middle America: a descriptive catalogue of the species of fishlike vertebrates found in the waters of north America, north of the isthmus of Panama. United States national Museum, Washington, DC.

_____. 1902. American food and game fishes: a popular account of all the species found in America north of the equator, with keys for ready identification, life histories and methods of capture. Doubleday Page and Co., new York.

_____. 1904. American food and game fishes: a popular account of all the species found in America north of the equator, with keys for ready identification, life histories and methods of capture. Doubleday Page and Co., new York.

_____. 1905. American food and game fishes: a popular account of all the species found in America north of the equator, with keys for ready identification, life histories and methods of capture. Doubleday Page and Co., new York.

Jordan, d. s., C. H. Gilbert, and C. L. Hubbs. 1882. A synop-sis of the fishes of northAmerica. Smithsonian institution, Washington, DC.

kaczynski, v. w. 1998. Marine survival of opia Hatchery coho salmon related to marine temperatures. Pages 131-147 in Proceedings of the 49th Annual Pacific northwest Fish Culture Conference. idaho Department of Fish Game, Boise.

kaczynski, v. w., and F. Alvarado. 2006. Assessment of the southern range limit of north American coho salmon: dif-ficulties in establishing natural range boundaries. Fisheries 31(8): 374-391.

Leidy, r. A., G. Becker, and B. n. Harvey. 2005. Historical sta-tus of coho salmon in streams of the urbanized San Francisco Estuary, California. California Fish and Game 91: 219-254.

moring, J. r., and r. L. Lantz. 1975. The Alsea watershed study: effects of logging n the aquatic resources of three headwater streams of the Alsea River, oregon. Part i-Biological Studies. oregon Department of Fish and Wildlife, Fishery Research Report 9. Corvallis, oR.

moyle, P. B. 2002. inland fishes of California. University of California Press, Berkeley.

odFw (oregon department of Fish and wildlife). 1982. Comprehensive plan for production and management of oregon’s anadromous salmon and trout. Part ii. Coho salmon considerations. oDFW Fish Division, Portland.

Payne, s. m. 1978. A howling wilderness: a history of the Summit Road area of the Santa Cruz Mountains 1850-1906. Loma Prieta Publishing Co., Santa Cruz, CA.

Peterson, w. P., and F. schwing. 2003. A new climate regime in northeast Pacific ecosystems. Geophysical Research Letters 30 (17):1896.

redmond, k. t., and r. w. koch. 1991. Surface climate and streamflow variability in the western United States and their relationship to large scale circulation indices. Water Resources Research 27:2381-2399.

sandercock, F. k. 1991. Life history of coho salmon (Oncorhynchus kisutch). Pages 395–446 in C. Groot and L. Margolis, eds.,

Miller Net company, Inc.

sonotronics, Inc.

Fisheries • vol 32 no 9 • september 2007 • www.fisheries.org 451

Pacific salmon life histories. University of British Columbia Press, vancouver.

schonher, t., and s. e. nicholson. 1989. The relationship between California rainfall and EnSo events. Journal of Climate 2:1258-1269.

schulz, P. d. 1995. Prehistoric fish remains, thicktail chub, from the Pajaro River system. California Fish and Game 81(2): 82–84.

scofield, n. B. 1916. The humpback and dog salmon taken in San Lorenzo River.California Fish and Game Quarterly 2(1): 41.

_____. 1919. Commercial fishery notes. Salmon at Monterey Bay. California Fish and Game 5(4):198.

_____. 1920. Commercial fishery notes. Silver salmon at Monterey in 1920. California Fish and Game 6(4):175.

shapovalov, L., and A. C. taft. 1954. The life histories of the steelhead rainbow trout (Salmo gairdneri gairdneri) and sil-ver salmon (Oncorhynchus kisutch) with special reference to Waddell Creek, California, and recommendations regarding their managment. California Department of Fish and Game Fish Bulletin 98:1–375.

shipley, w. 2002. The Awáswas language. Pages 173-181 in L. Yamane, ed. A gathering of voices. The native peoples of the central California coast. Santa Cruz County History Journal, issue 5.

smith, J. J. 1992. Distribution and abundance of juvenile coho and steelhead in Waddell, Scott and Gazos creeks in 1992. Department of Biological Sciences, San Jose State University, San Jose, CA.

_____. 1994. Distribution and abundance of juvenile coho and steelhead in Scott and Waddell creeks in 1988 and 1994: implications for status of southern coho. Department of Biological Sciences, San Jose State University, San Jose, CA.

_____. 1995. Distribution and abundance of coho and steelhead in Gazos, Waddell and Scott creeks in 1995. Department of Biological Sciences, San Jose State University, San Jose, CA.

_____. 1996. Distribution and abundance of coho and steelhead in Gazos, Waddell and Scott creeks in 1996. Department of Biological Sciences, San Jose State University, San Jose, CA.

_____. 1998a. Distribution and abundance of juvenile coho and steelhead in Gazos, Waddell and Scott creeks in 1998. Department of Biological Sciences, San Jose State University, San Jose, CA.

_____. 1998b. Distribution and abundance of juvenile coho and steelhead in Gazos, Waddell and Scott creeks in 1997 and the implications for status of southern coho. Department of Biological Sciences, San Jose State University, San Jose, CA.

_____. 2004. Distribution and abundance of juvenile coho and steelhead in Gazos, Waddell and Scott creeks in 2003. Department of Biological Sciences, San Jose State University, San Jose, CA.

_____. 2005. Distribution and abundance of juvenile coho

and steelhead in Gazos, Waddell and Scott creeks in 2004.

Department of Biological Sciences, San Jose State University,

San Jose, CA.

_____. 2006. Distribution and abundance of juvenile coho

and steelhead in Gazos, Waddell and Scott creeks in 2006.

Department of Biological Sciences, San Jose State University,

San Jose, CA.

snyder, J. o. 1912. The fishes of the streams eliminate of tribu-

tary to Monterey Bay, California. Bulletin of the U.S. Bureau

of Fisheries 32:47-72

_____. 1931. Salmon of the Klamath River, California. California

Division of Fish and Game Bulletin 34.

weitkamp, L. A., t. C.wainwright, G. J. Bryant, G. B.

milner, d. J. teel, r. G. kope, and r. s.waples. 1995.

Status review of coho salmon from Washington, oregon, and

California. U.S. Department of Commerce, noAA Technical

Memorandum nMFS-nWFSC-24.

wilcox, w. A. 1907. The commercial fisheries of the Pacific

Coast stocks in 1904. U.S. Bureau of Fisheries, Report of

the Commissioner of Fisheries for the Fiscal Year 1905 and

Special Papers 612. Washington, D.C.

yoshiyama, r. m. 1999. A history of salmon and people in the

Central valley region of California. Reviews in Fisheries

Science 7:197–239.