A reassessment of the phytogeographic characterization of Temperate Pacific South America

In 1980, a study of geographic distribution patterns and geographic affinities ofbenthic algae of temperate Pacific South America (5-55 °S) provided a first characterization of these coasts. High endemism and limited floristic exchange with the Tropical Pacific and with islands in the South Pacific at various distances from the continent were two outstanding features of the marine flora. A partial blockage of species exchange partially accounted for a peculiar latitudinal pattern of species richness that increases with increasing latitudes. In contrast, there was reduced species richness in comparison with other climatically equivalent regions that have contact with more effective routes of migration. Over the last two decades, more than 30 taxonomic and biogeographic studies on this benthic marine flora have been completed, adding a significant number of new records for the area. New studies test the above characterization and indicate that the new data add support to several key aspects of the phytogeographic characterization of this coastline. The relative importance of the various floristic components at different latitudes, the isolated character of this flora and the latitudinal pattern of increasing species richness to higher latitudes are shown.


INTRODUCTION
Twenty years ago, Santelices ( 1980) characterized the benthic phytogeography of the western coast of temperate South America from Northern Peru to Cape Horn in Southern Chile (4-55 °S). This characterization contained two components: a) the study of the floristic affinities of the species registered for the area in order to identify the relative contribution of species with different biogeographic affinities to the marine flora of the region and b) a quantitative study of latitudinal floristic changes along this coastline to define major biogeographic boundaries in the area.
Over the last two decades, a number of floristic studies have added a significant number of new records for the area. This study and a related paper in this volume (Meneses & Santelices 2000) evaluate the validity and significance of the above phytogeographic characterization in light of these new findings. In this report, we foccus on geographic patterns and floristic affinities while Meneses & Santelices (2000) evaluate the latitudinal floristic changes in light of the new evidence.

Phytogeographic characterization of the temperate coast of Pacific South America
The study of the biogeographic affinities of the benthic algal flora of temperate Pacific South America showed (Santelices 1980) a species assemblage of 380 taxa which comprise five different groups (Fig. 1). About 33.2% of the regional flora (some 123 species) were endemic to this coast and only a few of them were also found in geographically close oceanic archipelagos, such as Juan Fermindez (with six of these 380 species) or Galapagos (with 13 species). About 34.4% of the temperate Pacific South American benthic flora (130 species) had definite subantarctic affinities, as many of these species had also been reported from the subantarctic islands, New Zealand, Tasmania, Southern Australia and even from the Antarctic Peninsula. Only 3.4% of the flora (13 species) was found to correspond to Central American Pacific species, most of which also had limited latitudinal range extension into South America. A fourth geographic group, comprising 22.8% of the flora (87 species) exhibited wide geographic distribution, being present also in different parts of the world, such as the Atlantic coast of Europe and North America, the Caribbean, Pacific North America, Australia, New Zealand and the subantarctic islands. The fifth group was represented by 27 species (7 .1%) endemic to the Pacific coast of America, with anphitropical pattern of distribution in which the species are distributed along the Pacific coast of North and South" America, but are absent from Central America.
The relative importance of these five floristic elements varied at different latitudes of the Pacific coast of South America (4-55 °S). The number of endemic species (Fig. 1) for different latitudes was similar across the entire range, with three exceptions. The first of these was a short stretch of coastline north of Paita, Peru (4-5 os) where the number of endemic species was much reduced. The southern most extreme of South America (53-55 os) and also Central Peru ( 11-12 os) in the region of Callao, exhibited more endemic species, most of which presented short latitudinal distributions and many were known from only a single recorded collection.
The numerical importance of the species with subantarctic affinities along the South American coast ( Fig. 1) was found to decrease steadily to the north and none were found to extend north of 5 os latitude nor in the Galapagos. On the other hand, and as indicated above, the contribution of Central American Pacific elements was restricted both in number of species and in latitudinal range (Fig. I). Only six of these species reached as far as 5 os, and only two of them extended beyond Callao (c. 12 os). None were found to extend south of Coquimbo, in Central Chile (30 os). The latitudinal distributions of the remaining two groups (Fig. 1) were contrastingly different, but numerically complementary. The relative importance of cosmopolitan or widely distributed species increased while the importance of bipolar species decreased to the south.
Since the summation of endemic, bipolar and widely distributed species did not significantly change along temperate Pacific South America, the total species richness at different latitudes along this coastline was heavily influenced by a northward reduction in subantarctic species and a restricted southward incursion of Central American tropical species. The resulting pattern ( Fig. 1) was an increase in the total number of species with increasing latitudes. A major phytogeographic generalization that emergl(d from this study (Santelices 1980) was the apparent isolation of the marine benthic flora of temperate Pacific South America. Floristic exchange across the northern rim of the Peru Current appeared small as the contribution of Central American Pacific &pecies to temperate South America was limited in both number of species and latitudinal range. Floristic exchange between South America and nearby offshore islands (e.g., Juan Fernandez or Galapagos) also seemed limited, as indicated by the small number of common elements. This partial blockade to species migration and colonization was thought to explain the high representation of endemic elements and the peculiar latitudinal species diversity gradient found along Pacific South America.
Expanding these findings to species richness, Santelices (1980; also suggested that the relative isolation of this coastline could partially account for the reduced species richness of this area in comparison to other climatically equivalent regions which generally have greater contact with more effective routes of migration.

