GLOBAL SAMPLED RED LIST INDEX
OF THE ASCOMYCOTA
Baseline Evaluation
Guide
Selecting microfungi for the Sampled Red List
Sample size. In accordance with the IUCN Red List Index Guidelines for the Sampled Approach the sample size of the present work was 1,500 species, and the sample was not stratified.
The population to be sampled. The microfungi evaluated in this work all came from one taxonomic group: the ascomycetes (including lichen-forming and non-lichen-forming species, and anamorphic fungi with presumed ascomycete teleomorphs). The taxonomic classification used was that of the tenth edition of Ainsworth & Bisby's Dictionary of the Fungi (Kirk et al., 2008). The population to be sampled was the ascomycete subset of IndexFungorum as it existed on 3 July 2007. The total number of ascomycete names received from IndexFungorum was 202,912. This included significant quantities of synonyms. From that total number, the following categories were selected to comprise the population to be sampled: all accepted names of ascomycetes and their conidial states (i.e. one name for each species concept) [22,272 names]; basionyms for all names where the current application of the name was, on the day of sampling, uncertain [106,713 names]; all names where the current application of the name was, on the day of sampling, uncertain and the basionym was not known [7,983]. The total size of the population to be sampled was thus 136,968 names.
Making the sample. Random numbers for use in sampling came from Microsoft Basic's random number generator. Because many ascomycete names are of uncertain application, whatever the size of the sample, some selected names were expected, on closer study, to be synonyms of other names already selected. To allow for this, the initial random sample was for a larger number of names than the 1,500 required for compatibility with the IUCN Sampled Red List Index project. Because the digitizing of IndexFungorum was not a random process (for example, many lichen-forming fungi are strongly grouped within the database), every record in the population to be sampled was allocated a unique randomly generated number. The population was then sorted in numerical order by those numbers, and the process was repeated, so that the population had twice been sorted randomly. Every record in the population was then allocated a unique record number from 1 to 136,968 (the total size of the population to be sampled). 3,000 random numbers were generated, ranging from 1 to 136,968. The records with record numbers corresponding to those 3,000 random numbers were then selected. Those 3,000 records were then randomized by allocating a further unique random number to each, and they were sorted in numerical order by that unique random number. The first 1,500 names in that order were then taken, to form the ideal sample size recommended by the IUCN Sampled Red List Index project. These original names may be viewed in the unedited list.
Checking the sampled names. Those names were then scrutinized. Names not already in the Cybertruffle Nomenclature & Taxonomy Database were added. A search was carried out for missing basionyms and an attempt was made to determine the identity of names of unknown application. Where such a name was found to be a synonym of some other accepted name, that accepted name replaced the name of unknown application in the list. If this resulted in duplication within the 1,500 names, a replacement was taken from the remaining 1,500 names (the first duplicate name found being replaced by name number 1,501, the second by name 1,502 etc.). Names not effectively published, and names known only from catalogue entries, with no known place of publication were similarly replaced. The following categories received particularly close scrutiny: species known only as fossils, and species with nomenclaturally invalidly published / illegitimate names which may, however, represent genuinely different species. After these issues were resolved, the final list of 1,500 names was prepared, with a supplement showing replaced or rejected names. These 1,500 names were the species which were evaluated.
Information Sources
Manual searches. For each of those 1,500 names it was necessary to find out as much as possible about when, where and on what the species had been observed. This was challenging. First of all, because fungal nomenclature is not stable, it would be necessary to search not only under the currently accepted name of each, but also under every known synonym. On average there were about three synonyms for every name. That meant searches using about 6,000 names to be sure of getting information about the 1,500 species. In an ideal world, every fungal reference collection on the planet should have been checked for specimens or cultures of every one of those 1,500 species. More than 150 such collections are listed on the International Mycological Directory website: looking for 6,000 names in 150 collections would have meant about 900,000 searches. Every published reference in every publication should also have been checked. Taking as a single example the journal Mycotaxon, which has issued just over 100 volumes to date, that would have meant searching for 6,000 names in every index - another approximately 600,000 searches. That would then have had to be repeated for all of the other main mycological journals, and not all of them have indexes. It would then have been necessary to search for further information through the Internet: 6,000 searches of Google (which is not the only Internet search engine). Finally, a request should have also gone to all individual mycologists for any field observations of these species which they had not published. None of this was possible with the resources available for the work.
