From: PO3::"frithjof@PI.NET" "F.A.S.Sterrenburg" 18-MAR-1996 16:23:32.86 To: Multiple recipients of list DIATOM-L CC: Subj: Taxonomic note GYROSIGMA SPENCERI - a summary ------ To judge by my correspondence and recent publications, it appears helpful to summarise the taxonomic and ecological status of one of the most commonly recorded Gyrosigma taxa, Gyrosigma spenceri (Quekett) Griffith et Henfrey. For the formal papers see the references. G. spenceri is on record as a species from fresh, brackish and marine habitats - an ecological range that would appear to be extraordinary in the genus. None of the previous publications, including the most recent taxonomic and ecological treatment in the flora of Krammer & Lange-Bertalot, has been based on examination and designation of type material. Critical examination of the morphological descriptions in reference works (e.g. Cleve 1894, Hustedt 1930, Krammer & Lange-Bertalot 1986) reveals only a single true difference from descriptions of Gyrosigma acuminatum (Kuetzing) Rabenhorst in the same books: in G. spenceri the transverse striae are said to be "slightly coarser" than the longitudinal, in G. acuminatum the longitudinal and transverse striae are said to be "approximately equally fine". In the dichotomic keys to Gyrosigma species in these manuals this is the criterion that determines whether the search path will lead to either spenceri or acuminatum. From the outset, it is evident that such subjective terms as "slightly coarser" and "approximately equally fine" have never been acceptable as morphological species criteria, because it cannot be determined where the dividing line lies. Authors have never specified either their technique of stria density measurement (significant errors can result from methodological shortcomings) or the size of the database leading to their conclusions. Bailey's original material (see Reimer in Patrick & Reimer 1966) of G. spenceri has been preserved and this has been designated as the type slide (Sterrenburg 1995). A type slide of G. acuminatum was designated by Schoeman & Archibald 1986. Measurement of stria density ratios is discussed in Sterrenburg 1995. The number of valves in both type slides that lend themselves to examination is small, but inclusion of valves in isotype materials yields a total of a dozen for each. (It's only because of the sterling work by curators that we have anything left at all to examine!). FINDINGS In the type material of G. acuminatum, the transverse stria density is from "approximately equally fine" as the longitudinal stria density (difference comparable to error of measurement) to "slightly coarser" (difference 10%). The latter specimens are G. spenceri according to the existing reference works. In the type material of G. spenceri, the transverse stria density is from "slightly coarser" (difference 10%) to "approximately equally fine" (comparable to error of measurement) as the longitudinal stria density. The latter specimens are G. acuminatum according to the existing reference works. W. Smith slides of G. spenceri (BM) contain G. acuminatum. Note that Smith had his "acuminatum" completely wrong (Sterrenburg 1992) - one of the very few slips of this superb diatomist. In wild populations whose specimens completely match the types of G. acuminatum or G. spenceri respectively, no difference in stria ratio is demonstrable. Neither are there other morphological differences justifying separation. The database consists of several thousand specimens from locations ranging from Scotland to New Zealand. The dominant species in the type material of G. spenceri is Aulaco- seira granulata, indicating at most a slightly brackish ecology. CONCLUSIONS The type specimens of G. spenceri wholly match the somewhat slender specimens seen in flourishing populations of G. acuminatum. The different stria ratio said to distinguish G. acuminatum and spenceri is not demonstrable in the types of these taxa. The ecology of the type material of G. spenceri matches the ecology described throughout the literature for G. acuminatum. G. spenceri is a later name for the species called G. acuminatum. WHAT NOW? "G. spenceri" is thus a synonym, to be avoided as an identification. All "varieties" named after "G. spenceri" are unacceptable as identifications. For the organism frequently referred to as "G. spenceri var. nodiferum" see Sterrenburg 1994. What the numerous records of "G. spenceri" in the literature - and in the unpublished reference databases of institutes! - may have been requires an individual examination. In some cases G. acuminatum may have been involved. HOWEVER: 1) There are indeed several freshwater to somewhat brackish Gyrosigma species whose populations show a transverse stria density that is always 10% coarser than the longitudinal stria density (typically 15% and more). There appears to be one such "non-spenceri" under Algae Images of Bowling Green University on the WWW, but I have not yet verified this. 2) Marine "G. spenceri" records are a special case. There is no evidence that G. acuminatum can survive in marine habitats. But I have seen G. acuminatum valves mixed as allochthons with brackish-marine diatoms in Baltic materials (courtesy Pauli Snoeijs). 