A r c h . B i o l . M e d . E x p e r . 12:179-190, 1979 L i z a r d s and rodents: an explanation for their relative species diversity in C h i l e Lagartijas y roedores: una explicación para su diversidad relativa en Chile E D U A R D O R. F U E N T E S , F A B I A N M . J A K S I C Instituto de Ciencias Biológicas, Universidad Católica de Chile. Casilla 114-D, Santiago de Chile (Recibido para su publicación el 3 de diciembre de 1977) F U E N T E S , E . R . , J A K S I Ó , F . M . L i z a r d s and rodents: an explanation for their relative species diversity in C h i l e . ( L a g a r t i j a s y roedores: una explicación para su diversidad relativa en C h i l e ) . A r c h . B i o l . M e d . Exper. It 72.179-190, 1979. is discussed that the relative richness o f l i z a r d fauna in C h i l e might be related attributes of this g r o u p ( m a r k e d altitudinal segregation, coarse-grained vagility associated w i t h e c t o t h e r m y ) , which w o u l d favor a rapid to ecological habitat selection, and l o w speciation pattern concomitantly w i t h Q u a t e r n a r y Glaciations. L i z a r d s w o u l d d i v e r g e on mountaintops during interglacial periods and w o u l d m e r g e into fewer forms o r further d i v e r g e on the l o w l a n d s during glacial periods' intermingling. It is shown that number of subspecies per species in mountaintops is greater than on l o w l a n d s , which is consistent with the above hypothesis hereafter referred to as " m o u n t a i n speciation m o d e " . H o w e v e r , it is also documented that lizards might exhibit a " v a l l e y speciation m o d e " , in which case they w o u l d d i v e r g e in non-icy valleys during glacial periods, intermingling w i t h lizards in adjacent valleys during interglacials. A s in the previous case, merging or speciating of divergent forms w o u l d depend on the degree of reproductive isolation attained. A s a contrast, it is discussed the situation o f the relatively poor Chilean rodent fauna. It is shown that although rodents exhibit some altitudinal segregation they do not mountaintops than on l o w l a n d s , which is possibly related to their and greater vagility associated w i t h endothermy. O n the other form m o r e subspecies on fine-grained habitat selection hand, number of subspecies per species is positively correlated w i t h the extension of its geographical range, which strongly suggests a " v a l l e y speciation m o d e " for rodents in C h i l e . SPECIATION THE BIOGREOGRAPHY QUATERNARY GLACIATIONS PROBLEM depauperated VERTEBRATES when compared CHILE with the fauna of an area of similar climate and resources, like C h i l e has a geographically isolated position in C a l i f o r n i a for e x a m p l e . South A m e r i c a . T h e high A n d e s on the east, the barren A t a c a m a desert on the north, and M a m m a l s , ants and bees are about half as the diverse in C h i l e than in C a l i f o r n i a ( 2 ) . Birds Pacific O c e a n on the west, are all strong barriers and frogs are about the same, but lizard species to the dispersal are about one and a half times m o r e numerous of most o r g a n i s m s ( 1 ) . C o n c o ­ mitantly, the Chilean fauna tends to be relatively in C h i l e than in California ( 2 ) . In fact, speciation 179 180 F U E N T E S and J A K S I C in Chilean Liolaemus species has been strong tend to exhibit strong altitudinal partitioning. enough to produce a mosaic of alternative c o m ­ Of munities in central C h i l e ( 3 ) . O n the other hand, 20 are predominantly found above 1000 meters endemism in m a m m a l s , ants, and bees also tends above sea level ( m a s l ) , and only 3 of them to be high ( 1 , 2, 4 ) . known The question w e will attempt is species known in C h i l e , to live above and this However, at varies. but complex archipelago with copious speciation and 27° latitude south by the A t a c a m a for there are 8 Liolaemus m o r e like an island for other latitudes Between are altitude. segregation groups. different below w h y C h i l e seems to be comparable to an isolated lizards, and to answer the 36 Liolaemus altitudinal aproximately 18° desert, species above 1000 masl and only 1 below it. Further south, as the valley Possible explanations for differences in the gets m o r e mesic, the proportion of species found number of species between C h i l e and C a l i f o r n i a below could be related to the amount of time the dif­ Between 27° and 37° latitude, the ratio of species ferent groups have had to speciate in C h i l e , o r above 1000 masl to species below is about 1.6; to between 37° and 42° latitude, the ratio is 1.0, different resource availabilities in the two 1000 masl increases monotonically. regions. H o w e v e r , recent studies ( 2 , 5, 6 ) have and further shown the l o w productivity in the northern valley, the resource organisms is California. O n evidence does supply quite aid several groups o f in Chile and the other hand, paleontological is weak or not for comparable to inexistent, explain conceivable though that the both and therefore patterns. of these It could be shown to be important. tive route which w e hope will at least offer explanation for the relative a species diversity phenomenon. W e will attempt to show that ecological differences between lizards other groups, manifest low average height o f the Andes south of 42° latitude (see 7 ) . T h u s , there is strong altitudinal partitioning only in the central part of the country, roughly around Santiago. F i g . 