16 G.Newcombe and F.M.Dugan Records of eastem filbert blight,caused by Anisogramma anomala,are also restricted to North America (Farr et al.n.d.).But possibly because there were records of this disease from the late 1800s.A.anomala is not thought to be alien to North America(Sinclair and Lyon 2005).In fact,the native host of A.anomala is hypothesized to be Corylus americana (Coyne et al.1998).But if resistance patterns were the sole criterion for native range,the resistance of some species of Corylus native to each of North America.Europe,and Asia would again be et al.1998:Sinclair and Lyor 2005 anomala is not pre nt in Europe and ia,and is trulyn nthe re ian species of Cory ts m the exapted or nonho variety rather than adaptive.As some individuals of the European hazelnut C.avellana,are susceptible,nonhost resistance would seem to be ruled out.So the dominant."Gasaway gene that is inherited from resistant individuals of the European hazelnut(Coyne et al.1998)would be interpreted here as exapted. Seiridium cankers of cypress are caused by three species of Seiridium.Rela- tively resistant species of Calocedrus,Chamaec th Ame (Sinclai and Lyon 2005).On aga the lack of any discret geographic n to be adaptive would thwart any attempt to pin any of the thre species of Seiridium to any particular native range,at least using the sole criterion of resistance. Fusiform rust,caused by Cronartium quercuum f.sp.fusiforme,is also instruc tive.The fusiform rust fungus is thought by many to be native to the region where it is currently most damaging:the southeastern U.S.However,after testing 45 species of Pinus for susceptibility/resistance to C.quercuum f.sp.fusiforme,an origin of me in Central Am ypothesized (Tain Anderson 1993)This hypoth 0g with the latively strong of Pinus from Central Americ ver,e rong re f Asian a consistent in applying the criterion of resistance,the authors would then have had to propose a native range for C.quercuum f.sp.fusiforme involving widely scattered,disjunct populations in Asia,Europe,and Central America.quite unlike that of any species of Pinus that hosts the fusiform rust fungus The oak wilt fungus,Ceratocystis fagacearum,of the middle and eastern United States.has already been mentioned.Its ge netic hom eneity has led some research othesize origin(Ju vik etal).But,it is in mp sible toinfer of C.fagacearum host one for h an that unde erstudied in some 30 species o Quercus (Mabberley 2008).So,in addition to the challenging need to distinguish adaptive and exapted/nonhost resistance,undersampling issues can be formidable One might imagine that difficulties in determining native ranges of fungi from resistance are only encountered when alien pathogens are obscure and of little importance.But not only are the examples just cited important,but even fungi as important as the Ophiosto ma species that have caused global pandemics of Dutch elm disease in the past c entury,re nain of uncertain,geographic prigin (Brasier and
Records of eastern filbert blight, caused by Anisogramma anomala, are also restricted to North America (Farr et al. n.d.). But possibly because there were records of this disease from the late 1800s, A. anomala is not thought to be alien to North America (Sinclair and Lyon 2005). In fact, the native host of A. anomala is hypothesized to be Corylus americana (Coyne et al. 1998). But if resistance patterns were the sole criterion for native range, the resistance of some species of Corylus native to each of North America, Europe, and Asia would again be problematic (Coyne et al. 1998; Sinclair and Lyon 2005). Furthermore, if A. anomala is not present in Europe and Asia, and is truly native to North America, then the resistance of Eurasian species of Corylus must be of the exapted or nonhost variety rather than adaptive. As some individuals of the European hazelnut, C. avellana, are susceptible, nonhost resistance would seem to be ruled out. So, the dominant, “Gasaway” gene that is inherited from resistant individuals of the European hazelnut (Coyne et al. 1998) would be interpreted here as exapted. Seiridium cankers of cypress are caused by three species of Seiridium. Relatively resistant species of Calocedrus, Chamaecyparis, Cupressus, Juniperus, Taxodium, Thuja, and Thujopsis are native to both Eurasia and North America (Sinclair and Lyon 2005). Once again, the lack of any discrete, geographic source of resistance known to be adaptive would thwart any attempt to pin any of the three species of Seiridium to any particular native range, at least using the sole criterion of resistance. Fusiform