Diversity of Myxomycetes in arid zones of Peru part II: the cactus belt and transition zones ; Diversidad de Myxomycetes en zonas áridas de Perú, parte II: el cardonal y las zonas de transición

The results obtained from a second survey for Myxomycetes in the arid areas of Peru are reported. A total of 37 localities from the cactus belt (‘cardonal’), between 1500 and 3000 m a.s.l., were sampled over six years. This survey is based on 601 identifiable collections of myxomycetes, developed in the field under natural conditions or those that were recovered from moist chamber cultures. In total, 84 taxa representing 19 genera were recorded. One new species, Cribraria spinispora Lado & D.Wrigley sp. nov., is described, three species new to South America and 22 additional new species for Peru are reported as well. Endemic plants, expecially cacti, had a particular relevance as myxomycete substrates. The predominance of the order Physarales T.Macbr. in arid areas is reinforced. Comments are included on some collections, as well as SEM micrographs of several species. A statistical evaluation of the diversity of myxomycetes in the cardonal area and a comparison of this area with the dry coastal desert of Peru and other Neotropical xeric environments are also included. Results show that the arid areas are rich in myxomycetes, and that each area has a unique assemblage. The differences appear to be related to the host plants. Resumen. Se presentan los resultados de un segundo estudio sobre Myxomycetes de zonas áridas de Perú. Un total de 37 localidades del cardonal, situadas entre 1500 y 3000 m s.n.m., fueron muestreadas durante seis años. Este estudio está fundamentado en 601 recolecciones de mixomicetes, obtenidas en el campo o por cultivo en cámara húmeda. En total, se han identificado 84 táxones que corresponden a 19 géneros. Se describe una nueva especie, Cribraria spinispora Lado & D.Wrigley sp. nov., además se registran por primera vez tres especies para Sudamérica y 22 para Perú. Las plantas endémicas, especialmente los cactus, han tenido una especial importancia como substrato para los mixomicetes. Se reafirma el predominio del orden Physarales T.Macbr. en zonas áridas. Se incluyen comentarios sobre algunas recolecciones, así como micrografías al MEB de algunas especies. Se realizó una evaluación estadística de la diversidad de mixomicetes en el cardonal y una comparación entre esta zona, la costa seca de Perú y otros ambientes neotropicales xéricos. Los resultados muestran que las zonas áridas son ricas en mixomicetes y que cada una tiene un conjunto exclusivo de especies. Las diferencias parecen estar relacionadas con las plantas anfitrionas. How to cite this article: Lado C., Wrigley de Basanta D., Estrada-Torres A., Stephenson S.L. & Treviño I. 2019. Diversity of Myxomycetes in arid zones of Peru part II: the cactus belt and transition zones. Anales del Jardín Botánico de Madrid 76 (2): e083. https://doi.org/10.3989/ajbm.2520 Title in Spanish: Diversidad de Myxomycetes en zonas áridas de Perú, parte II: el cardonal y las zonas de transición. Received: 11‒I‒2019; accepted: 30‒IV‒2019; published on-line: 28‒X‒2019; Associate Editor: M. Dueñas.


INTRODUCTION
The Myxomycetes G.Winter or plasmodial slime moulds are a group of microorganisms considered as a protozoan fungal analogue (Kirk & al. 2011). The Myxomycetes have been found in all terrestrial ecosystems, and more than 1,000 species are currently recognized (Lado 2005(Lado -2018. The myxobiota of the arid or semi-arid environments, despite the lack of water, is surprisingly high, since they develop adaptive strategies to survive in these extreme environments, as exemplified by the succulenticolus myxomycetes (Lado & al. 1999). A total of 191 different species have been reported by our team from arid regions of the Neotropics (Lado & al. 2007a(Lado & al. , 2013. The elevated numbers of myxomycete species is particularly interesting, since 24 of the species found in these surveys were previously unknown, and have been described as new to science (Lado & al. 1999(Lado & al. , 2007b The moist chamber cultures were prepared in the manner described by . The pH of each culture was determined after 24 h, and the cultures were maintained at room temperature (21-25 °C) in diffuse daylight and examined at regular intervals for a period of up to three months. As the myxomycete sporophores were formed, the portion of the substrate upon which they occurred was removed from the moist chamber culture, allowed to dry slowly in a closed empty Petri dish and then glued into a small cardboard box. All sporophores of a given species that developed in the same culture were considered to represent a single record. All microscope measurements and observations were made with material mounted directly in Hoyer's medium or polyvinyl alcohol (PVA). Differential interference contrast (DIC) microscopy was used to obtain descriptive data. Some specimens were examined, in the Real Jardín Botánico de Madrid CSIC, with a Hitachi S-3000N scanning electron microscope (SEM) at 10-15 kV. For all SEM-photographs the critical point dried material technique was employed. Colour notations in parenthesis are from the ISCC-NBS Color Name Charts Illustrated with Centroid Colors (Kelly & Judd 1976).