1 shows altitudinal species In this contribution w e have taken an alterna­ partial reflects m o r e mesic climate in the central part, and the is factors south it is 0.0. T h i s trend themselves in and different speciation patterns. Because available ecological sequences on the mesic Roble, L a Campana) slopes ( L a g u n i l l a s ) . Coastal Ranges (El and on the drier A n d e s T h e figure starts only at about 1000 masl and does not contain the species b e l o w the level (see 8 ) . Still, it can be seen that there is species turnover on the three mountains, although it is not strictly parallel in all o f them. It information about most taxa is scarce, w e w i l l is important to emphasize at this point restrict ourselves mostly to the comparison o f that in other transects covering a w i d e r altitudinal lizards with rodents. range, some species would probably be added, T h e r e is no doubt that these are " p a r t i c u l a r " groups, but as will be seen and later in this paper, their speciation modes could slope and exposure. T h e important Fig. offer clues for future w o r k on other taxa. First manner, w e w i l l document, in a fairly detailed the ecological attributes of lizards in the will then c o m p a r e these attributes with the available ecological information on rodents and their postulated speciation mode. 1 are the existence of species turnover lack of parallelism in The the distribution biological leopardinus lizards, country. The and of reasons for the altitudinal ed, but there are already some e m e r g i n g patterns. L. high A n d e s border its east side, and on the west on aspects in S o m e species in F i g . 1, like L. nigroviridis PARTITIONING C h i l e is basically a mountainous depending species replacement are currently being investigat­ L. ALTITUDINAL be dropped, species on nearby mountains. relation to their proposed pattern of speciation. We perhaps other are whereas schroederi other and L. species further typically species like L. nitidus and high-mountain are low south or north ( 8 , and tenuis, altitude personal observations). Species w o u l d thus be " t r a c k i n g " coast it is limited by the Coastal R a n g e s , with in altitude only a n a r r o w central valley in between. T h i s can survive and reproduce. H o w e v e r , what physiography is strongly reflected in the distri­ the restrictions which limit the use o f the w h o l e bution of the organisms. T h u s , Chilean altitudinal gradient by the species? lizards and latitude conditions w h e r e they are LIZARDS A N D Fuentes that and density selective and Jaksic pressure on open (in preparation) of grasses can ground show impose a such that lizards, analogue species strong grass can be 181 RODENTS lemniscatus ( F i g . 1). L. lemniscatus At El R o b l e , as grass density altitude, L. lemniscatus decreases munity, in each habitat are different, and thus presence nigrouiridis or is absent from the Coastal R a n g e s . absence of grass acts as a sieve alternative communities (9, separating 10) in each environment. We found with drops out of the c o m ­ distinguished. O p t i m a l morphologies for survival true is a grass- species which uses ground preferentially ( 5 , 11). being replaced by the j u v e n i l e s o f (see 5 ) . In this case L. L. leopardinus O t h e r species seem to be strongly associated with the presence or absence of a v e r y particular this mechanism to explain some environmental feature. Thus, L. monticola is latitudinal as well as altitudinal habitat partition­ a ing. presence of boulders and rocks ( 8 , 11, 12). T h e s e T h u s , L. nigrouiridis dweller, ground but can is preferentially a rock frequently if grass density be found Both species are leopardinus as grass density increases. preferentially rock dwellers, but have the option of hunting on the and L. leopardinus L. the is l o w . A t L a g u n i l l a s (east of Santiago) w e found that L. replaces L. nigrouiridis on ground, The dweller features strongly associated tend to be more frequent with the high in the mountains, close to tops and ridges, and concomi­ tantly L. monticola at l o w altitudes. drops out of the community Similarly, L. tenuis is a tree dweller ( 5 , 8 ) and its presence is strongly as­ sociated with the distribution of trees. is more of a grass-species than nigrouiridis. rock Besides these particular elements, some general metabolic difficulties seem to add more same grass-non grass phenomenon could be involved in setting an altitudinal EL 1000 limit on L. ROBLE 1200 1400 1600 T 1 constraints on high altitude hance zonation. CAMPANA 1800 2000 1 1 1000 1200 1400 1600 species and to en­ Shortening of the growing LAGUNILLAS I8D0 20CO I0CO 1 1400 1830 2200 1 1 1 1 1 1 1 L_ nigroviridis L leopardinus L_ schroederi L_ tenuis L. monticola L nitidus L. lemniscatus Lgravenhorsti L. fuscus C maculatus Fig. 1. Altitudinal species replacement of lizard species. Altitudinal species sequences at the closeby mountains El Roble, Campana, and Lagunillas are compared. Data for El Roble were obtained from Pinto el al. (43). Data for Campana from Jerez and Ortiz (44). Notice that species sequences differ on the three mountains. See the text for further explanation and discussion. 