rust, caused by Cronartium quercuum f.sp. fusiforme, is also instructive. The fusiform rust fungus is thought by many to be native to the region where it is currently most damaging: the southeastern U.S. However, after testing 45 species of Pinus for susceptibility/resistance to C. quercuum f.sp. fusiforme, an origin of C. quercuum f.sp. fusiforme in Central America was hypothesized (Tainter and Anderson 1993). This hypothesis was congruent with the relatively strong resistance of species of Pinus from Central America. However, equally strong resistance of Asian and Mediterranean species was evident in this study. To be consistent in applying the criterion of resistance, the authors would then have had to propose a native range for C. quercuum f.sp. fusiforme involving widely scattered, disjunct populations in Asia, Europe, and Central America, quite unlike that of any species of Pinus that hosts the fusiform rust fungus. The oak wilt fungus, Ceratocystis fagacearum, of the middle and eastern United States, has already been mentioned. Its genetic homogeneity has led some researchers to hypothesize an exotic origin (Juzwik et al. 2008). But, it is impossible to infer the native range of C. fagacearum from host resistance alone for no other reason than that this subject remains seriously understudied in some 530 species of Quercus (Mabberley 2008). So, in addition to the challenging need to distinguish adaptive and exapted/nonhost resistance, undersampling issues can be formidable. One might imagine that difficulties in determining native ranges of fungi from resistance are only encountered when alien pathogens are obscure and of little importance. But not only are the examples just cited important, but even fungi as important as the Ophiostoma species that have caused global pandemics of Dutch elm disease in the past century, remain of uncertain, geographic origin (Brasier and 16 G. Newcombe and F.M. Dugan
1 Fungal Pathogens of Plants in the Homogocene 1 Buck 2001).Surveys in China(Brasier 1990)and in the Himalayas(Brasier and Mehrotra 1995)were undertaken because resistant species of Ulmus are naturally distributed there.However.o.ulmi and o.novo-ulmi.were not found in eithe surveyed region.Absence of the relevant pathogens implies that the r resistance of elm exapted or nonhos rather than adaptive.It should be noted tha nd spec elated to the Dutch elm fu ssible gre debated. cause (summarily reviewed in Dugan 2008). The above examples should not only encourage caution in inferring native ranges of fungi from resistance.They should also provoke questions about the threat of novel pathogens of plants.Even if adaptive resistance is obviously,by definition,absent in naive plants encountering novel pathogens,could not exapted or nonhost resistance protect them,and if so,at what frequency? 1.7 First Encounters Between Evolutionarily Naive Plants and Novel Pathogens As previously mentioned,predicting outcomes of invasions is one of the central objectives of invasion biology(Kolar and Lodge 2001).With plants and animals relational hypotheses fo either on the relative invasive ess of potential the ive invasibility yof po tially invaded ities (Hege and T ).Relational hypoth ses consic er both.In all cases,h owever,it i considered essential to know the original,pre-Homogocene,geographic ranges of the organisms in question to develop and test hypotheses.To test the pathogen release hypothesis(Keane and Crawley 2002),for example,one needs to compare the enemies of a particular plant or animal in its native and invaded ranges.Novel weapons might aid a plant invader but only when wielded against species that are evolutionarily naive in the sense that they have never faced the weapons in question (Callaway and Ridenour 2004).Even Darwin's hypothesis,the first hypothesis of invasive ess (Rejmanek 1996).that invasive ies ar rom alier enera than from genera a found in both ranges is,of course,predicated on knowing se ranges are (Darwin 1859) For fungal pathogens involved in pathogen reunions,prediction of outcomes is straightforward.Plant species"ABC,"known to be susceptible in its native range to pathogen"abc."is likely to be susceptible everywhere else to"abc."provided that the environment is conducive to infection and disease expression.We might assume that the native range of"abe"is the same as that of"ABC,"but all we really need to know for predictive purposes is that the latter is susceptible to the former some where else.But,pa reunions are e not first enc unt Outc st enc ers b ve plants vel pathog be much mo enging to predict. the assumption that for predictive purposes the native ranges of phytopathogenic