The ACE and the CHAO1 abundance indices (Colwell & Coddington 1994;Colwell & al. 2004) were used to estimate how complete the sampling effort was. The accumulation curve was adjusted according to the Clench function, where 'S n = (a · n) / [1 + (b · n)]', where 'S n ' is the number of species accumulated for a unit of collecting effort ('n') (Jiménez-Valverde & Hortal 2003). Each collecting site was considered as the unit of collecting effort, using the total number of species found with the program EstimateS v. 9.1.0 (Colwell 2013), adjusted so that '3' was the upper abundance limit for rare or infrequent species, following the criteria of Stephenson & al. (1993) and considering the species with a relative abundance less than or equal to 0.05 to be rare. With the same program, the richness of species with 1.5 times more sampling effort was calculated. The adjustment according to the Clench function was carried out with the program Statistica v. 12, using the Simplex and Quasi-Newton method of adjustment (Jiménez-Valverde and Hortal 2003). ACE and CHAO1 abundance indices were calculated for all results, only field results and only moist chamber culture results.
To compare the results from this area with those of other studies carried out in arid areas in the Americas by the same team, the complementarity has been calculated, using the formula proposed by Colwell & Coddington (1994): 'C jk = U jk / S jk ', where 'U jk = S j + S k -(2V jk )' and 'S jk = S j + S k -V jk ', where 'U jk ' is the number of species that are different for the two communities, 'S jk ' is the total richness of species for both communities, 'S j ' is the number of species from the first community, 'S k ' is the number of species from the second and 'V jk ' is the number of species common to both.
A test of independence was done using a contingency table and a χ² test (Zar 1996) to determine whether the proportion of the collections, and the species from different orders of myxomycetes was dependent on the sampling method used, or whether these were due to the vegetation type of the survey area. If there was a difference, a standardized residual analysis was done to see if the proportions were statistically significant between the observed and expected values (Sheskin 2004). All specimens are deposited in the herbarium MA-Fungi (sub C. Lado) with duplicates in the herbaria TLXM (sub A. Estrada-Torres), UARK (sub S.L. Stephenson) or in the private collection of D. Wrigley de Basanta ('dwb').

RESULTS
As a result of the survey of these arid and semiarid areas carried out in six different years, 601 collections of myxomycetes, either specimens that had developed in the field or those that were recovered from moist chamber cultures, were identified. In total, 84 taxa representing 19 genera of myxomycetes were recorded. Of these, one is a new species of the genus Cribraria Pers., 1794 that is described below, three are new records for South America, namely Didymium umbilicatum D. Wrigley & al., Hemitrichia succulenticola G. Moreno & al. and Stemonitopsis amoena (Nann.-Bremek.) Nann.-Bremek., and 22 more are new records for Peru.
The myxomycete collections from this survey are arranged alphabetically in the list that follows by genus and then species. Nomenclature follows Lado (2005Lado ( -2018 unless otherwise stated. Information is provided on the locality from which each record was collected, the substrate upon which it was collected, the source of each record (either a field collection or a collection obtained from a moist chamber culture), and the pH of the moist chamber culture in which the specimen appeared. All dates given are substrate collection dates, with the exception of the new species, in which the dates are those obtained in moist chamber culture. All identified collections are included with comments on their distribution. The distribution of the species in the Neotropics is based on Lado & Wrigley de Basanta (2008), but the information on the area of study has been updated with added references. New records for South America or Peru are marked with a degree symbol or an asterisk, respectively. Species of particular interest have additional taxonomic comments.
Notes.-In the Neotropics, it was previously reported from Ecuador, Mexico and Chile, but this is the first record from Peru. Only two collections from the same locality, the specimens agree with the description provided by Nannenga-Bremekamp (1991), except for the more prominent ornamentation of the capillitium. Notes.-Without doubt, it is the commonest species of the study area, linked virtually with all the cacti and succulent substrata examined. It has been collected in more than 50% of the localities sampled, and it was also the commonest species of the arid lands of the coastal desert of Peru  Notes.-This is the second record of this species from Peru and enlarges its distribution, since previously it was reported from the north, in the arid lands of Piura . In the Neotropics it is also known from Argentina, Ecuador, Mexico and Panama (Illana & al. 2000;Moreno & al. 2001Moreno & al. , 2007Moreno & al. , 2015Lizárraga & al. 2015b;Lado & al. 2017;Rojas & al. 2017). Only one collection of several minute sporocarps, with a columella reaching about ½ of the sporotheca, dividing to form the main branches of capillitium, and with the dark capillitium branching and anastomosing ( fig. 3a, b) and forming the typical surface net with few free ends ( fig. 3c) that matches the original description of Härkönen (1977Härkönen ( , 1978. Notes.-It is widely distributed in Peru and was previously reported from Ancash, Arequipa, Ayacucho, Lima, Madre de Dios and Tacna .