182 F U E N T E S and J A K S I C season with altitude imposes strong on some species. Fuentes and restrictions D i Castri ( 1 3 ) noted that high altitude Liolaemus species are 1. Species which live on mountains and do not form subspecies. H o w e v e r , this is strictly not an exception but sustaining evidence, since forced to be o m n i v o r o u s , incorporating vegetable these species are known from only one locality tissues and therefore give strong evidence for genetic into their diet. They interpreted this plant consumption as compensation for l o w and isolation unpredictable category are L. signifer, ments. insect densities in those e n v i r o n ­ Furthermore, large Callopistes its ( 5 ) found that eats vegetation at quardi, L. not paulinae, (see 8 ) . of altitudinal favorable season distribution, but shortens with altitude, lizards mountaintops. L. L. 2. M o u n t a i n Species L. buergeri, constanzae, species at in L. ornatus, pantherinus, lorenzmulleri, d o w n . A g a i n , it w a s suggested that as limit further Fuentes maculatus on L. L. L. knegi, and high L. darwxni latitudes do form pensate the l o w consumption of insects by eating mountains are very l o w , and distinction plant material. T h e p r o b l e m is that only large ween mountain lizards can all use vegetation efficiently, plants are physiological edible (13, difficulties 14). for (see 7 ) . A g a i n , this is expected from the h y ­ the pothesis lizards of fitzingeri 17). T h i s seasons at indicates higher that altitudes be viviparous shorter could favorable impose alti­ tudinal restrictions to egg development of some Liolaemus species. O n the other side, l o w alti­ tudinal limits are likely to be associated with particular habitat features ( i . e . rocks and trees), and diffuse competition ( 9 ) in these m o r e stable environments. In and does not contradict it. Species in this category are L. magellanicus, altitude (8, bet­ and valley becomes arbitrary not L. to not latitudes and most also tend H o w e v e r , at high Besides eating l o w caloric plant material ( 1 5 , 1 6 ) , high Liolaemus moc- fitzgeraldi, have to be less insectivorous and have to c o m ­ subspecies. this L. lineomaculatus, L. L. d'orbignyi, kingi, and L. (see 8 ) . T h u s , the pattern of divergence on mountains but homogeneity at l o w altitudes, confirms the suggestion isolation based on of lizard altitudes. ecological data populations A s w e saw regarding living before, habitat at high selection (most likely in relation with interspecific c o m ­ petition) can set barriers to movement of lizards on an altitudinal gradient. It is likely then that sum, in central C h i l e there is altitudinal species turnover of lizards and this pattern seems to be associated with alternative these barriers have a l l o w e d for intraspecific dif­ ferentiation in mountain species (see Fig. 2 ) . morphologies and behavior of the species. In general, species Interestingly, the of nal parable t o , or greater than differences observed habitats, are not suggesting equally adapted the presence to other of ecological barriers to lizard movement between mountains. populations differentiation between gradient mountain degree with optimal fitness at one point in the altitudi­ has been c o m ­ between mainland and island populations. T h e r e are no island species absent from the mainland, and the existing island forms are o n l y subspecies of mainland species ( 8 ) . H e r e w e refer only to islands connected with M O V E M E N T S BETWEEN MOUNTAINS the mainland during the last glaciation, that is H e l l m i c h ( 1 2 , 18) noted that high altitude species separated tend to diverge more than species inhabiting the meters deep channel valley. In fact, using the data by H e l l m i c h ( 1 2 ) excludes mainly the Juan Fernández archipelago and and some large islands on the southern channels, D o n o s o - B a r r o s ( 8 ) , w e have been able to from the continent by at (19, 20). T h i s most 100 restriction show that species living under 800 masl tend to both of which are not known to h a v e lizards. It have fewer also excludes Easter subspecies than species above that level (Fischer's exact probability test, P < . 