Fungal infection, decline and persistence in the only obligate troglodytic Neotropical salamander

The fungal pathogen Batrachochytrium dendrobatidis (Bd) is implicated in global mass die-offs and declines in amphibians. In Mesoamerica, the Bd epidemic wave hypothesis is supported by detection of Bd in historic museum specimens collected over the last century, yet the timing and impact of the early stages of the wave remain poorly understood. Chiropterotriton magnipes, the only obligate troglodytic Neotropical salamander, was abundant in its small range in the decade following its description in 1965, but subsequently disappeared from known localities and was not seen for 34 years. Its decline is roughly coincident with that of other populations of Neotropical salamanders associated with the invasion and spread of Bd. To determine the presence and infection intensity of Bd on C. magnipes and sympatric amphibian species (which are also Bd hosts), we used a noninvasive sampling technique and qPCR assay to detect Bd on museum specimens of C. magnipes collected from 1952 to 2012, and from extant populations of C. magnipes and sympatric species of amphibians. We also tested for the presence of the recently discovered Batrachochytrium salamandivorans (Bsal), another fungal chytridiomycete pathogen of salamanders, using a similar technique specific for Bsal. We did not detect Bd in populations of C. magnipes before 1969, while Bd was detected at low to moderate prevalence just prior to and during declines. This pattern is consistent with Bd-caused epizootics followed by host declines and extirpations described in other hosts. We did not detect Bsal in any extant population of C. magnipes. We obtained one of the earliest positive records of the fungus to date in Latin America, providing additional historical evidence consistent with the Bd epidemic wave hypothesis. Genotyping results show that at least one population is currently infected with the Global Panzootic Lineage of Bd, but our genotyping of the historical positive samples was unsuccessful. The lack of large samples from some years and the difficulty in genotyping historical Bd samples illustrate some of the difficulties inherent in assigning causality to historical amphibian declines. These data also provide an important historical baseline for actions to preserve the few known remaining populations of C. magnipes.


48
Introduction 49 Chytridiomycosis, a fungal disease caused by Batrachochytrium dendrobatidis (Bd) and 50 Batrachochytrium salamandrivorans (Bsal), is recognized as a driver of amphibian population 51 declines and extirpations on a global scale, which many believe is indicative of a sixth mass 52 extinction on earth (Wake and Vredenburg 2008; Scheele et al. 2019). One particular lineage of 53 Bd, the Global Panzootic Lineage (Bd-GPL), is particularly virulent and associated with most 54 documented declines (Farrer et al. 2011). In many of the most affected areas, such as Central 55 America, South America, and Australia, Bd-GPL appears to have been historically absent (prior 56 to the 1970s) and amphibian declines followed after Bd invaded and rapidly increased in 57 prevalence and in geographical distribution (Cheng et al. 2011;Lips 2016). Despite the apparent 58 recent and rapid spread of Bd-GPL to many parts of the world, recent genomic studies indicate 238 specimens of C. magnipes sampled, we found a Bd prevalence of 2.5% (7/279) ( Table 2). The 239 earliest record of a Bd-positive animal came from a single specimen collected in 1969, 240 representing 1.2 % (1/81) of the total sampled specimens from that year. We obtained Bd-241 positive specimens in subsequent years (1970, 1975, 1981 and 1982). For the period prior to the 242 first detection of Bd (1952Bd ( -1968 and for the years with relatively large sample sizes (1964 and 243 1968), we calculated the probability of a false negative (assuming that Bd had been present at a 244 prevalence of 5%). The probability of a false negative was 0.007 for 1952-1968, 0.09 for 1964 245 and 0.12 for 1968. 246 247 Current Bd and Bsal prevalence in caves of Mexico 248 A total of 29 skin swabs of salamanders and frogs were collected from cave study sites; we did 249 not observe dead amphibians or individuals with skin lesions at any site. Chiropterotriton 250 magnipes was found in two of the 15 sites surveyed, with both extant populations representing 251 new localities; no individuals were found in sites with historical records of the species. Of the six 252 C. magnipes we found at El Coni, Hidalgo, we sampled a minimum of five different individuals 253 based on age and sex of the specimens and of the nine C. magnipes from La Trinidad, we 254 sampled a minimum of six different individuals (Table 1). We detected Bd at three of the six 255 sites where we found amphibians (Table 3)   Manuscript to be reviewed 298 Despite our near-complete sampling of available museum specimens, the small sample size for 299 all years following 1975 (resulting directly from declines in salamander abundance) make it 300 difficult to infer trends in Bd prevalence with confidence.

301
Our contemporary surveys for Bd show that the fungus is present in both sites with extant 302 C. magnipes populations, although not at especially high prevalence (Table 3). Individuals of C.  Manuscript to be reviewed 344 Bd-positive specimens, argues against this being a major problem, but it cannot be discounted 345 with the sampling methodology we used.

346
The results of our field surveys across the range of C. magnipes complement Bd detection 347 data in revealing apparent population extirpation at historical sites and low abundance in extant 348 populations. We visited some sites only once during our survey and, given the low numbers of 349 salamanders seen in extant populations and the fact that we did not always detect C. magnipes 350 where it remains present, it is possible that salamanders are still present at other historical sites 351 but went undetected. Prior to this study, however, we (and others) have made multiple visits to  Table 2). The low number of 356 salamanders detected at these two sites could either be the result of ongoing small population 357 size following declines due to epizootic Bd infection (or other causes) or a reflection that they 358 naturally occur at low abundance at these sites. We never found more than five individuals on 359 repeated visits to both sites from 2010 to the present, making it unlikely that the low number of 360 individuals detected is a sampling artifact. The lack of precise estimates of C. magnipes 361 abundance at historical sites (because not all individuals seen were necessarily collected) makes 362 a direct comparison with previous abundance of the species impossible, but it is notable that 363 neither of the contemporary population seems to reach the abundance reflected in museum 364 samples from historic sites. 365 We note that almost all known localities of C. magnipes are along roads or near towns, 366 where detection by local people or scientists is most likely; the karstic geology of the region 367 means that there are likely hundreds or thousands of other small caves within more remote 368 portions of the species' range that have not yet been searched. Furthermore, many parts of the 369 caves we did explore cannot be accessed easily (at least not without professional equipment) and 370 cave systems could be connected by subterranean passages, providing more habitat than is 371 currently known. In most cases where we did find C. magnipes, they were relatively close to the 372 mouths of caves, but we cannot discount the possibility that we were not able to detect 373 salamanders in inaccessible or deeper portions of caves.

374
Taken together, these results suggest that the invasion of Bd may have played a role in the 375 decline and apparent disappearance of C. magnipes from most known sites. Our data show that 376 Bd was most likely absent from these populations prior to 1969. After that, it likely arrived 377 several years before observed declines in salamander abundance, similar to declines in other      Map of caves where searches for C. magnipes were conducted.
Historical sites where C. magnipes was collected are shown as blue circles, sites where C.
magnipes was recorded during our surveys as blue stars, and caves where C. magnipes was not previously reported and was not detected in our survey as red squares. Locality numbers correspond to those in Table 1. Historic and new records of C. magnipes encompass entire range of the species, and no populations are known between population 13 in Hidalgo and the rest of the species' range.