Computerized searches. The only realistic solution was to make computerized searches, and that was the route taken, not least because it meant that sequential searching using every known synonym was merely a matter of programming. Instead of looking at individual collections (and thanks to generous collaboration) records from three main on-line fungal databases (Cybertruffle, the Fungus Records Database of Britain and Ireland, and the USDA Fungal Databases) were aggregated. To this core was added information from the editorial queue of Cybertruffle (about 500,000 records which were not yet available on-line because they were still being edited). These databases and the editorial queue of Cybertruffle provide a remarkably extensive coverage because they include so many digitized records from national and regional checklists and other major mycological works dealing with individual countries, as well as records from some major fungal reference collections. In addition, the indexes of several of the main mycological journals, and of various other mycological publications were digitized using a scanner. The resulting images were then converted to text using Optical Character Recognition software, and a searchable database of the resulting records was made (that database is now openly available on Cyberliber). Finally, where available, information from the great mycological catalogues (Index of Fungi, Petrak’s Lists, Saccardo’s Sylloge, and Zahlbruckner’s Index Universalis) was added. A full list of the sources used can be found on the Information Sources page of this website. The several million records of all types of fungi which these collective sources represented was the source information on which our evaluations were based.
Applying IUCN criteria to source data
Some practical difficulties. The IUCN criteria, as they currently stand, are difficult to use with the fungi because they are incompatible with some aspects of the biology of these organisms. When evaluating randomly sampled species, this problem becomes acute, because there is a strong pressure to make some real evaluation (not just Data deficient), and in general there are very few records on which to base the decision. In the present work, as a first step, it was important not to over-estimate the number of records available for a species. For example, there might be several catalogue entries all based on the same original collection. Counting each of these as a separate record would introduce bias. To prevent this, the sum of all catalogue entries for a given species was divided by three and rounded to the nearest whole number, and that figure was used in estimating the total number of records. Another potential source of bias was detected in respect of dates. Unlike the other main database sources, information kindly supplied by the USDA Fungal Databases contained no collection dates, so that it was impossible to tell from those records when the fungus had been observed. In cases where a species was backed by large numbers of records from the USDA Fungal Databases, but not from other sources, the result was that the number of dated records was disproportionately small, and this in turn sometimes led to the species being placed in a more endangered category than was realistic. Wherever such cases were detected, a correction was made.
Additional criteria. It is doubtful whether even one of the 1,500 species evaluated had sufficient information to comment intelligently on population trends and, for almost all of them, distributional information was basic. The following additional criteria were therefore also taken into account.
| total number of records; |
| number of those records with a known date; |
| proportions of dated records before and after 1960 (a decline in dated records could, in some circumstances, be evidence of a decline in populations); |
| numbers of associated organism genera and species (this factor detected species with a narrow range of associated organisms); |
| numbers of continents and countries (this factor detected potential endemism); |
| whether or not the country was well- or poorly-explored for mycology (this subjective categorization of countries was only used when the sampled fungus was potentially an endemic); |
| whether or not the fungus was lichen-forming (a narrow range of associated organisms could be expected for lichen-forming species). |
These criteria were applied in a very cautious and conservative manner to supplement the IUCN criteria when allocating each species to a particular IUCN category. Full descriptions of how the criteria were applied to different groups are available as follows:
Results
View list. All 1500 evaluations have been made available on-line and can be accessed using the "View list" option on the menu to the left. Selecting that menu option leads to five further options: (1) the full list in alphabetic order, (2) the full list in taxonomic order, (3) the list of all species with evaluations other than Data Deficient, (4) possible endemics listed by country, and (5) species with only one known associated organism genus.
Presentation of information. For each evaluated species, there are three separate web pages. The first is the evaluation itself, set out in a format compatible with the IUCN Red List Assessment questionnaire, but with additional hyperlinks providing direct access to relevant information in Cyberliber, Cybernome and IndexFungorum. The second contains a distribution map. The third is a detailed listing of all the evidence used to arrive at the evaluation.
Breakdown of the evaluations. For most species, the amount of available information was very low indeed. Out of the 1500 species evaluated, 1071 were represented by fewer than 11 records, 366 by between 11 and 100 records, 57 by between 101 and 1000 records, 4 by between 1001 and 10,000 records, and 2 by more than 10,000 records. Most species were therefore evaluated as Data Deficient. Out of 1500 evaluations, 1423 (about 95%) were evaluated as Data Deficient. The evaluations of 1308 of these were unqualified. For the remainder, some brief comments were possible, as follows: 51 were evaluated as Data Deficient [possibly Least Concern], 43 were evaluated as Data Deficient [probably Least Concern], 3 were Data Deficient [probably Vulnerable], 6 were Data Deficient [possibly Endangered], 1 was Data Deficient [possibly Critically Endangered], 9 were Data Deficient [possibly Extinct], and 2 were Vulnerable. All the other 77 (about 5%) evaluations were Least Concern.