Comatricha laxa
Comatricha pulchella (C.Bab.) Rostaf. Notes.-In Peru it was previously reported from Madre de Dios and Ancash by  and , respectively.  Notes.-It is a widely distributed species, known in Peru from Ancash, Cajamarca and Madre de Dios  s. y Br) by transmitted light, darker at the base and paler and concolourous with the peridium at the tip. Peridium single, pink grayish (8. gy. Pink-18. l. gy Pink-32. gy. y Pink) by transmitted light, partially evanescent, remaining in the basal 1 /3-1 /2 as a calyculus, and as a net without nodes or with few fusiform nodes above; calyculus membranous, with radial striations, without perforate margin; net lax mostly isodiametric, threads 2-3 µm in diameter, sometimes with few slightly expanded areas but not thickened, with occasional free-ending threads; dehiscence apical by disintegration of the peridium to leave the peridial net. Columella absent. Capillitium absent. Spores 9-11 µm in diameter, free, dark purplish (224. d. P) in mass, pale pink (8. p. Pink-32. gy. y Pink) by transmitted light, slightly angular or polygonal, densely spinulose, the spines in the centre of the polygons by SEM. Calcic granules 1-2 µm in diameter, concentrated in the peridial net and the upper border of the calyculus, dispersed in the rest of the calyculus, pinkish (4. l. Pink-8. gy. Pink). Plasmodium not observed.
Etymology.-The epithet 'spinispora' refers to the spines present as ornamentation on the spores. Notes.-The principal characters that distinguish this minute species from others in the genus are the larger diameter of the spores, 9-11 µm, that are pinkish, angular or polygonal and ornamented with spines, the spines more concentrated in the centre of the polygons, clearly visible by SEM ( fig. 4i, j). The peridium is evanesent in the upper part but leaves a mostly isodiametric net, with bands filled with calcic granules ( fig. 4g). Occasionally, the bands are slightly flattened or fusiform but do not form well-defined plates or nodes, and have only occasional free ends. The calyculus reaches one third to one half of the diameter of the sporotheca ( fig. 4a-c, k-n), and the upper border is filled with calcic granules, darker and well-defined ( fig. 4g). The colour of the sporotheca is dark purple to dark violaceous. The combination of these characters does not match any described species of Cribraria Pers. and so we describe it here as a new species.
Cribraria confusa Nann.-Bremek. & Y.Yamam., Cribraria minutissima Schwein. and Cribraria rufa (Roth) Rostaf. have an evanescent peridium that also leaves an isodiametric net, but differ in the size of the spores (6-8 µm in diameter), and the brown colour of the sporotheca (Martin & Alexopoulos 1969;Nannenga-Bremekamp 1991;Nannenga-Bremekamp & Yamamoto 1983). Cribraria paucidictyon Yu Li also has a peridial network and a size of the sporocarps similar to Cribraria spinispora sp. nov., but differs in the colour of the sporotheca, that is brown to chestnut-brown, and in the size (diameter 6.3-7.5 µm) and the ornamentation of the spores, illustrated by SEM and defined by Li (2002) as "minutely and densely covered with long warts, with incomplete and faint reticulation combined by long warts". Regarding the colour of the sporotheca the new species resembles Cribraria violacea Rex, Cribraria tecta Hooff and Cribraria purpurea Schrad., but the spores in all three cases are smaller, with a diameter less than 8 µm. In addition, in Cribraria violacea the colour of the sporotheca is violet ( fig. 4h) and the peridial net forms flat plates with many free ends, the spores by SEM, according to the illustration provided by Moreno & al. (2001), are baculate with small crests instead of spiny. Cribraria tecta is differentiated by the apical peridial plate that remains as a parachute (van Hooff 2009) and the spores by SEM are ornamented with dispersed spines with a broad base. Cribraria purpurea is also distinguished by the larger dimensions of the fruiting bodies (Martin & Alexopoulos 1969), 1.5-2.5(3) mm in total height vs. 0.5-0.9 mm in the new species, as well as the sporotheca, with a diameter 0.6-1 mm vs. 0.1-0.3 mm in Cribraria spinispora sp. nov. Cribraria fragilis Lado & Estrada and Cribraria zonatispora Lado, Mosquera & Beltrán-Tej. also appear on dead remains of cacti and have fruiting bodies with purplish shades, but the small dimensions of the sporotheca (diameter 0.05-0.13 mm and 0.08-0.2 mm respectively, instead of 0.1-0.3 mm in Cribraria spinispora sp. nov.), and the different size and unique shapes of the spores (Lado & al. 1999;Estrada & al. 2001), differentiate these species. Cribraria lepida Meyl. also has a purple colour of the sporotheca and has been recorded in arid lands of South America ), but it is distinguished by the larger dimensions of the sporocarps (2-5 mm in height), the peridial net with pulvinate nodes, and the smaller and almost smooth spores, (6)7-8 µm in diameter (Martin & Alexopoulos 1969).