3) Finally, Gyrosigma representatives with clearly coarser transverse striae do indeed live by the million in the marine littoral. They have been identified as spenceri but do not match its type. I'm sorry I had to zap one of our most popular Gyros, but actually the taxonomic situation gets much clearer. The original confusion of acuminatum and spenceri finds a logical explanation if we remember that people like Quekett, Kuetzing, Rabenhorst and Smith were unlikely to have access to, or even linguistic mastery of, each other's publications. How the confusion has been able to persist that long is something else again. I'll be happy to try and identify any of your "spenceris". REFERENCES Schoeman and Archibald 1986: Nova Hedwigia 43 (1-2), p. 129-157 1992: Diatom Research 7/1, p. 137-155 (no reprints left!) 1994: Proc. Acad. nat. Sc. Philad. vol 145, p. 217-236 1995: Proc. Acad. nat. Sc. Philad. vol 146, p. 467-480 ***** From: PO2::"frithjof@PI.NET" "F.A.S.Sterrenburg" 20-MAR-1996 06:52:55.84 To: Multiple recipients of list DIATOM-L CC: Subj: Re: Taxonomic note I received several messages referring to my taxonomic note on G. spenceri asking what we shall do with "G. spenceri var. curvula" as described by Reimer in Patrick & Reimer, so here's a collective reply. Because spenceri is a synonym, we may no longer derive varieties from it. All the putative vars. have therefore to be examined in the type materials to decide what they are. There are 3 theoretical possibilities: 1) if the mooted differences from "spenceri" sensu somewhat slender specimens of acuminatum are minor and inconstant we have forms of acuminatum. The diatoms should then have no more than 10% (average for the population) difference in long./transv. stria density (section Acuminati sensu Peragallo 1890/1891, defined in Sterrenburg 1995). If the diatom in question shows a distinct (average typically 15% and more) difference in long./transv. stria density (section Strigiles sensu Peragallo, defined in Sterrenburg 1995) it may be: 2) a validly described species such as G. kuetzingii, for instance, or: 3) a good species that has not yet been validly described. Stria ratio has so far been found stable in the sense that specimens in populations do not cross the boundaries between the sections of Peragallo, as defined in Sterrenburg 1995. Database: ca. 50 Gyrosigma species, thousands of valves. There is, therefore, so far a good disjunction between Attenuati, Acuminati and Strigiles although it is always possible that one may find a species which happens to straddle one of the borders naturally (say a stria difference ranging from 6% to 13%, straddling Acuminati and Strigiles). These artificial sections are excellent aids to identification, but were never intended as systematic entities. The situation as regards G. "spenceri var. curvula": - No type has been designated and deposited. A definitive conclusion is thus not possible at the moment and the following is provisional. See further! - I have seen lots of specimens that match Reimer's concept of "var. curvula". They also appear to match Grunow's description of that taxon, but see further. - Reimer's CONCEPT is completely valid: the "var. curvula" he described has distinctly coarser transverse striae and belongs to the Strigiles. It does NOT match any other described Gyro known to me. - In contrast to all previous descriptions of the past 150 years, spenceri does NOT belong to the Strigiles, however, but matches acuminatum. Reimer's NAME is not valid, therefore. The problems regarding Reimer's var. curvula are thus nomenclatural rather than taxonomic. The organism should be elevated to species rank. Let's assume it matches Grunow's organism (in the absence of a type, it does at least match Grunow's morphological data, but see further). It could then be renamed as G. curvula. This raises such fierce nomenclatural problems that I have not yet been able to find a solution. It is the most daunting nomenclatural issue I have encountered so far. Rabenhorst 1853 described Gyrosigma curvulum. According to Van Landingham this equals spenceri!! If the types would match, curvulum would be another synonym of acuminatum. But it is in NO WAY assured that Grunow indeed