Possible endemics. Some species are apparently known only from one country. As part of the evaluation process, a note was made of those sampled species which were possibly endemic (those species recorded only from one country). In all, 914 of the 1500 randomly sampled species were known only from one country. These were distributed as follows: USA (110 potentially endemic species), Brazil (69), France (51), India (44), Germany (43), Italy (37), Australia (36), Japan (32), Argentina (27), South Africa (25), Czech Republic (21), Philippines (19), Russia (19), Indonesia (18), New Zealand (18), China (18), Canada (16), Austria (14), Chile (14), Portugal (14), UK (12), Sweden (11), Ukraine (10), Sri Lanka (9), Taiwan (9), Venezuela (9), Ecuador (8), Spain (8), Antarctica (7), Cuba (7), Finland (7), Mexico (7), Costa Rica (6), Hungary (6), Tanzania (6), Kazakhstan (5), Malaysia (5), Poland (5), Sierra Leone (5), former Czechoslovakia (5), Belgium (4), Colombia (4), Madagascar (4), Netherlands (4), Papua New Guinea (4), Puerto Rico (4), former USSR (4), Algeria (3), Cameroon (3), Democratic Republic of Congo [Zaïre] (3), Denmark (3), Greece (3), Norway (3), Pakistan (3), Paraguay (3), Rumania (3), Trinidad & Tobago (3), Uruguay (3), Bosnia & Herzogovina (2), Bulgaria (2), Congo (2), Egypt (2), Georgia (2), Greenland (2), Kenya (2), Luxembourg (2), Morocco (2), New Caledonia (2), Perú (2), Sao Thomé & Principe (2), Tunisia (2), Turkmenistan (2), Afghanistan (1), Angola (1), Armenia (1), Belize (1), Bermuda (1), Central African Republic (1), Dominica (1), Estonia (1), Ethiopia (1), French Guiana (1), Guinea Bissau (1), Guyana (1), Iraq (1), Korea (1), Laos (1), Libya (1), Nigeria (1), Norfolk Island (1), Panamá (1), Réunion (1), Serbia (1), Singapore (1), Suriname (1), Tadjikistan (1), Togo (1), Turkey (1), Uganda (1), Uzbekistan (1), Vanuatu (1), Yemen (1), Zambia (1), former French Equatorial Africa (1) and former Yugoslavia (1).
Species apparently with only one known associated organism genus. A note was also made of those sampled species for which only one associated organism genus was recorded (in a few cases, particularly for associations with organisms other than flowering plants, one associated organism family, order or even class was also noted). Just over 700 of the 1500 species were in these categories. Associations were very widely spread. The genera with the most fungi uniquely associated with them were: Pinus (16 uniquely associated species), Quercus (15), Homo (12), Acer (9), Fagus (8), Prunus (8), Solanum (8), Salix (7), Alnus (6) and Citrus (6). Although not specifically noted as part of this work, most lichen-forming species also, by implication, had only one associated organism: their symbiont.
Discussion
A tenfold increase in the number of fungi evaluated globally. At the time of writing, only three species of fungi are present in the IUCN Red List: one is the basidiomycete Pleurotus nebrodensis (Inzenga) Quél., the other two are the lichen-forming ascomycetes Cladonia perforata A. Evans and Erioderma pedicellatum (Hue) P.M. Jørg. In addition, 109 other fungal species have been evaluated in various individual numbers of the IMI Description Sheets of Fungi and Bacteria since 2006. Apart from these, very few if any other fungi have been evaluated at a global level using the IUCN categories and criteria. The present work has thus increased by about tenfold the number of globally evaluated species.
First published examples of fungal conservation status evaluations. The information used to make the assessments of the IUCN Red List species Cladonia perforata, Erioderma pedicellatum and Pleurotus nebrodensis, and the reasoning leading to those assessments appear not to have been published (at least not in the format of the IUCN Red List assessment questionnaire). Similarly, none of the 109 species evaluated in recent IMI Description Sheets have had the assessment information and reasoning published. They are, in every case, a simple statement that, using a defined version of the IUCN categories and criteria, the species was assessed globally at a particular level. The 1500 evaluations of the present work thus appear to be the first, or at least very early published examples of how the evaluation process is being carried out for fungi. As a result, they have value as examples which can be copied or, better, improved when subsequent evaluations of species are carried out.