Buck 2001). Surveys in China (Brasier 1990) and in the Himalayas (Brasier and Mehrotra 1995) were undertaken because resistant species of Ulmus are naturally distributed there. However, O. ulmi and O. novo-ulmi.were not found in either surveyed region. Absence of the relevant pathogens implies that the resistance of Asian elms is exapted or nonhost, rather than adaptive. It should be noted that Ophiostoma species related to the Dutch elm fungus are one possible, and greatly debated, cause of the great mortality of elms in Europe during the Neolithic (summarily reviewed in Dugan 2008). The above examples should not only encourage caution in inferring native ranges of fungi from resistance. They should also provoke questions about the threat of novel pathogens of plants. Even if adaptive resistance is obviously, by definition, absent in naive plants encountering novel pathogens, could not exapted or nonhost resistance protect them, and if so, at what frequency? 1.7 First Encounters Between Evolutionarily Naive Plants and Novel Pathogens As previously mentioned, predicting outcomes of invasions is one of the central objectives of invasion biology (Kolar and Lodge 2001). With plants and animals, nonrelational hypotheses focus either on the relative invasiveness of potential invaders or the relative invasibility of potentially invaded communities (Heger and Trepl 2003). Relational hypotheses consider both. In all cases, however, it is considered essential to know the original, pre-Homogocene, geographic ranges of the organisms in question to develop and test hypotheses. To test the pathogen release hypothesis (Keane and Crawley 2002), for example, one needs to compare the enemies of a particular plant or animal in its native and invaded ranges. Novel weapons might aid a plant invader but only when wielded against species that are evolutionarily naive in the sense that they have never faced the weapons in question (Callaway and Ridenour 2004). Even Darwin’s hypothesis, the first hypothesis of invasiveness (Rejma´nek 1996), that invasive species are more likely from alien genera than from genera found in both ranges is, of course, predicated on knowing what those ranges are (Darwin 1859). For fungal pathogens involved in pathogen reunions, prediction of outcomes is straightforward. Plant species “ABC,” known to be susceptible in its native range to pathogen “abc,” is likely to be susceptible everywhere else to “abc,” provided that the environment is conducive to infection and disease expression. We might assume that the native range of “abc” is the same as that of “ABC,” but all we really need to know for predictive purposes is that the latter is susceptible to the former somewhere else. But, pathogen reunions are not first encounters. Outcomes of true, first encounters between naive plants and novel pathogens appear to be much more challenging to predict. Thus far, we have been operating under the assumption that for predictive purposes the native ranges of phytopathogenic 1 Fungal Pathogens of Plants in the Homogocene 17
18 G.Newcombe and F.M.Dugan fungi must be known.However,the status of a first encounter may be ascertained merely by knowing that the two parties differ in their native ranges.In some cases we may know their native ranges:the host switching that.for instance.occurred when naive eucalyptus was introduced to south america and novel puccinia psidii switched to Eucalyptus from native Myrtaceae.But,in the case of Dutch elm disease or dogwood anthracnose,as discussed above.neither the native ranges of the novel pathogens nor the identities of their adaptive hosts have ever been determined.Nev rtheless,it is clear tha t the ese diseases do represe ent firs encounters switching ch encou ters al orepresent the fu ungal co mponent of what invasion biologists call "biotic resistance(Parker and Gilbert 2004).Any and all organisms in a native community can provide biotic resistance to repel invaders.When pathogenic fungi switch from native plants to naive,alien plants,as in rust of Eucalyptus,they contribute to biotic resistance.This scenario represents one of three categories of first encounters between naive plants and novel pathogens that we emphasize here (Table 1.1):(1)alien plants versus native pathogens:(2)native plants versus alien ；a the of alien thei (J nes and Bat ative en on alie en but natur er 2007)(3)native plants versus alien ogens not ed ir pathogen reunion nrst tw categories allow pathogens considerable lag periods during