0 0 1 ) . Typical Furthermore, braltstroemi, species 8 0 0 masl (L. nitidus, living above and L. fuscus, and L. below lemnisca- major island Island, since it is oceanic. subspecies L. pictus are L. chiloensis, cyanogaster L. pictus on the southern C h i l o é archipelago (43° tus) do not form subspecies. T h e r e are t w o types latitude south), L. nigromaculatus of exceptions to this rule: nigromaculatus ater, and sieversi at about 30° latitude. L. 183 LIZARDS A N D RODENTS Fig. 2. Altitudinal islands in central Chile. Distribution of heights above 1000 masl (black) or above 600 masl (dashed) are comparable to an altitudinal mainland on the east, with an archipelago on the west. Glacial snows could have been as low as 900 masl (see 21) and therefore the density of altitudinal islands is a measure of alternative recolonization opportunities in central Chile during interglacial times. See the text for discussion and more explanations In sum, adaptations of lizards to their habitats seems to be fine enough to entail reproductive isolation between populations on nearby moun­ tains. would have formation isolation in the European Alps, Hell- mich ( 1 2 ) suggested the Andean races o f mus would develop on different Liolae- mountains during interglacial periods and eventually come into contact with each on glaciations, ranges Based on research other during glacial resource partitioning By the above mechanism, races and species A T W O W A Y SPECIATION MODEL favored and hence n e w forms capable o f coexistence. would separate populations follow perhaps interglacial mountains. formerly w o u l d intermingle and During on different further diverge, or if differences achieved are not strong enough, they w o u l d m e r g e into one species. A t any rate, this is a speciation model which in repeated glaciation events could generate great diversity. In central C h i l e there have occurred three to periods. T h a t is, as snow level approaches l o w e r four altitudes, ranges along the A n d e s , as well as in the Coastal high altitude lizard species and sub­ glaciations (21) and there are separate closer to R a n g e s ( 7 ) , thus a l l o w i n g for multiple species the valley, with many of them c o m i n g in contact formation within even a small latitudinal range. with each other. If reproductive difference achiev­ The essential factors in this speciation mode is ed during the isolation in the interglacial period that species were and that during interglacial periods, movements species w o u l d shift their distributions pronounced enough, natural selection partition the habitat altitudinally 184 F U E N T E S and JAKSIC between mountains are restricted and comparable On to by valley speciation movements between on the continental of this mainland and islands shelf. In the remaining part contribution we shall refer to this speciation mode as " m o u n t a i n speciation". A second mode of speciation, the other number side, the same design of species per cycle. T h e point is that speciation rate depends isolates and amount not exclusive operating would at most double the directly on number o f of intermingling. C l e a r l y , speciation rate of groups using both modes C h i l e . D u r i n g glaciations, pockets of non snowed speciation rate of groups using only one mode. been inferred to occur in the southern be considerably higher than For all the reasons given above, lizards seem habitats or at least with short g r o w i n g seasons, have would speciation of the former, is also likely to have operated in to fall into the mountain speciation mode. T h e A n d e s ( 2 1 ) . T h e s e pockets being more or less evidence given on their altitudinal isolated from each other, can lead to reproductive and on the isolation inter- presence of subspecies and even species on some glacial periods, the formerly isolated populations mountains during the present interglacial, point between populations. w o u l d intermingle and further During differentiate, or m e r g e into fewer forms. F o r analogous reasons stratification isolation between mountains, in this direction. H o w e v e r , in addition with lizards could have also used the valley speciation mode. as before, w e will refer to this speciation mode Unfortunately, since w e are living on interglacial as " v a l l e y speciation". In valley speciation, the period at present and therefore during an essential features mingling phase for valley between valleys are or restricted pockets movements during glacial possible to clearly determine its importance. periods, and m i x i n g during interglacial periods. Fig. 3 shows the t w o speciation modes in a Possible lizards examples could patterns of L. t w o valleys, as shown in F i g . 3, each cycle o f L. platei mountain 26° diagram. number Note that speciation for three mountains would at most triple and the of species (species 1 and 2 in Fig. 3 ) . BEGINNING INTERGLACIAL END OF INTERGLACIAL inter­ speciation, it is not platei and be of the platei valley speciation geographical and L. nigromaculatus. is roughly distributed 32° latitude south, whereas known subspecies L. platei GLACIAL in subspecies curicensis between its only is restrict- INTERGLACIAL MOUNTAIN / 1 SPECIATION, DIFFERENTIATION DIFFERENTIATIONS OF POPULATIONS OF POPULATIONS VALLEY SPECIATION DIFFERENTIATION OF POPULATIONS Fig. 3. Speciation modes. Mountain and valley speciation are asynchronous with regard to the phases of differentiation and of mixing of differentiat­ ed populations. In mountain speciation, differentiation occurs during interglacial periods and mixing in valleys during glaciations. T h e opposite is true of valley speciation, where differentiation occurs in isolated habitat pockets or val leys during gl acial periods. See th e text for further discussion. 