Testing the hypothesis
A critical analysis of the data supporting the above hypothesis suggested a rather spotty knowledge of the marine algal flora of the region (Ramfrez & Santelices 1981, Santelices 1982. Thus, the geographic patterns observed could reflect incomplete sampling rather than true species absence. Therefore, new taxonomic studies in the region would contribute relevant information to test the above characterization. More complete floristic studies in the northern part of the region (e.g., Northern Chile, Peru) and in the Oceanic Islands (Easter Island and Juan Fernandez Archipelago) were thought to be specially critical because they would test the respective absence of species from the Central American Pacific and from the Western Pacific and Central Pacific Islands. If the described pattern was due to incomplete sampling, those taxonomic studies should increase the number of species with tropical affinities in Northern Chile and Peru and the number of species shared between the Continent and the Central Pacific Islands.

Floristic studies in Northern Chile and Peru
Intensive taxonomic studies were carried out (Ramfrez & Santelices 1981, Ramfrez 1982 between 1976 and 1980 in intertidal and subtidal marine habitats around Antofagasta in Northern Chile (23 °40'S; 70 °25'W). Eight ecologically different sampling stations were regularly sampled over the 4 years, with 60% of the total number of gatherings being collected in subtidal habitats.
A total of 70 species was found, 14 of which constituted new records for the area. The intensive sampling, however, failed to significantly modify the total number of species known for the area or to find species with geographic affinities not considered in the original biogeographic hypothesis (Santelices 1980). Sampling also failed to significantly increase the number of species with tropical affinities. Therefore, the relative representations of algal groups with different geographic affinities remained as predicted by the phytogeographic hypothesis, while the latitudinal pattern of species richness was only slightly modified.
Intensive seaweed sampling was later completed (Pinto 1989) further north, between Iquique (20° 12'S) and Loa River (21 °26'S). A total of 24 sampling sites, including intertidal and subtidal stations (to 10 m depth) were regularly sampled between June 1987 and May 1988, which corresponded to a inter-El Nifio period.
Sampling in this case yielded 89 species, 19 of which represented range extensions. Similar to the Antofagasta study (Ramfrez & Santelices 1981) and in spite of the new additions, the geographic affinities of the flora found near Iquique agreed well with the proposed phytogeographic characterization of temperate Pacific South America.
No special intensive sampling programs to increase taxonomic know ledge have been undertaken in any locality along the Peruvian coastline that would provide data to test the role of sampling effort on the above phytogeographic characterization. However, Acleto ( 1988) completed a preliminary analysis of the phytogeographic affinities of the Peruvian flora. Integrating results of the most important taxonomic contributions to the Peruvian littoral (Howe 1914;Collins 1915;Taylor 1947;Dawson et al. 1964;Acleto 1973Acleto , 1980Acosta 1976;Acleto & Endo 1977), Acleto (1988) recorded 225 species extending between 3°30'S and18 °30'S. A total of 28 species (12.4%) was recognized as endemic to Peru, 15 species (6.6%) had affinities with the marine flora of Chile and the subantarctic region; 24 species (1 0.6%) had tropical and subtropical affinities with the American Pacific, 26 species (11.5%) exhibited bipolar distribution and the remaining 132 species (58.6%) had widespread geographic distribution. Acleto (1988) concluded that the distribution of the marine flora of Peru agreed well with the previous (Santelices 1980) phytogeographic characterization done of the area.
Intertidal sampling in specific places of the Peruvian littoral generally yields low numbers of species. For example, Acosta (1976) listed 67 species for the Pisco Province in Central Peru. More recently, Benavente ( 1994) found only 36 species ofbenthic macroalgae in intertidal habitats ofCaleta Yasila, south ofPaita (5°S). Interestingly most of these species were endemic to the area (32.3% ). Species with wide distributions or bipolar distributions had roughly similar importance (23.5% and 20.6% respectively) while the species with subantarctic affinities (5.9%) were the least abundant. Although species with tropical affinities were proportionally more abundant than in localities further south (17 .65% ), their quantitative significance still was very minor.
In summary, results of floristic studies conducted in Northern Chile and Peru over the last 20 years are consistent with several aspects of the original phytogeographic predictions made by Santelices (1980) for the area.