which host switching may occur from their adaptive hosts.It is important to recall that alien organisms frequently become invasive only after considerable lag periods(Mack et al.2000).The third category is distinct because it involves alien pathogens that must switch to the naive host immediately upor introduction because their adaptive host is absent. The null h othesis might be ued that these distinctions of thre e cate unn cess ily naiv are clo s be de b alien path partic cularly if the pathogen in question er rus t and c surely suggest as much.In the absence of selection for resistance,is not genetic susceptibility to pathogens of exotic congeners inevitable and complete?The shor answer is no.Much of what we have already discussed implies this.The easiest wav to expand that answer is to further discuss resistance to pathogens that were clearly alien.starting with the white pine blister rust fungus.Cronartium ribicola.The latter was a"Category 3"alien pathogen in North America(Table 1.1).but with ar risk.it wa adaptiv host bu D strobus,to w had already de North An SW rica (Kinloch 2003 onartium ribi a is d n ve to eas sia where are thought to include Pinus sibirica. wallichia and P.pumila(Kinloch 2003:Kinloch and Dupper 2002:Sinclair and Lyon 2005) The host range of C.ribicola spans the species of Pinus belonging to subgenus Strobus that includes sections Quinquefoliae and Parrya (Gerandt et al.2005). These sections are especially speciose in North America and Asia,and less so in Europe where P.cembra and P.peuce are native.When the latter two European spe species. and eight North america species tested for
fungi must be known. However, the status of a first encounter may be ascertained merely by knowing that the two parties differ in their native ranges. In some cases, we may know their native ranges: the host switching that, for instance, occurred when naive Eucalyptus was introduced to South America and novel Puccinia psidii switched to Eucalyptus from native Myrtaceae. But, in the case of Dutch elm disease or dogwood anthracnose, as discussed above, neither the native ranges of the novel pathogens nor the identities of their adaptive hosts have ever been determined. Nevertheless, it is clear that these diseases do represent first encounters that must involve host switching. Such encounters also represent the fungal component of what invasion biologists call “biotic resistance” (Parker and Gilbert 2004). Any and all organisms in a native community can provide biotic resistance to repel invaders. When pathogenic fungi switch from native plants to naive, alien plants, as in rust of Eucalyptus, they contribute to biotic resistance. This scenario represents one of three categories of first encounters between naive plants and novel pathogens that we emphasize here (Table 1.1): (1) alien plants versus native pathogens; (2) native plants versus alien pathogens involved in pathogen reunions, i.e., the majority of alien pathogens, which exist in their new, non-native environments on alien but naturalized hosts (Jones and Baker 2007); (3) native plants versus alien pathogens not involved in pathogen reunions. The first two categories allow pathogens considerable lag periods during which host switching may occur from their adaptive hosts. It is important to recall that alien organisms frequently become invasive only after considerable lag periods (Mack et al. 2000). The third category is distinct because it involves alien pathogens that must switch to the naive host immediately upon introduction because their adaptive host is absent. The null hypothesis might be argued that these distinctions of three categories are unnecessary because evolutionarily naive plants will always be decimated by alien pathogens, particularly if the plants are closely related to the host of the alien pathogen in question. The examples of white pine blister rust and chestnut blight surely suggest as much. In the absence of selection for resistance, is not genetic susceptibility to pathogens of exotic congeners inevitable and complete? The short answer is no. Much of what we have already discussed implies this. The easiest way to expand that answer is to further discuss resistance to pathogens that were clearly alien, starting with the white pine blister rust fungus, Cronartium ribicola. The latter was a “Category 3” alien pathogen in North America (Table 1.1), but with an asterisk; it was introduced without an adaptive host but on a naive host, Pinus strobus, to which it had already switched outside North America (Kinloch 2003). Cronartium ribicola is considered native to eastern Asia where its adaptive hosts are thought to include Pinus sibirica, P. armandii, P. koraiensis, P. wallichiana, and P. pumila (Kinloch 2003; Kinloch and Dupper 2002; Sinclair and Lyon 2005). The host range of C. ribicola spans the species of Pinus belonging to subgenus Strobus that includes sections Quinquefoliae and Parrya (Gernandt et al. 2005). These sections are especially speciose in North America and Asia, and less so in Europe where P. cembra and P. peuce are native. When the latter two European species, seven Asian species, and eight North American species were tested for 18 G. Newcombe and F.M. Dugan