The only known species of Cribraria described with spores with a diameter of 9-11 µm are Cribraria macrospora Nowotny & H.Neubert, Cribraria microcarpa var. megaspora Shuang L.Chen & M.Q.Guo and Cribraria stellifera Nowotny & H.Neubert, but in all these taxa the sporothecae are ochraceous, nut brown or reddish brown, not dark purple or dark violaceous as in Cribraria spinispora sp. nov. Additionally, Cribraria macrospora has larger sporocarps, 2.5-4 mm in total height, sporotheca with a diameter of 1-1.6 mm, and the peridial net is more dense, intricate, and has free ends (Neubert & al. 1993). In Cribraria microcarpa var. megaspora and Cribraria stellifera the peridial net is dense and intricate, and has numerous rounded and pulvinate nodes with several free ends (Neubert & al. 1993;Guo & Chen 2014).
Cribraria spinispora sp. nov. was collected in the field, where it fruited under natural conditions, and was isolated from moist chamber cultures.
The new species seems to be exclusive to the 'cardonal'. In all the cases, the new species was growing on dead remains of Peruvian columnar cacti, Espostoa lanata (Kunth) Britton & Rose and Weberbauerocereus weberbaueri (K.Schum. ex Vaupel) Backeb., in extremely arid lands and can be closely linked to these substrates, since it has not been observed in the rest of the many cacti analyzed in this and other studies. The collections are from two different regions in Peru and from different dates and substrates, the characters are stable and common in all the collections. This is the third small succulenticolous species of Cribraria developing only on cacti, like Cribraria zonatispora and Cribraria fragilis, suggesting that the microhabitat in these substrates may have favoured diversification in the group. Notes.-A minute species reported previously in Peru from Cajamarca and Madre de Dios . The long stalked sporocarps, the violet colour of the sporotheca and the peridial net with plates and many slender free end ( Notes.-In Peru it was previously reported only by  and  from the tropical forest of Madre de Dios. Notes.-It is the second record of this species from South America, which was previously reported from Ancash (Peru) by . In the Neotropics it is only known from Mexico, where the species was described. The distinctive characters of this specimens are the crowded and stalked sporocarps ( fig. 5a-d), the hemisphaeric, depressed sporotheca, angular from mutual pressure ( fig. 5a, b), the stalk that is calcareous, short and wide ( fig. 5d), giving the appearance of being an extension of the hypothallus, the double peridium, the outer layer thick, crustose, calcareous, smooth, and with an irregular to almost circumscissile dehiscence. The capillitium is branched and anastomosed forming a net ( fig. 5c), and the typical large spores, with a diameter of 16-20 µm, ornamented with few, prominent, sharp spines, of almost 3 µm in length. All these features agree with the original description of the species provided by Estrada-Torres & al. (2001). Notes.-In Peru it was only known from Cajamarca and Madre de Dios . Notes.-This species has only been reported previously from Argentina, Costa Rica, Ecuador and Venezuela. Farr (1976), in her monograph of Myxomycetes from the Neotropics, also reported a collection from Peru preserved at BPI herbarium, but in a search of the Fungal database of this herbarium no specimens from Peru were found. Several of our specimens agree with the commented by Farr (1976: 205) from the Venezuelan Andes, since the capillitium is elastic and the peridium, in the basal zone, where it covers the columella, is brown instead of white.

Diderma effusum (Schwein.) Morgan
Diderma hemisphaericum (Bull.) Hornem. Notes.-It is a widely distributed species throughout the Neotropics, but not reported from Peru. The taxonomic characters of our collections exhibit considerable variation and it is difficult to establish satisfactory diagnostic criteria. In the specimen C. Lado 24479 leg., for example, the columella is white very prominent and almost cylindrical, the capillitial threads are expanded in zig-zag, and the spores has a diameter of 9-10 µm, and are ornamented with prominent and dispersed spines. In other Peruvian collections, some of these features are not evident. Farr (1976), in her monograph of Myxomycetes of Flora Neotropica also pointed out the variability of the specimens examined and provides a broader description. We follow her criteria and have included all these collections under the same binomen. Notes.-In Peru this species has been reported from Arequipa, La Libertad, Lima and Tumbes , linked always to arid areas. The specimen C. Lado 22442 leg. has only five stipitate sporocarps, and does not match the description provided by Farr (1976) for specimens from the Neotropics, but it is placed in this species tentatively because of the short and black stalks of the sporocarps, the colourless peridium, and the bigger spores, with a diameter of 10-12 µm, that are pale brown and with groups of darker warts. This collection resembles Didymium clavus (Alb. & Schwein.) Rabenh. and Didymium melanospermum (Pers.) T.Macbr., but in these species the peridium is areolate or brown. Also, the spores in Didymium clavus are smaller (6-7 µm in diameter), and in Didymium melanospermum are bigger (10-14 µm in diameter). In the specimen C. Lado 23500 leg., the dehiscence is clearly circumcissile. Notes.