185 LIZARDS A N D RODENTS ed to a few valleys of C u r i c o , at 35° C o d y et at. latitude ( 2 ) used beta-diversity, that is (see 2 2 ) . Both D o n o s o - B a r r o s ( 8 ) and M u l l e r species turnover over comparable habitat tran­ and H e l l m i c h ( 2 2 ) insist on the subspecies level sitions, to describe species turnover of m a m m a l s of differentiation, although for the L. platei the argument. this is not It is reasonable to crucial assume and lizards in California and C h i l e . T h e y found that beta-diversity altitudinal is in central C h i l e . T h i s means not only that lizards the continuous 32° and last 35° latitude glaciation, the south. previously distribution of L. platei could have and mammals rodents lizards between gap for transects quite During comparable on once had a wider distribution, covering change with and altitude, but that in central C h i l e — at least around the latitude of been fragmented. L a t e r , during the interglacial, Santiago — they seem to change at about L. platei same rate. w o u l d have persisted only in its known distribution, 35° with extinctions between 32° and latitude. Lwlaemus Here, competition with other species could have been very important, T h e biological compromises involved altitudinal segregation by rodents are the in the not as clear as in lizards. In N o r t h A m e r i c a , structural since they reach high species densities in that habitat ( 2 7 , 28, 2 9 ) and productivity ( 3 0 , 3 1 ) latitudinal have been range preparation) L. platei ( 8 ) . Fuentes offer evidence and Jaksic (in suggesting that — a non grassy species ae is in fact replaced by L. lemniscatus the permanence of L. platei last glaciation is not curicensis clearly since the understood, but could be related to the local physiography. The second evidence for valley in lizards is the kreiss of L. L. nigromaculatus and its numerous W h a t e v e r the reasons are which explain the altitudinal partitioning phenomenon in rodents, o n e requirement of mountain nigromaculatus. subspecies, evidence for isolation of lizards is the race formation between between localities in a we data by O s g o o d manifestation lemniscatus of the secondary discussed L. contact platei (12). — L. transition occurs at about this same tend to be relatively isolated, the same is not true for rodents. form a latitudinal cline ( 2 , 12), suggested to be Interestingly, speciation is not met by these vertebrates. W h e r e a s lizards living on different mountaintops speciation par­ there is no published information. — a grassy species — at about 32° latitude south. O n the other hand, invoked in relation to habitat titioning, but about the Chilean situation used Rottmann to be ( 3 2 ) , and statistically mountaintops the valley. found For ( 1 ) and Our greater than rodents, Miller and this difference not significant (Fischer's exact latitude ( 3 2 ° ) and shows clear signs of h y b r i d i z ­ probability test, P > . 3 4 ) . T h a t is, species restric­ ation and secondary contact (Fuentes and Jaksic, ted to l o w altitudes tend to form as many sub­ in p r e p a r a t i o n ) . ' It is very likely therefore, that species ( o r to be as isolated) as species living there w a s a glacial barrier to dispersal at about on mountaintops. that latitude. N o w a d a y s w e w o u l d be o b s e r v i n g the merging and differentiation of The formerly isolated populations. M o r e evidence on areas o f secondary contact, and k n o w l e d g e o f which valleys w e r e glaciated (see 21) are needed before w e can postulate a set of barriers to dispersal in the above result is particularly important for our speciation model, since the three m a m malogically better known islands off the coast of C h i l e ( T i e r r a del Fuego, C h i l o e and M o c h a ) w e r e all connected to the mainland during the last glaciation ( 2 1 ) , and n o w a d a y s they exhibit past. subspecies of mainland forms. T i e r r a del Fuego Rodents present a clearly contrasting situation in comparison to lizards. Rodents exhibit alti- has three subspecies, C h i l o e four, and Mocha three (see 1, 3 2 ) . tudinal habitat partitioning (see 2 3 , 24, 25, 2 6 ) T h e evidence suggests then, there has but, as w i l l be seen, also m o r e movement between time for rodent subspeciation on mountains, mountains. well as intermingling of populations during the dinal F i g . 