Floristic studies with Islands in the Central Pacific
A new taxonomic survey of the intertidal and shallow subtidal marine algal flora of Easter Island (27 °07'S; 109 °22'W) was undertaken (Santelices & Abbott 1987) to test whether increase in sampling effort would increase the total number of species known for the island and later, how this increase in species richness would affect the level of floristic similarity between the island and the South American Continent.
The Las afinidades floristicas de la flora marina de Isla de Pascua. A) Patron de distribucion a lo largo del Pacifico de cien especies colectadas en la Is la de Pascua. B) Especies de Is la de Pascua (30 spp) con patron de distribucion del Pacifico Indooccidental. C) Patron de distribucion de especies colectadas en Isla de Pascua y en otras aguas tropicales y subtropicales. D) Patron de distribucion de especies colectadas en Isla de Pascua y en otras aguas calidas y temperadas. Notese que elmimero de especies en A) es el resultado de la suma de Ios datos en B), C) y D) mas 14 especies endemicas y 6 otras especies con distribucion en otros lugares (ver Figs. 3B y 3D). species was reached, 69.4% of which (100) were considered for further geographic analysis (Fig.  2). Thirty three species exhibited wide distribution in tropical and subtropical waters (Fig. 2C), 17 were widely distributed in warm and temperate waters (Fig. 2D), 30 species had an Indo-West Pacific affinity (Fig. 2B); 14 were endemic to Easter Island and six species corresponded to taxa found only in a few other localities (e.g., Juan Fernandez). When these 100 species are mapped according to their representation in other localities of the Pacific ( Fig. 2A), it becomes evident that the number of species the island has in common with another locality decreases as distance increases. However, even the most remote Western Pacific locality considered (Nha Trang in Vietnam) maintained a higher similarity value with Easter Island than nearby Juan Fermindez. Thus, this new survey (Santelices & Ab bott 1987) characterized the flora of Easter Island as having a greater affinity with islands in the West Pacific and almost no relationship with islands in the East Pacific or with Continental Chile, regardless of distance and despite the significant increase in new records and species numbers achieved through the study.
A later survey on the benthic marine algae from Easter Island was undertaken by Ramirez & M tiller ( 1991 ), in which the authors made collections on three opportunities at different localities around the island. They added six new records and confirmed the affinities and biogeographical patterns established by Santelices & Abbott ( 1987), likewise concluding that the flora of this island is very different from that of the nearest Pacific South America coasts.
These marine flora findings of Easter Island motivated a new assessment of the geographic affinities of the marine flora of the Juan Fernandez Archipelago (Santelices 1992), a group of islands located at about 650 km to the west of Central Chile (33°45'S; 79°22'W). Data were initially recorded for 110 taxa. However, taxonomic uncertainties and lack of information on geographic distribution of species elsewhere reduced the number of species for further consideration to 89 (Fig. 3A). This flora exhibited a high degree of endemism (about 30% ), a large number of widely distributed species (45%; Fig.  3D) and a small group of species with circumpolarsubantarctic affinities (about 13.5%; Fig. 3B). The sources for this flora, therefore, seem to be distant localities in the Southern Pacific, including the southern tip of South America, Southern Australia, New Zealand and several subantarctic islands. The West Wind Drift and the Humboldt Current appeared as the route most probably followed by many species reaching the islands. Thus, the geographic affinities of this flora are as high with Southern Chile (19-21 of the 89 species) as with Central and Northern Chile ( 19-23 species) and higher than with Peru (12-19 common species). Considering effective dispersal distances of marine algae, the Juan Fernandez Archipelago appears to be more isolated than Easter Island, which explains the very different number of endemic species found in both islands (14% in Easter Island; 29.2% in Juan Fernandez). These results also are consistent with predictions of limited species exchange across the northward flow of the Chile-Peru current system. The presence of 5 species that only eo-occur in the Archipelago and in Central Chile (Fig. 3C) attests to the possibility of migration across the Chile-Peru current system. But, the low number of these species suggests that although the possibility exists, its frequency is low. Rock lobster fishing boats moving between the Archipelago and the continent or ENSO events of such an intensity as to reach the Archipelago (e.g., 1973; Silva & S ievers 197 3; Arana 19 8 7) are potential transporting mechanisms between the Juan Fernandez Archipelago and Central Chile.
More recently, Meneses & Hoffmann ( 1994) studied drift and hand collected materials from the Desventuradas Archipelago (26°20' S, 80°'W). They added 10 new records to the 12 previously known species (Levring 1942;Etcheverry 1960). Although these islands are located 972 km from the Chilean coast, only five of the 22 species found there also occur in the continent. On the other hand, six of the species found in the islands also occur in temperate zones of the South Pacific (South Africa, Australia, Tasmania, New Zealand) reproducing the pattern found for Juan Fernandez. Meneses & Hoffmann ( 1994) stressed the floristic similarity between the Desventuradas and the Juan Fernandez Archipelago and the potential transport and introduction of marine algae by lobster fishing boats that roam among these islands.
In summary, results with the new floristic studies of the oceanic islands off Central Chile support the prediction of a very limited floristic exchange between the islands and the continent. These studies also have characterized phytogeographically the marine flora of these islands, described the large floristic differences between them and called attention to effective dispersal ranges, which seem to be much more extensive for Juan Fernandez than for Easter Island, regardless of their respective distance from the Continent.