-It is widely distributed throughout the Neotropics as well as Peru (Ancash, Ica, Lima, Madre de Dios, Moquegua and Tacna), usually linked to litter and dead leaves  Notes.-This is the first record of the species from the continental Neotropics, although previously reported from the Galapagos Islands (Ecuador) by Eliasson & Nannenga-Bremekamp (1983). Notes.-It is a widely distributed species in the Neotropics, but has not been reported previously from Peru. All the collections are uniform in therir features, the peridium has poorly marked areoles, the stalks are scantily developed, blackish but occasionally whitish, the columella is small, flattened, as a basal disc, ochraceous-whitish, and the spores are dark brown and uniformely warted, with a diameter of 10-12 µm. Macroscopically our specimens resemble the small forms of the variable species Didymium squamulosum. Notes.-It is widely distributed in the Neotropics but has not been reported previously from Peru. It was found in the upper limit of the 'cardonal' (2800 m a.s.l.), growing on the shrubs. Notes.-This species, originally described from India (Nannenga-Bremekamp & al. 1984), is widely distributed in the arid lands of Peru . Notes.-The distribution of this species in Peru is enlarged considerably with these collections, since it was limited to Piura in the north of the country , and now it is also known from La Libertad and Lima. Notes.-In Peru it was previously reported from Ancash, Arequipa, Ica and Cusco . Notes.-In Peru it was previously reported from Ancash and Piura . Notes.-In Peru it was previously reported from tropical forests of Madre de Dios ), but also from the arid lands of Lambayeque and Piura . Notes.-It is widely distributed throughout the Neotropics and usually associated with the wood of tropical forests. In Peru previously reported from Madre de Dios . In BPI there are two collections from Peru, one (BPI 838053) that is probably the specimen mentioned by Farr (1976: 96) in the monograph of Flora Neotropica, without a locality, and the second (BPI 837992) from Tingo María (Huánuco). Notes.-This species, described by Léveillé (1863), originally from Colombia as Enerthenema muscorum Lév., has also been reported from Costa Rica, Brazil, Ecuador, El Salvador, Mexico and Venezuela. This is the first record for Peru. The specimens studied have grouped sporocarps ( fig. 6a), with a black, shining and flat stalk ( fig. 6b); the peridium is membranous, very persistent, with blue or green shades, and fragmented in patches ( fig. 6a-c). The capillitial threads, radiating from the apex of the columella, are dichotomously branched, but with very few anastomoses, dark brown to blackish, uniform in diameter, only paler and attenuated at the tips ( fig. 6d), and the spores are black in mass, dark brown to dark purplish brown by TL, ornamented with prominent and irregularly distributed, blunt spines of around 1 µm long. The collections C. Lado 22486 leg. and C. Lado 22487 leg. have the primary branches of the capillitium paler ( fig. 6e, f) and resemble Lamproderma scintillans, but the rest of the characters, including the dark and spiny spores, agrees with the description provided by Farr (1976: 250-251) for Lamproderma muscorum. Notes.-In the Neotropics it has also been reported from Brazil, Mexico, Costa Rica (Rojas & al. 2015) and Paraguay (McHugh 2009). This is the first record for Peru. Notes.-Only two previous records are known from the Neotropics, in Mexico (Rojas & al. 2010) and Argentina (Lado & al. 2014  Notes.-The species, described from arid lands of Argentina , is also distributed in Brazil (Araújo & al. 2015;. Our records are the first from Peru (Lima, Tacna, La Libertad, Ancash and Huancavelica) and enlarge the distribution to the Pacific coast of South America. All the specimens examined show the distinctive characters of this species such as the dehiscence of the peridium by very well marked polygonal plates and the spiny capillitial threads  Notes.-In Peru it was previously reported from Ayacucho, Cajamarca and Madre de Dios . Notes.-In Peru it was previously known from Ancash, Arequipa, Cajamarca, Cuzco, Lima and Tacna .

Perichaena depressa
Perichaena quadrata T.Macbr. Fig. 7a-c. Notes.-In the Neotropics it is only known from Argentina, Chile, Mexico, Panama and Peru (Lado & al. , 2013(Lado & al. , 2014Lizárraga & al. 2015aLizárraga & al. , 2015b. The delimitation between this species and Perichaena depressa was very difficult in the Peruvian material, since intermediate forms were common. Under this taxon we have grouped all the collections with depressed to pulvinate or globoid sporocarps, not markedly flattened as occurs in Perichaena depressa, with small dimensions of the sporotheca (diameter 0.1-0.5 mm vs. 0.5-1.5 mm in Perichaena depressa) and the capillitial threads ( fig. 7a, b), mostly with reticulate ornamentation (very difficult to see in some cases), all these features are noted by Keller & Eliasson (1992)  Notes.-In Peru it was previously reported from the arid coastal lands of Arequipa and Lima ).