4 shows the patterns of altitu- segregation at about 18° latitude south, and M a l l e c o ( 3 9 ° latitude). It can be seen that in both to mountain down. situations tops, there and are species species present further as interglacial, and hence the lack of dif­ ferentiation. restricted living been This difference in mobility between lizards and rodents is likely to be related to their mode of thermoregulation, and energy requirements. 186 F U E N T E S and J A K S I C T A R A F » C A ( I 8 * - 2 0 " L A T SUR) M A L L E C 0 ( 3 8 ' - 3 9 ° L A T SUR ) 1000 2000 3000 4030 5000 T ~r A cirwco 1500 2000 0. glifoidas P oaqoodi A. botiviaww A sublimit E typu» L- viicqcio Ch. tahomqe A. «don Ch. brevicouda N. ebriwus C. lepidus Fig. 4. Altitudinal species replacement of rodents .'pedes. Altitudinal species sequences in central Chile (Malleco) and northern Chile (Tarapacâ). Data for Malleco were obtained from Greer (24). Data for Tarapacâ were taken from Miller and Rottmann (25), Spotorno (26), and records of Museo Nacional de Historia Natural (Chile). See the text for discussion. Rodents are endotherms and do not have the of geographical distribution measured in degrees capability to hybernate, at least in C h i l e ( 1 , 2 4 ) , (r = .82, P < . 0 0 0 1 ) . T h a t is, species so over long latitudinal stretches tend to have m o r e they their are more altitudinal likely range to periodically and diet. In shift contrast, lizards are temporal opportunists, have relatively l o w e r energetic requirements, and restrict activity to the season of m a x i m u m not shifting in altitude. their productivity, use one environmental whereas type preferentially, rodents w o u l d be more finely grained than isolation during In towards speciation m o d e for rodents. compensate than In sum, of their high for do not mountain comply speciation mobility during the valley mode. Number is true w o u l d thus have for the more been diverse for than able to centers of other with the because of subspecies with opposite moun­ interglacials. rodents is significantly correlated the origin of t h e ' C a l i f o r n i a fauna. R o d e n t s , on the On the other hand, rodents seem to have speciated in latter, and hand, speciation rode whereas moun­ tain lizards. requirements In agreement w i t h T h e former w o u l d speciate faster the tains, and hence less subspeciation our this is the fact that C h i l e has m o r e lizards than being less specialized than lizards, w o u l d have between secondary a valley exhibited more between-year i n t e r m i n g l i n g probabilities is valley evidence points rodents. transition This for glacial periods, and conclusion, California, greater ones. conclusive) a t w o - m o d e speciation for lizards, but to the p r o p o r t i o n o f their occurrence. and not contacts during interglacials. tending to use more environmental types, related Rodents, m o r e restricted (certainly In the sense described by L e v i n s ( 3 3 ) , lizards w o u l d be coarsely grained and subspecies evidence distributed of breadth depending conditions, on more w o u l d have stringent speciated at a slower rate and w o u l d therefore not have been able to compensate for the diversity of centers of origin in the Californian rodent fauna. T h e by n o w well k n o w n radiation of altiplano rodents ( 2 3 , 26, 3 4 ) w o u l d thus be linked to the existence of numerous v a l l e y s and surrounding 187 LIZARDS A N D RODENTS mountains, which w o u l d h a v e provided during glacial times the reproductive isolation needed for speciation. T h e geographical and ecological proximity of these valleys during the inter- south and southwest deserts), the relatively low. For speciation (Sonoran Chilean in the and immigration several country Mexican has faunistical has been copious. glacials, their isolation from each other by the A m o n g these, high mountains, and the vaster areas associated an excelent e x a m p l e o f adaptive radiation. T h e i r with the altiplano, would have conditions for greater provided the speciation of rodents in lizards (Liolaemus) been groups, seem to be speciation has been explained as a result of the greater speciation rate as compared to other the northern region than in the more extended groups, particularly rodents. T h e basic difference southern A n d e s . between these t w o types of vertebrates seems to Other groups besides rodents and lizards are more difficult to explain, and the required information w e do not have to examine tlieir reside in their different mobilities, ultimately related to their endothermy or ectothermy. A s a consequence of this difference, lizards w o u l d speciation patterns. H o w e v e r , it is possible that speciate during and between glaciations, whereas both birds and frogs also have valley speciation rodents would do so only during glacial periods. patterns. Birds have high mobilities and are not expected to show the large phenotypic dif- It is possible the speciation of birds and frogs could be explained by the same scheme. ferences between mountains as lizards d o . C o d y ( 3 5 , 3 6 ) has shown Chilean birds tend to segregate more Speciation by regions than in these by habitat animals would operate not only between valleys, but distant and