Latitudinal patterns of species richness
In addition to the taxonomic studies mentioned above, others (Etcheverry et al. 1980;Acleto 1981Acleto , 1984Acleto , 1986Westermeier 1981;Santelices & Montalva 1983;Contreras et al. 1984;Ramfrez & Rojas 1986Ramirez et al. 1986;Westermeier & Rivera 1986;Villouta & Santelices 1986;Santelices & Abbott 1985;Bird et al. 1987;Wynne 1988;Santelices et al. 1989) have added new records and described new species for the coast of Peru and Chile. It is therefore pertinent to evaluate whether or not these additions have significantly modified the pattern of species richness along temperate Pacific South America.
The emerging pattern of latitudinal species richness varies slightly when single findings are included or excluded (Fig. 4). If only repeated findings are considered (Fig. 4B) there is a gra-

A. Total number of species
.. dual increase of species richness from 4°S to 42°S, species richness decreases between 42°S and 44°S, and it remains stable down between 42°S and the southernmost tip of South America. Considering the number of species involved, these data indicate that it would be expected to find 50-70 species in representative habitats between 5°S and 10°S, around 90-95 species between 12°S and 20°S; 120-130 species between 20°S and 32°S and 130-140 species at latitudes between 3 3 and 40°S, decreasing thereafter to 120-130 species south of 42°S (Chiloe).
The inclusion of single findings (equivalent to one degree findings; Fig. 4C) slightly modifies the above pattern. Along the Peruvian and Chilean coastline there is a larger number of species in places where more intensive taxonomic studies have been completed, suggesting that the species richness of the whole area is perhaps higher and that many of these single gatherings could become more extended ranges with more intensive sampling along the entire coast. Still, the increments are slight in most places and follow the general latitudinal pattern already discussed. Thus, as many as 100 species, rather than 50-70, may be expected between 5 and l0°S; tip to 120 species may occur between 12 and 20°S; and up to 200 species between 33 and 40°S. The high values gathered for 54 and 55°S are especially important.  ) have been especially useful to understanding the species pools that form in wave exposed and sheltered rocky habitats so typical of Central and Southern Chile. Here, they are complemented by a diversity of taxa representative of the sheltered sounds that become ecologically important south of 42°S. Regression analyses of the latitudinal trends in species richness (Fig. 4) in all cases indicate significant increases in species richness to the south. Compared to the distributional data used 20 years ago (Fig. 4A), the taxonomic data added over the last 20 years (Fig. 4B) has slightly reduced the r 2 value but have not modified the statistical significance of the pattern (p <0.0001). Incorporation of single findings (one degreespecies) further increases the statistical significance of the southward increase in species number (Fig. 4C) .
In summary, the additional data contributed by taxonomic studies in Peru and Chile, have supported the hypothesis that the number of benthic algal species increases to the south in temperate Pacific South America, in sharp contrast to the widely accepted paradigm of increased species richness to the tropics. Compared to the data available 20 years ago, new species additions and phytogeographic registries additions have allowed for statistical analyses supporting the pattern. Further sampling between 41 os and 53°S now appears to be most important to complete the overall picture of species richness along temperate Pacific South America.

Relative importance of geographically different groups in the marine flora of temperate South America
Given the addition of new records and species to this marine flora, it is necessary to evaluate whether or not the relative contribution of geographically different groups of species to this flora has changed in relation to the pattern predicted by Santelices (1980).
Compared to the original characterization (compare Fig. 5 with Fig. 1), endemic and bipolar species now show a reduction in species abundance towards the northermost and southermost distribution limits. Species with wide geographic distribution remains equally represented along temperate Pacific South America. The very unequal representation of tropical species versus the abundance of species with subantarctic affinities is notable and still appears as the principal factor determining the latitudinal pattern of I 30spp.
~ . .  from these studies have added support to several key aspects of the phytogeographic characterization formulated 20 years ago for this coastline. This coastline appears to be effectively isolated from the Western Pacific, the Central Pacific Islands and the Eastern Tropical Pacific and presents very unequal contribution of tropical and subantarctic elements. These regional characteristics are likely to have noticeable effects on the biodiversity and ecology of these algae, some of which have been explored elsewhere (Santelices 1980(Santelices , 1982.