Notes.-In Peru it was previously reported from Ancash, Cajamarca, La Libertad and Piura . The spores of our collections have an ornamentation, by SEM ( fig. 7g, h), which coincides with those mentioned by  for material from the Peruvian coastal desert, but slightly differ from that illustrated by Moreno & al. (2009) of the holotype of this species, since the bacula are more dense, the apex is not widened and warty to coralloid, and the spore surface is not covered by tiny densely and regularly distributed warts. Notes.-In South America, this variety was previously reported only from Argentina (Lado & al. 2014 Notes.-In Peru it was previously reported from Loreto and Piura (Wrigley de Basanta & al. 2008a;.

DISCUSSION
The results of the survey of these habitats yielded 601 collections of myxomycetes representing 84 species in 19 genera and in all the five orders recognised in this group. These collections include one species that is new to science. This biogeographical area was virtually unexplored for myxomycetes and the results have added 26 species to the myxobiota reported from Peru. This increases the number of myxomycetes recorded in the country so far to 153 species, and extends the known distribution of many species.
In order to more effectively compare these results with the first set of data from the dry coastal areas and lomas formations , the arid belt below the area under study here, we have used the same structure of analysis as that paper. The predominant order in this survey was the Physarales T.Macbr. with 50 species, almost a 60% of the recovered taxa. This order was followed by the Trichiales T.Macbr. and the Stemonitidales T.Macbr. (fig. 8). The orders with the lowest number of species were the Cribrariales T.Macbr. and Echinosteliales G.W.Martin. The predominance of the order Physarales is even more obvious when the total number of collections is taken into account, where more than 60% of the total belonged to this order, almost 30% to the Trichiales and only 8.2% to the remaining three orders. When compared with the results from the coastal desert of Peru , the distribution of the orders and also the total numbers of species are similar ( fig. 8), and show no statistical difference (χ² 0.11, df 3, P 0.9906). There are approximately ten species from the Physarales for every seven species from the remaining orders. However at the species level there are some significant differences. The most commonly collected species were Badhamia melanospora (77 collections), Trichia agaves (29), Physarum spectabile (27), Perichaena vermicularis (24), Didymium vaccinum (22), Perichaena quadrata (22), Perichaena depressa (22), Didymium nigrisporum (20) and Physarum clavisporum (20). The three first species can be considered to be particulary abundant (> 4% relative abundance) and the most representative of this Peruvian habitat, since they make up 21.8% of all collections. On the other hand, 30% of the studied species were represented by only a single collection, an indicative percentage of a diverse assemblage.
In the coastal desert and lomas formations ) the most commonly collected species was also Badhamia melanospora (139 collections), followed by Physarum atacamense (135), Didymium nigrisporum (65), Physarum pusillum (47), Physarum clavisporum (35), Perichaena vermicularis (33), Arcyria insignis (26) and Physarum compressum (21). The only species commonly appearing in both studies were Badhamia melanospora, Perichaena vermicularis and Physarum clavisporum. This indicates a high level of diversity in both the areas since, in spite of the fact that they are neighbouring territories, many species appear to be replaced by others in the same genus. From the total of 84 species reported for the 'cardonal', 42% are common to this area and the coastal desert, but 32% appeared only in the 'cardonal' area, and 26% were only collected in the coastal desert.
The mean number of species per genus (S/G), a measure of taxonomic diversity used in other studies of myxomycetes (Stephenson & al. 1993), shows overall higher diversity the lower the value. S/G ratio in the 'cardonal' was 4.42, whereas that of the coastal lomas showed more diversity since the ratio was 3.5. This is possibly due to the presence of a greater variety of plant substrates on the coast, where endemic and coastal plants, such as Tillandsia purpurea Ruiz & Pav. or Capparis scabrida Kunth, are frequent as well as many seasonal herbaceous plants. These provided microhabitats for more different genera of myxomycetes and so increased the diversity. In the 'cardonal' there were columnar cacti such as Armatocereus matucanensis Backeb. ex A.W.Hill, Corryocactus brevistylus (K.Schum. ex Vaupel) Britton & Rose, Weberbauerocereus weberbaueri (K.Schum. ex Vaupel) Backeb., as well as succulent plants such as Agave americana L., Furcraea andina Trel. and perennial shrubs like Baccharis latifolia (Ruiz & Pav.) Pers., providing possibly more stable moist conditions but for fewer genera.