1 Fungal Pathogens of Plants in the Homogocene 19 blister rust resistance in Europe,only P.cembra and the Asian species,P.armandii and P.pumila were completely resistant (Stephan 2001).Among the North American species.P.aristata of section Parrya was more resistant than the P.lamb enting section Quinquefoliae:P.strobifor rtiana.P.albicaulis.P.flexilis.P.m cola,and P. robus.When bliste t resista d in North Americ strobif was more resistant than other species of North The point of emphasis here is the fact that none of the North American species encountering C.ribicola for the first time were completely susceptible because resistant individuals have been found in each (Stephan 2001:Sniezko et al. 2008).In fact,four species (i.e.,P.strobiformis,P.monticola,P.fexilis,and P.lambertiana)have been shown in separate studies to possess major genes for resistance.albeit at low frequencies (Kinloch 1992:Kinloch and Dupper 2002 Kinloch et al.1999.2003).Altho gh phe evidence for these Cr not de ected in whitebark pine (P.ic lis)Mex can white pir e(P.ayace wa foxtail pine (.rna),and Great Basin bristlecone pine longaeva).all of these species might possess such genes at low frequencies that would simply require additional sampling for their discovery (Kinloch and Dupper 2002).The authors conclude that although"blister rust traditionally is considered an exotic disease in North America,these results,typical of classic gene-for-gene interac tions,suggest that genetic memory of similar encounters in past epochs has been retained in this pathosystem"(Kinloch and Dupper 2002). Befor onsidering mples of exapted resistance other than the Crge nes the implicat gene nla tion.Dis in plant often tackl require by using som eptual dich my. Van Der Pla famously distinguished between vertical and horizontal res nce for example (Van Der Plank 1975).Gene-for-gene interactions characterize Van Der Plank's vertical resistance (Briggs and Johal 1994:Flor 1971;Thompson and Burdon 1992).Flor(1971)developed the gene-for-gene theory by performing correlated studies of the inheritance of both host resistance and pathogen virulence using cultivated flax and flax rust.Flor is typically quoted for defining these interaction in this way:"for each gene that conditions reaction in the host there is a ling ge e tha ons path ons can als nferred,some h.ce city in th e p and F -gen 992).by proving that there are a number of nct,major genes resistan that allow pathogen isolates to be differentially distinguished as pathotypes.By this definition,gene-for-gene interactions do characterize white pine blister rus (Kinloch and Dupper 2002)and also poplar leaf rust that we shall discuss next as it also involves the sudden appearance of "genetic memory"in first encounters (Newcombe et al.2001). Recall that gene-for-gene interactions are thought to be the product of continu ous coevolution (Person 1959.1967).But both the poplar leaf rust pathosystem of e bliste st pathosystems of the whit pine America appear be the produd exapted resistance genes,and recent,adaptive changes in the pathogen populations
blister rust resistance in Europe, only P. cembra and the Asian species, P. armandii and P. pumila were completely resistant (Stephan 2001). Among the North American species, P. aristata of section Parrya was more resistant than the seven species representing section Quinquefoliae: P. strobiformis, P. balfouriana, P. lambertiana, P. albicaulis, P. flexilis, P. monticola, and P. strobus. When blister rust resistance was tested in North America, P. strobiformis was more resistant than other species of North American origin (Sniezko et al. 2008). The point of emphasis here is the fact that none of the North American species encountering C. ribicola for the first time were completely susceptible because resistant individuals have been found in each (Stephan 2001; Sniezko et al. 2008). In fact, four species (i.e., P. strobiformis, P. monticola, P. flexilis, and