had the same organism in his var. curvula as Rabenhorst did in his curvulum! The quality of Rabenhorst's data is notoriously bad - especially for his time, when W. Smith's data were vastly superior. There is no designated type, but in my opinion designation of a (neo-, iso-) type would be wholly arbitrary because Rabenhorst's data are so bad that ANY small Gyro forms a match. Even worse: Ehrenberg 1838 described Navicula curvula, which according to Van Landingham also equals spenceri!! Ditto: Nav. curvula Kuetzing 1844. Do we have more helpings of acuminatum here? But again, it is in no way assured that Rabenhorst's G. curvulum indeed constitutes a generic transfer of N. curvula Ehrenberg, because the latter author's data are even worse. Ditto for Kuetzing's Nav. curvula, which may or may not be something completely else again... Typification would meet the same problems as described above. Reimer in Patrick & Reimer indicated this horrifying situation and said it had to be cleared up in future studies. Gleefully, perhaps? I'm putting this up for discussion: One might either rename "G. spenceri var. curvula" as described by Reimer as something completely different, say G. horrida, OR use the escape clause in the ICBN where usage prevails and call it G. curvula, with the express proviso that continued usage of "curvula" does not imply continuity of identity with Nav. curvula and G. curvulum. Until a solution based on designation of a type (this MUST be Grunow material and MAY NOT be Reimer's material, see further) has been formally published, the best option I can suggest is to say: "the taxon called G. spenceri var. curvula sensu Reimer". Reimer's data permit unambiguous identification, so your identification will be both verifiable and reproducible. Even if the name would change later, the organism's identity is then satisfactorily defined. Always include an informative picture of "spenceri-like" things in papers, see the horrific examples of Archibald 1984 in Bacillaria 7, p. 173. That is precisely what has NOT been done for the mass of "spenceris" in the literature. An interesting aside: both verbally and in correspondence, diatomists immediately come up with the question "how about curvula?" BUT ONLY IF THEY ARE AMERICAN. The taxon is well-represented in records from the USA, but not in any recent publication from elsewhere I am aware of (it is not mentioned in Hustedt 1930 or Krammer & Lange-Bertalot 1986, for instance). Is it much more common in the USA, or has it been much more commonly described from the USA because US diatomists are prime users of Reimer in Patrick & Reimer, which gives the only good recent description? Tongue in cheek: diatom species are where diatomists are... But the final issue in the curvula saga is ecological. Grunow (1880) mentions the North Sea littoral and saline inland waters as prime habitats, US authors describe this diatom from at most slightly brackish inland waters. I am not at all sure that we do indeed have an "indifferent?" species as Reimer suggests - COMPLETE WITH QUERY, it should be noted. If Grunow's "common" North Sea curvula is the same as I have in my samples by the million, IT IS ENTIRELY DIFFERENT from the organism identified as "var. curvula" I have seen in large numbers in samples from US colleagues (including the ANSP samples of Reimer in Patrick & Reimer 1966)... Now add to this the curious absence of Gyrosigma peisonis (or "spenceri var. peisonis") in the US data (including Reimer in Patrick & Reimer) and a waft of a delicious taxonomic/ecological can of worms is in the air! A study on kuetzingii, peisonis and some other confused species has been finished and is intended as the next installment of "The Spenceris" on the Proc. Acad. nat. Sc. Philad. channel (unsponsored, alas). I am fairly confident the curvula vs. marine "non-spenceri" thriller will find a happy end next. My prediction is that by scrapping "spenceri" we'll end up with about half a dozen easily identified and ecologically unambiguous species. This is a classic case of a bevy of simulacrum species. Examination of lots of specimens in populations from about everywhere has continued to show itself as an essential prerequisite, so if any of you have "spenceris" or whatever to share from Klondike to Namaqualand I'd be delighted. ***** From: PO2::"frithjof@PI.NET" "F.A.S.Sterrenburg" 7-MAY-1996 14:39:51.05 To: Multiple recipients of list DIATOM-L CC: Subj: Gyrosigma problems I'll gladly help with your Gyro puzzle and the very first requirement is one or more slide(s) of your stuff. This is the only way to assure ourselves of the identity of the organism we're talking about. I am glad you put your original message on the List, because it clearly shows the problems that can arise in practice and calls for some general comments. These relate to very deep issues of semantics and to the quintessence of taxonomy, which is why I also answer through the "List". Two notes to ensure that we start with the same toolkit: 1) Taxonomy is NOT the attachment of a label to specimens. Taxonomy is the search for a set of scientifically valid parameters by which an organism can be unambiguously and reproducibly separated from other organisms, as a biologically distinct entity - a species for instance. 