Of the 84 species identified from this survey, 51 were only collected in the field and seven were only obtained from moist chamber culture, while 26 were recovered from both. In the field collections, 71% of the specimens found were represented by 20 species, whereas in moist chamber cultures this percentage was made up of 14 species. Six additional species, abundant in field collections, were   and cardonal. rarely collected in the moist chamber cultures or not at all. Both techniques showed that Badhamia melanospora is the most dominant species. The results show that the technique of moist chamber culture in conjunction with field collecting are complementary methods and more likely to reflect the actual myxomycetes present in the area. In such arid extreme environments where field collecting may not reveal all the myxobiota, especially the more minute species, the dual methodology is essential, as has been stated previously . For instance, the order Physarales had fewer than expected collections from moist chamber cultures (z -2.39, P 0.0084), and more than expected collections of other orders (Stemonitidales, Echinosteliales and Cribrariales, z 6.39, P < 0.0001). The latter groups had fewer field collections than expected, (z -4.25, P < 0.0001), probably because the smaller species in those orders, such as Echinostelium arboreum or Echinostelium colliculosum are very difficult to detect in the field. The benefits of the dual methodology are reinforced once more by the tests of independence done on the proportion of collections and species (table 1) recovered from the field or from moist chamber culture in each survey. For the proportion of collections from different orders χ² = 53.19 (df 2; P < 0.0001), indicating that the number of collections of different groups of myxomycetes depends on whether they were collected in the field or from moist chamber culture. For the proportion of species from different orders, however, χ² = 3.28 (df 2; P 0.194), which reiterates the fact that the distribution of orders is independent of the method used to collect the specimens (field or culture).
When the number of collections from each order in each of the surveys is compared the statistical analysis shows that the vegetaion type studied does affect the relative abundance of myxomycetes in each order (χ² 53.76, df 4, P < 0.0001). The standardized residuals indicate no significant difference from the expected frequency for orders Stemonitidales, Echinosteliales and Cribrariales (P > 0.05). However the Physarales in the 'cardonal' show a lower than expected frequency (z -2.61, P 0.0045) and higher than expected in the coastal desert (table 2). The opposite occurs for the Trichiales, where the frequency in the 'cardonal' is higher than expected and lower than expected in the coastal desert. This indicates how the differences in vegetation shape the myxomycete communities within them, with decaying cacti and succulent plants favoring the development of species of the Trichiales.
The remains of cacti from 13 genera and 21 species were examined or cultured and the most productive of them were species of genera Haageocereus and Opuntia, although overall they were only the third most productive substrates with 34 collections each. However the remains of all species of cacti taken together produced 26 species of myxomycetes in 161 collections, and these made up 26.8% of the results. This is slightly more than the 23% of collections representing only 15 species recovered from all the cacti present in the coastal desert . The dead leaves and other remains of the bromeliad genus Puya were the most productive substrate, with 89 collections from 40 different myxomycete species and so more diverse as substrates than cacti. The species of this plant genus are distributed exclusively in the Neotropics, and in Peru they are a characteristic element of the western slopes of the Andes, growing mostly on steep-sloped rocky terrain ( fig. 1i-m). The arrangement, in the form of a rosette, of the leaves of these plants, channels the water,  from the scarce precipitation, towards the stem, where it progressively moistens the dead leaves that still remain attached and whose decomposition is very slow, given their coriaceous nature. This arrangement guarantees increased humidity since it prevents rapid desiccation, creating a microhabitat that is used by myxomycetes for their development. The populations of Agave, that in Peru, are usually associated with agricultural lands, produced 81 collections of myxomycetes (13.5%) in this survey with 17 species. In the coastal desert, the most productive and diverse genera of plants were Capparis L. and Tillandsia L.
The most productive locality was Plazapalapa in the administrative department of La Libertad at 7°59′ S latitude and an elevation of 1742 m a.s.l. which produced the most collections (72) and the greatest number of species (25). The productivity of this locality may be due to the greater density of vegetation associated with its more protected position in the valley and its proximity to a watercourse, that provides more moisture and, therefore, more vegetation and plant remains in decomposition for the development of myxomycetes. More than half of the collections were produced by only seven sampled localities (304 collections), located between 5º and 13° S, whereas in the coastal survey 14 localities produced a half of the collections (375). Nevertheless, some of the localities showed very unequal values over the six years of study. This may be due to the environmental conditions of each year, as well as to the physical geography of the territory and the role that the microhabitat has in the appearance of myxomycetes. The localities with 'cardonal' from the north of Peru, between the latitudes 5° and 11° S, had a greater number of collections (372) than the south (229), located between the latitudes 12° and 18° S. This may be due to proximity to the moisture of the tropical areas and the subsequent increased variety in vegetation, whereas the aridity is higher in the southern Peru due to the proximity to the Atacama Desert and the greater influence of the Humbolt stream. The elevation range where the greatest number of collections was made (196) was 2400-2700 m a.s.l., followed by 1500-1800 m a.s.l. (166), but the number of collections taking into account the number of collecting localities, was highest at the elevations between 1500-1800 m a.s.l. (33.2 collections per locality) and 2400-2700 m a.s.l. (16.33).
The comparison of the pH of the substrates reveals that the moist chamber cultures in this survey had a similar range to those of the coast ( fig. 9) and the majority of myxomycetes collected from them were from substrates around pH 6 or 7. However there were more species collected from substrates in the acid range in the coastal deserts (pH 5-7) and more in the slightly basic range in the 'cardonal' (pH 6-8). As it has been pointed out (Wrigley de Basanta 2004), some myxomycetes are very pH sensitive and some have pH optima, and the internal tissues of cacti, like those in the 'cardonal', have a basic pH at certain times in their process of decay (Mosquera & al. 2003;).