P. lambertiana) have been shown in separate studies to possess major genes for resistance, albeit at low frequencies (Kinloch 1992; Kinloch and Dupper 2002; Kinloch et al. 1999, 2003). Although phenotypic evidence for these Cr genes was not detected in whitebark pine (P. albicaulis), Mexican white pine (P. ayacahuite), foxtail pine (P. balfouriana), and Great Basin bristlecone pine (P. longaeva), all of these species might possess such genes at low frequencies that would simply require additional sampling for their discovery (Kinloch and Dupper 2002). The authors conclude that although “blister rust traditionally is considered an exotic disease in North America, these results, typical of classic gene-for-gene interactions, suggest that genetic memory of similar encounters in past epochs has been retained in this pathosystem” (Kinloch and Dupper 2002). Before considering examples of exapted resistance other than the Cr genes, the implications of gene-for-gene interactions require explanation. Disease resistance in plants is often tackled by using some conceptual dichotomy. Van Der Plank famously distinguished between vertical and horizontal resistance, for example (Van Der Plank 1975). Gene-for-gene interactions characterize Van Der Plank’s vertical resistance (Briggs and Johal 1994; Flor 1971; Thompson and Burdon 1992). Flor (1971) developed the gene-for-gene theory by performing correlated studies of the inheritance of both host resistance and pathogen virulence using cultivated flax and flax rust. Flor is typically quoted for defining these interactions in this way: “for each gene that conditions reaction in the host there is a corresponding gene that conditions pathogenicity in the pathogen.” Gene-for-gene interactions can also be inferred, somewhat less rigorously (Thompson and Burdon 1992), by proving that there are a number of distinct, major genes for resistance that allow pathogen isolates to be differentially distinguished as pathotypes. By this definition, gene-for-gene interactions do characterize white pine blister rust (Kinloch and Dupper 2002) and also poplar leaf rust that we shall discuss next as it also involves the sudden appearance of “genetic memory” in first encounters (Newcombe et al. 2001). Recall that gene-for-gene interactions are thought to be the product of continuous coevolution (Person 1959, 1967). But both the poplar leaf rust pathosystem of the Pacific Northwestern region and the blister rust pathosystems of the white pines of North America appear to be the product of recent pathogen introductions, exapted resistance genes, and recent, adaptive changes in the pathogen populations. 1 Fungal Pathogens of Plants in the Homogocene 19
20 G.Newcombe and F.M.Dugan Attempts to explain the Crgenes in terms of selection have been made:the highest frequencies of Crl and Cr2 are in the American Southwest near overlaps with pinyon pines of section Parrya and pinyon blister rust caused by an American native rust fungus,Cronartium occidentale(Kinloch and Dupper 2002).However, to positively demonstrate that C.occidentale was the selective agent that explains evolutionary retention of Cr genes,one would have to show that Cr genes protect species in section Quinquefoliae against C.occidentale (Kinloch and Dupper 2002).Otherwise,the ention es in No th Ame ca i Dupper nes appear par doxical given th of selection 02 the cas ce of poplar leaf rust also.resistanc e genes were revealed by pathogen introductions.raising again the question of their retention in the absence of selec tion.Complex pathogenic variation indicative of gene-for-gene interactions also appeared very quickly in this system,once the pathogen population had undergone hybridization to match that of its hybrid host (Newcombe et al.2001).Some background is needed to explain current gene-for-gene interactions in poplar leaf h Pacific Northwes of No Americ Popls trichocarpa cottonw ood,is tive to the 0 ，o with its Whe nP. vith P trichocarpa (T) wood m section cros les (D). h om growing F hybrid clones can b e selected.These TxD have been the mainstay of commercial poplar plantations in the region for nearly three decades. Initially TxD hybrids were rust-free.The resistance of P. deltoides to M.occidentalis (Newcombe et al.2000).was transmitted to all TxD F hybrids indicating that these P.deltoides parents are dominant homozygotes in this respect. In 1991.Melampsora medusae.the coevolutionary rust of p.deltoides.was found in the on.It ickly be nt that s TxD F:hybrids w ble to M. An th medusae in a TxD ar pedig cdcmonstatedh t the M e