Evidence of how little is known about the conservation status of fungi. About 95% of all ascomycetes evaluated here were Data Deficient. Being based on a random sample, the present evaluations enable us with some confidence to make certain observations about the conservation status of ascomycetes in general. Thus it seems reasonable to suppose that if evaluations of all other known ascomycetes were made using the same data resources, this percentage would remain roughly the same. The tenth edition of Ainsworth & Bisby's Dictionary of the Fungi (Kirk et al., 2008), traditionally an authority on these matters, estimated that a little over 97,300 species (two thirds of them ascomycetes) had been described up to the time of its publication, and acknowledged the most widely accepted estimate, of 1.5 million fungal species globally (Hawksworth, 1991), which includes an element for those not yet discovered. On the basis of those figures, only about 7% of all fungal species have so far been described. The random sample of 1500 species underwriting the present work is thus a sample only of that 7%, making it hard to avoid the conclusion that well over 90% of all fungi (described and undescribed), and possibly over 99%, cannot yet be properly evaluated for conservation because we simply do not have enough information.
An important political aspect. This exercise has therefore very clearly and objectively demonstrated how little is known about the fungi. There is an important political slant to this demonstration. To the uninitiated, the IUCN categories Not Evaluated and Data Deficient seem very similar. "If there is not enough information to make an evaluation, why bother?" The answer is that moving species from Not Evaluated to Data Deficient constitutes a shift in responsibilities. In simple terms, Not evaluated means "lazy scientists haven't done their job", while Data Deficient means "hard working scientists have tried to evaluate, but need more resources". From that point of view, the more fungi which can be moved from Not Evaluated, even to Data Deficient, the stronger will be the case when asking for support.
The quality and quantity of source information is important. It is instructive to compare the 1,500 evaluations of the present work with the 109 evaluations made in the IMI Description Sheets series. Of those 109 evaluations, 55 (about 51%) were evaluated as Data Deficient, 51 (about 46%) as Least Concern, 1 (about 1%) as Near Threatened and 2 (about 2%) as Vulnerable. Like the present results, virtually all are either Data Deficient or Least Concern. The proportions of these two categories are, however, very different. That difference is surely a reflexion of the greater amount of source information used when preparing IMI Description Sheets, which have used new and much more rigorous standards since 2006. Those standards include consultation of a range of sources significantly wider than it was possible to use for the current work, including (in addition to those used in the present work) the GBIF, New Zealand Landcare Fungi and Bacteria, and New York Botanic Garden Virtual Herbarium websites, specimens in the fungarium at the national British fungus collection in Kew, the CABI database, and the first 50 pages or all pages (whichever is fewer) found by a search of Google. Furthermore, the IMI Description Sheets are in all cases written with the involvement of an expert in the relevant fungal group. Thus, if the current work were repeated, but with sources and expert attention levels similar to those provided for the IMI Description Sheets evaluations, the proportion of evaluations resulting in Data Deficient would probably decline to about 50%, and the proportion resulting in Least Concern would probably rise to nearly 50%. That change in proportions provides an insight into how IUCN categories and criteria work.
At very low levels of information, IUCN criteria cannot be used to establish conservation status. Most IUCN categories express a view about the conservation status of the organism (Extinct, Extinct in the Wild, Critically Endangered etc.), but one (Data Deficient) merely comments on the quantity and quality of the information. Data Deficient is the only evaluation possible at very low levels of information. As more information becomes available, the first real conservation status category which can be used seems to be Least Concern: it's easiest to detect which species are common and look likely to stay that way. Substantially more information is needed, however, before all Least Concern species can be identified (that category is also used for uncommon species which have no obvious threats), and even more information is needed before the other categories can be considered. For the ascomycetes, even with access to all existing information, at best about 50% would remain Data Deficient: the amount of information needed before other categories can be identified is currently simply not available anywhere. The matrix of additional factors used in the present evaluations did not alter this general picture. At most, it enabled us to add comments like “possibly Least Concern” as suffixes to the overall Data Deficient evaluations. It is also evident that, unless new information is generated - new observations of successful and unsuccessful searches for these 1500 species - a re-evaluation in five or ten years time (such as has always been envisaged for this scheme) will simply result in more species slipping back to Data Deficient.