According to the estimators CHAO1 and ACE, the expected species richness if the sampling effort had been exhaustive would be 115 and 107 respectively ( fig. 10), therefore, the 84 registered species represent 73% and 79% of the estimate. Based on the Clench function, the estimated number of species, with an r 2 0.999, was 122, indicating a good fit of the data to the curve ( fig. 11), and giving a value slightly higher than that estimated using CHAO 1. According to the Clench function, this survey registered almost 70% of all possible species in the sampling area, and very similar to that of the coastal desert survey (63-69%). Considering the extremely arid conditions of the 'cardonal' area, this is a most satisfactory result. The extrapolation of the species richness, using the program EstimateS, showed that if 18 more collecting localities had been visited, under the same collecting conditions as those used in the survey, this would have returned only 15 species more.
Using the above estimators, the field collections alone, by both CHAO1 and ACE, showed that 79% and 75%, respectively, of the species theoretically possible were recorded. Using moist chamber cultures alone, both CHAO1 and ACE showed that 79% and 83% of possible species were recovered, a slightly more efficient result probably on account of the more stable favourable conditions that existed in the culture. This productivity is similar to that of the coastal area , but in the 'cardonal' 97% of the localities visited yielded identifiable myxomycetes.
To examine the similarities between the area studied and other dryland surveys in the Americas, done by the same team and using the same methods, the percentage complementarity was calculated for each (table 3).
The assemblages of myxomycetes in the different areas compared in table 3 showed fairly low similarity between the species recovered, with less than 50% of the same species collected in any two studies. From a total of 191 species reported, only 17 species (8.9%) of these were recovered in all of the areas compared here, and 104 species (54.45%) appeared in only one.
Unsurprisingly the highest number of species shared in common with those found in the 'cardonal' belt of the Peruvian Andes was for the study of the neighbouring arid coastal area of lomas in Peru, with the species assemblage of these areas the most similar of those in table 3. This was followed by the Monte desert in Argentina and Tehucán in Mexico, and the least similar to this area of Peru of those studied was Chile. Several factors may help to explain the results. The relative proximity of the study sites in the two Fig. 9. Number of collections (lines) and species (bars) of Myxomycetes recorded, at each different substrate pH in moist chamber cultures from this survey and from lomas in the coastal desert . Peruvian surveys results in an exchange and interchange of species, but differences such as the low elevation and coastal fogs of the lomas allow totally different plant substrates to grow, such as plants from the very productive genera Capparis and Tillandsia, and no columnar cacti. The Monte desert in Argentina is on the other side of the geographical barrier of the Andes mountains, but there are some corridors for exchange of species from one side to the other. However the elevation gradient was of a wider range (500-4500 m a.s.l.) in the Argentinian drylands study and the rain shadow has possibly a greater effect on the eastern slopes, and the plant species are different so there was a broader range of plant genera as substrates to account for the differences. On the other hand, the Tehuacán-Cuicatlán valley in Mexico had a similar range of elevation and has large columnar and candelabra cacti such as species of Cephalocereus, Neobuxbaumia, Stenocereus, Pachycereus, Pilosocereus and Myrtillocactus, which dominate over extensive areas ), affecting the number of common myxomycete species (46) with the present survey. However Tehuacán is in the northern hemisphere and therefore the distance for species exchange is greater. The most different assemblage was in the very dry areas of Chilean Regions I-IV which, being further   Lado & al. (2007Lado & al. ( , 2013 and  from the equator than the Peruvian deserts, are even more arid. The methodology may also have affected the results, since the Chilean surveys were from 0-4800 m a.s.l. and so, covered the range of both Peruvian studies.
In terms of species some, such as Badhamia melanospora, Didymium wildpretii, Licea succulenticola or Physarum spectabile seem to appear more readily in arid areas than others suggesting an assemblage specific to these xerophyllous conditions or to the plants that inhabit these environments. This indicates that some species are associated with specific environments or specific plants in those environments and reinforces the idea that each area has its own characteristic myxobiota as suggested previously (Wrigley de Basanta & al. 2010a, 2013. The influence of microhabitat conditions on these microorganisms and the indirect effect of macrohabitat factors have been pointed out in many studies before Wrigley de Basanta & al. 2010a, 2013 and is confirmed by these results. In particular, the analysis of the results from the two surveys of extreme xerophyllous overlapping environments in Peru, using similar methods, has revealed that each vegetational belt has some similarity in the myxomycete assemblage of species. However more surprisingly there are striking differences, especially at the species level, between the two types of vegetation with an exchange of species of more than 57%, as shown above. The differences appear to be more related to the biotic factors of each area than to abiotic factors. The climatic, orographic and edaphic factors influence the availability of host plants but it is these in turn that influence the myxomycete assemblage of the area.