ce was ocarpa parent (N 96.n note that this gene had gone unnoticed in previous studies of the resistance of P.trichocarpa to M.occidentalis.The gene-for-gene explanation for the detection of Mmdl with M.medusae is that the latter evidently possesses the matching avirulence allele,unlike the coevolutionary rust,M.occidentalis.Both P.trichocarpa and M.medusae could even be fixed for this gene-for-gene pair as their inter action phenotype was alwavs resistant,altho gh testin was limited to nin individuals the latte m the be et al ).R istance to M. me e was ob xD F cause the P trichocarpa parent was heterozygous at Mn gregate in the et al 190 Until 1995.there was no pathogenic varation in the rust population that simply consisted of M.medusae.F clones either possessed the dominant Mmdl allele fo resistance,or not.Emergence in the mid-1990s of the hybrid of M.medusae and M.occidentalis.M.x columbiana.changed this situation.Previously resistan FI clones became susceptibyopareni that there way abundant pathogenic variation in the new of M.x columbiand
Attempts to explain the Cr genes in terms of selection have been made; the highest frequencies of Cr1 and Cr2 are in the American Southwest near overlaps with pinyon pines of section Parrya and pinyon blister rust caused by an American native rust fungus, Cronartium occidentale (Kinloch and Dupper 2002). However, to positively demonstrate that C. occidentale was the selective agent that explains evolutionary retention of Cr genes, one would have to show that Cr genes protect species in section Quinquefoliae against C. occidentale (Kinloch and Dupper 2002). Otherwise, the retention of Cr genes in North American white pines appears paradoxical given the absence of selection (Kinloch and Dupper 2002). In the case of poplar leaf rust also, resistance genes were revealed by pathogen introductions, raising again the question of their retention in the absence of selection. Complex pathogenic variation indicative of gene-for-gene interactions also appeared very quickly in this system, once the pathogen population had undergone hybridization to match that of its hybrid host (Newcombe et al. 2001). Some background is needed to explain current gene-for-gene interactions in poplar leaf rust in the Pacific Northwest of North America. Populus trichocarpa, the western black cottonwood, is native to the region, along with its coevolutionary rust, Melampsora occidentalis. When P. trichocarpa (T) from section Tacamahaca is crossed with P. deltoides (D), the eastern cottonwood, from section Aigeiros, fastgrowing F1 hybrid clones can be selected. These TxD hybrids have been the mainstay of commercial poplar plantations in the region for nearly three decades. Initially TxD hybrids were rust-free. The resistance of P. deltoides to M. occidentalis (Newcombe et al. 2000), was transmitted to all TxD F1 hybrids indicating that these P. deltoides parents are dominant homozygotes in this respect. In 1991, Melampsora medusae, the coevolutionary rust of P. deltoides, was found in the region. It quickly became apparent that some TxD F1 hybrids were susceptible to M. medusae. Analysis of the inheritance of resistance to M. medusae in a TxD hybrid poplar pedigree demonstrated that the Mmd1 gene for resistance was inherited from the P. trichocarpa parent (Newcombe et al. 1996). It is important to note that this gene had gone unnoticed in previous studies of the resistance of P. trichocarpa to M. occidentalis. The gene-for-gene explanation for the detection of Mmd1 with M. medusae is that the latter evidently possesses the matching avirulence allele, unlike the coevolutionary rust, M. occidentalis. Both P. trichocarpa and M. medusae could even be fixed for this gene-for-gene pair as their interaction phenotype was always resistant, although testing was limited to nine individuals of the former and four isolates of the latter from the southeastern U.S. (Newcombe et al. 2000). Resistance to M. medusae was observed to segregate in the TxD F1 because the P. trichocarpa parent was heterozygous at Mmd1 (Newcombe et al. 1996). Until 1995, there was no pathogenic variation in the rust population that simply consisted of M. medusae. F1 clones either possessed the dominant Mmd1 allele for resistance, or not. Emergence in the mid-1990s of the hybrid of M. medusae and M. occidentalis, M. x columbiana, changed this situation. Previously resistant TxD F1 clones became susceptible. It rapidly became apparent that there was abundant pathogenic variation in the new hybrid population of M. columbiana 20 G. Newcombe and F.M. Dugan