postulated isolated ones. types. between Vuilleumier that the Nothofagus RESUMEN probably bird fauna (37) as- C h i l e es un país relativamente rico en lagartijas, pero pobre en roedores, en comparación lugares que exhiben características con similares de sociated with forest speciated phyletically, which clima y recursos. En este trabajo se intenta res- strongly suggests that long distances are requir- ponder por qué nuestro país ha sido análogo a un archipiélago para las lagartijas (alta diferen- ed for genetic isolation in these animals. Birds would thus be expected to exhibit ciación específica e intraespecífica), pero relatively lower endemism than lizards or rodents. bien como una The diversidad y diferenciación intraespecífica). similarity of C a l i f o r n i a and Chile bird faunas remains to be explained, but it is probably a compound phenomenon and relatively more with local speciation immigration from Se muestra isla para los roedores que las lagartijas más (menor exhiben una fuerte segregación altitudinal, que es particular- distant mente patente en la zona central de C h i l e (véase Frogs might have speciated locally ( 3 8 ) and nal de especies estaría relacionado con la selec- localities. F i g . 1). L a razón para este r e e m p l a z o altitudi- could exhibit valley and perhaps some mountain ción speciation. F r o g s seem to c o m p l y with the l o w parte de las lagartijas, que les proporcionarían mobility requirement, condiciones óptimas para su sobrevivencia y re- but to our is not k n o w n h o w much altitudinal k n o w l e d g e it stratification they exhibit. It is suggestive though that total species counts for C a l i f o r n i a and de ciertas características del habitat producción. En el texto se analizan por numerosos casos en que esta situación es evidente. C h i l e frogs Intimamente ligados a la selección del habitat are about the same (see 39, 4 0 ) . M o r e research se encontrarían on this point is necessary. miento de las las morfologías y el comportalagartijas, por lo que fenotipos óptimos en un cierto lugar no lo serían en o t r o , CONCLUSIONS en donde podrían ser reemplazados por fenotipos (especies) It has been assumed — and there are diferentes. L a estrecha relación que good existiría entre fenotipo y ambiente seleccionado, reasons for it — that C h i l e and C a l i f o r n i a offer determinaría la partición altitudinal del gradien- resources for comparable faunas ( 4 1 , 4 2 ) . H o w - te ever, whereas C a l i f o r n i a has received immigration el flujo de individuos desde ciertas porciones del of species from centers in the north ( G r e a t B a s i n ) , gradiente (en las cuales constituyen el fenotipo de habitats en las montañas, restringiendo 188 F U E N T E S and JAKSIC óptimo) a estratos similares en montañas cer- citan posibles ejemplos de especiación "de v a l l e " . L a suma de estos dos m o d o s de especia- canas. Este r a z o n a m i e n t o basado en datos ecológi- ción, actuando durante sucesivos ciclos glacia- cos, se ajusta al hecho que las lagartijas que están les, restringidas a lagartijas presentes en C h i l e . divergencia (mayor q u e aquellas las montañas muestran cantidad que viven en de mayor El m a y o r la gran diversidad dores chilenos, que son bastante menos diversos aislamiento de las cumbres habría entonces per- q u e las lagartijas. Se discute que, aun mitido los la mayor diferenciación intraespecífica de C o m o contraste se presenta el caso de los roe- subespecies) el valle. explicaría entonces roedores exhiben las un m a y o r m o v i l i d a d entre ellas. A m o d o de a p o y o conjunto de islas en las cuales se puede producir a esta proposición, se muestra que los roedores divergencia de intraespecífica, asemejarían a debido a la restric- ción del flujo de individuos entre aislados conti- montaña no forman 4), también en que se observa en las montañas. En este sentido, se Fig. cuando altitudinal montañas regiones montañosas (véase partición presentan significativamente más subespecies que los de valle. Esto, junto a otras evidencias, sugiere que los roedores especiarían nentales (véase F i g . 2 ) . al modo " d e v a l l e " . A n t e r i o r m e n t e se ha sugerido que las de Lwlaemus se desarrollarían en razas montañas En conclusión, los autores proponen que la diferencia en diversidad de lagartijas y roedores durante los períodos interglaciales, entrando en podría deberse a sus distintos modos de especia- contacto ción: " d e m o n t a ñ a " y " d e v a l l e " para el primer con otras razas similares cuando el descenso de la nieve (durante los períodos glaciales) las empujara a bajar al valle. g r u p o , pero sólo " d e v a l l e " para el segundo En esta situación, o se refuerzan los mecanismos de aislamiento reproductivo, o las especies incipientes se refunden en una sola. En C h i l e han tres o cuatro glaciaciones, de manera que el m e canismo de especiación propuesto podría generar gran diversidad en de las montañas. sucesivas recolonizaciones L o s factores esenciales en este modelo d e especiación requieren d