Potential endemics. Given their size, it is not surprising that the USA and Brazil should have high levels of apparent endemism among the sampled species. If an attempt is made, very roughly, to group countries into regions of approximately similar size, a different picture emerges. Temperate Europe has about 250 apparently endemic species, south Asia about 170, the neotropics about 130, temperate North America about 130, subsaharan Africa about 60, Australasia about 60, north Eurasia (Kazakhstan, Russia, Scandinavia) about 50, temperate South America about 50, and other parts of the world (Antarctica, northern Africa, western Asia) about 30. Since the sample was random, it’s likely that all possible endemics among known ascomycetes will be similarly distributed. It would be dangerous, however, to jump to the conclusion that this is evidence for greater endemism in Europe: that may be true but, more probably, these figures may simply indicate that more ascomycetes have been described from Europe than elsewhere. There seems to be very little published about fungal endemism. It clearly exists, but there may have been a reluctance to address the question, given the very low levels of data. Some very rough statistics about potential fungal endemics can be obtained from various websites on the Cybertruffle server, as follows:
| Brazil, 3,377 species of fungi, of which 2,047 are potential endemics = possibly 61% endemism; |
| Chile, 3,886 species of fungi, of which 1,955 are potential endemics = possibly 50% endemism; |
| Cuba, 5,538 species of fungi, of which 2,200 are potential endemics = possibly 40% endemism; |
| Dominican Republic, 2,366 species of fungi, of which 699 are potential endemics = possibly 30% endemism; |
| Georgia, 6,515 species of fungi, of which 2,598 are potential endemics = possibly 40% endemism; |
| Puerto Rico, 3,192 species of fungi, of which 789 are potential endemics = possibly 25% endemism; |
| Trinidad & Tobago, 1,647 species of fungi, of which 407 are potential endemics = possibly 25% endemism; |
| Ukraine, 6,684 species of fungi, of which 2,217 are potential endemics = possibly 33% endemism; |
| Venezuela, 3,886 species of fungi, of which 1,334 are potential endemics = possibly 34% endemism. |
These figures are reasonably compatible with the results from the current work. It is far from clear whether fungi share the same patterns of endemism as animals and plants, and there is not, at present, enough information to start making specific comments about levels of endemism in the fungi. The day when such comments will be possible can now, however, be seen on the horizon.
Associated organism genera. Much the same caution is needed when looking at information about fungi with apparently only one associated organism genus. If lichen-forming species are included, these fungi comprise well over 50% of all ascomycete species sampled. If that proportion is reflected throughout all ascomycetes, it will surely have an impact on calculations of total species numbers in the fungi. It is a pity that major biodiversity initiatives like GBIF do not provide information about associations, since that information is enormously important in the ecosystem approach to conservation. More information about the conservation status of associated organisms would be very useful. In cases where a fungus was only known in obligate association with an endangered species, the fungus could automatically be given the same conservation status as its associated organism. Although more information is clearly needed, it is telling that almost all the genera with the highest number of uniquely associated fungi are temperate forest trees. It is almost heresy to suggest that biodiversity hotspots in temperate regions can compare with tropical rainforest, but in the case of fungi, this may be the case. An insect can find its next host in a tropical rainforest by flying down, for example, a scent gradient, but fungi do not have directed flight: finding the next correct associate in an area of high plant diversity is extremely problematic for them. Is it possible that tropical rainforest encourages a high diversity of generalist fungi (able to colonize different associated organisms), but not of specialists? Perhaps temperate forests, with their extensive stands of few species are a better place to look for high diversities of specialist fungi.
Conclusions
This baseline Sampled Red List Index evaluation of ascomycetes is ground-breaking. It is the first large-scale attempt to evaluate the conservation status of fungi using IUCN categories and criteria, and may be the first where the source information and the reasoning behind the evaluations has been published. The very high proportion of species evaluated as Data Deficient is the first concrete evidence of how little is known about the conservation status of this phylum. Care is needed to ensure that the size of the problem - this level of ignorance - does not deter funding bodies from supporting further work. Fascinating information about endemism and specialist associations with other organisms is starting to emerge. The matrix of additional factors was of some use in helping to identify IUCN categories, but is likely to become valuable only when information levels are higher. If signatories of the Convention on Biological Diversity are serious about using sampled red list indexes, there needs to be funding for field studies and research into these 1,500 species.
References
| Hawksworth, D.L. (1991). The fungal dimension of biodiversity: magnitude, significance, and conservation. Mycological Research 95 (6): 641-655. |
| Kirk, P.M.; Cannon, P.F.; Minter, D.W.; Stalpers, J.A. (2008). Dictionary of the Fungi 10th edition. 771 pp. UK, Wallingford: CABI. |