e la partición altitudinal del habitat en las montañas, y la restricción del m o v i m i e n t o de poblaciones que v i - ven distintas en montañas durante ACKNOWLEDGMENTS habido los inter- Johanna Filp and Patricio Sánchez made useful suggestions on the manuscript. W e are also grateful to Angel Spotorno for his advice on altitudinal distribution of Chilean rodents, and to José Yáñez, curator at the Museo Nacional de Historia Natural (Chile), for allowing us to use localities information on mammals under his care. Research reported in this contribution has been generously supported by L'niversidad Católica de Chile through grant DIUC 113/76 to Eduardo R. Fuentes. glaciales. A este modo de especiación, los autores llaman "especiación en montañas". Un podría de segundo modo también ocurrir períodos glaciales través que de divergencia durante e "bolsones" quedarían de REFERENCES la filética alternancia interglaciales, pero a o valles libres de hielo aislados unos de otros, p e r m i - tiendo así la diferenciación de las poblaciones. A este tipo de especiación, los autores "especiación en interglaciales, aisladas podrían valles". las Durante poblaciones reunirse llaman los períodos previamente y divergir aún más, o refundirse en una sola especie. L o s dos tipos de especiación figura propuestos se ejemplifican en la 3. En base a las premisas de estos dos modos de especiación, y a la ecología de las lagartijas, se postula q u e éstas especiarían en C h i l e a la m a nera "de montaña", aunque también se 1. Osgood, W.H., Field Mus. Nat. Hist., Zool. Ser. 30:1, 1943. 2. Cody, M.L., Fuentes, E.R., Glanz, W.E., Hunt, J.H., Moldenke, A., In "Convergent Evolution in Chile and California". Mooney, H . A . , Ed. Stroudsburg, Pennsylvania, Dowden, Hutchinson and Ross, 1977, p. 144. 3. Fuentes, E.R., In "Maquis and Chaparral". Di Castri, F., Ed. Netherlands, Elsevier, (In press). 4. Glanz, W.E., In "Chile-California Mediterranean Scrub Atlas". Thrower, N . J . W . , Bradbury, D . E . , Eds. Stroudsburg, Pennsylvania, Dowden. Hutchinson and Ross, 1977,p. 232. 5. Fuentes, E.R., Ecology 57:3, 1976. 6. Glanz, W.E., Ph. D . Thesis, Berkeley, University of California, 1977. 7. Anonymous, "Atlas de la República de Chile", Santiago, Instituto Geográfico Militar, 1970. 8. Donoso-Barros, R., "Reptiles de Chile", Santiago, Ediciones de la Universidad de Chile, 1966. 189 LIZARDS A N D RODENTS 9. MacArthur, R.H., "Geographical Ecology", N e w York, 10. Cody, M.L., "Competition and the Structure of Bird Communities". Princeton, Princeton University Press, 11. Fuentes, E.R., Ph. D . Thesis, Berkeley, University of California, 1974. 12. Hellmich, W., A b . Bay. Ak. Wiss. 24:1, 13. Fuentes, E.R., paraiso 8:66, 1934. Di Castri, F., An. Mus. Hist. Nat. Val­ R.T., 29. Rosenzweig, 15. Szarski, H, Evolution 76:529, 1962. 16. Ostrom, H.J., Evolution 77:368, 1963. 17. Hurtubia, J., Di Castri, F., In "Mediterranean Type Ecosystems". Di Castri, F., Mooney, H . A . , Eds. N e w York, Springer-Verlag, 1973, p. 349. 18. Hellmich, W.C., Evolution 5:359, 1951. 19. Williams, EE., Quart. Rev. Biol. 44:345, 1969. 20. Rand, A.S., Breviora 319:1, 1969. 21. Simpson, B., Science 173:771, 22. Muller, L., Hellmich, 1971. W., Zool. Anz..9:225, 1938. 23. Pearson, O.P., Univ. California Publ. Zool. 56:391, 1958. 24. Greer, J.K., Publ. M u s . Mich. State Univ., Biol. Ser. 1965. S.D., Rottmann, J., "Guía para el Reconocimien­ Mistral, 1976. A.E., WinakurJ., Ecology 55:558, 1969. Ecology 52:312, 1971. 31. Meserve, PL, J. Mammal. 57:300, 1976. 32. Miller, Rottmann, S.D., J., Informe Interno, CONAF (mimeog.), 1976. 33. Levins, R., "Evolution in Changing Environments". M . Hershkovitz, P., Fieldiana: Zool. 46:\, 35. Cody, M.L., Ecology 57:455, 1970. 36. Cody, M.L., 1962. In "Mediterranean Type Ecosystems". Di Castri, F., Mooney, H . A . , Eds. N e w York, SpringerVerlag, 1973, p. 307. 37. Vuilleumier, F. Nature 275:247, 1967. 38. Vuilleumier, F., Nature 2 79:87, 1968. 39. Cei, J.M., "Batracios de Chile". Santiago, Ediciones Universidad de Chile, 1962. 40. Stebbms, R.C., " A Field Guide to Western Reptiles and Amphibians". Cambridge, Mass., Riverside Press, 1966. 41. Di Castri, F., Mooney, H.A., Eds. "Mediterranean Type Ecosystems". N e w York, Springer-Verlag, 1973. 42. Thrower, N.J. W., Bradbury, D.E., Eds. "Chile-Califor­ nia Mediterranean Scrub Atlas". Stroudsburg, Pennsyl­ to de Mamíferos Chilenos". Santiago, Editorial Gabriela 26. Spotorno, Ecology 57:728, 1976. M.L., Princeton, Princeton University Press, 1968. 1975. 14. Pough, F.G., Ecology 54:837, 1973. 25. Miller, 28. M'Closkey, 30. Heller, H.C., 1974. 3:51, Cody, M L . , Diamond, J . M . , Eds. Cambridge, Mass., BelknaD Press, 1975, p. 315. Harper and Row, 1972. vania, Dowden, Hutchinson and Ross, 1977. 43. Pinto, A., Hermosilla, W., Di Castri, F., Astudillo, V., Inv. Zool. Chilenas (III Congreso Latinoamericano Zoo­ An. Mus. Hist. Nat. Valparaíso 9:141, 1976. 27. Brown, J.H., In "Ecology and Evolution of Communities". logía), 1965, p. 32. 44. Jerez, V., Ortiz, <S:58, 1975. J.C., An. M u s . Hist. Nat. Valparaiso