2) The only valid standard of reference for such an entity in the field we are discussing (plants, in particular algae, and more especially diatoms) is the type. This is what the ICBN stipulates. Now first of all, my studies have NOT led to the conclusion that "G. acuminatum and G. spenceri are probably synonyms". That conclusion would be a non-event. What HAS been done is this: - the type of Frustulia acuminata (which in the course of time was named G. acuminatum) has been directly compared with the type of Navicula spenceri (which in the course of time was named G. spenceri); - this comparison demonstrates that not a single scientifically valid parameter permits unambiguous and reproducible separation of the specimens in these two type materials; - it showed that the morphological parameter traditionally claimed to separate G. acuminatum and G. spenceri ("slightly finer" vs "approximately equally fine" longitudinal and transverse striation) has never been validated; - but that at any rate the two types do NOT show any difference in this respect; - and finally, that the assemblages in the type materials show no evidence whatsoever of the ecological difference people have claimed for G. acuminatum and G. spenceri. The conclusions: - G. acuminatum is a fresh to at most slightly brackish diatom and it was validly described (as Frustulia acuminata) as a new species, as proven by examination of Kuetzing's original material - designated as the type. - G. spenceri is a superfluous and illegitimate name, because examination of the original material of Navicula spenceri, designated as the type, shows that the name relates to a diatom that had already been validly described - viz. as Frustulia acuminata. - The superfluous and illegitimate label "G. spenceri" may not be used to identify a biologically distinct entity (i.e., as a species name) because a species may not bear a superfluous and illegitimate name. It is thus incorrect to say "I need to have some more information about these TWO SPECIES". There is only ONE species involved, G. acuminatum. "G. spenceri" is a superfluous and illegitimate label, NOT a species. One of the next sentences is quoted verbatim here, and with some emphasis: ********* " Since G. spenceri has a wider tolerance ******* ********* to salt than G. acuminatum..." ******* because it turns the issue upside down. Abandon all hope, ye who enter through that particular gate, because this line of argument will preclude ANY meaningful conclusion. The statement that "G. spenceri has a wider tolerance to salt than G. acuminatum" starts by postulating as given the existence of a species called G. spenceri. As we have seen, the existence of such a species is precisely what its type material DENIES. Next, the statement infers that this G. spenceri is a biologically distinct entity because of its wider salt tolerance, for instance. As we have seen, such an ecology is precisely what the type material REFUTES. In fact, this is tantamount to claiming that GOD CREATED G. SPENCERI. This cannot be true because God has surely been so busy creating good species that no time was left for creating superfluous and illegitimate labels... What happened is that Quekett saw a nice little thing which he named after Charles Spencer, not being aware of the fact that Kuetzing had already named the same nice little thing. In the next 150 years, confusion continued so that in the course of time the label has been used for various entirely different organisms. How could one ever interpret the statement that "G. spenceri has a wider tolerance to salt than G. acuminatum"? VERSION 1: The type of G. spenceri has a wider tolerance to salt than the type of G. acuminatum. This statement is false, the type materials match in all respects, ecology included. VERSION 2: G. spenceri sensu Okuno 1974 or Jackson & Lowe 1978 has a wider tolerance to salt. This statement is false: both these G. spenceris are incompatible with the type of G. spenceri. We're now talking about organisms that correspond to G. kuetzingii's type. Salt tolerance is closely comparable to that of G. acuminatum. VERSION 3: G. spenceri sensu Hustedt 1930 has a wider tolerance to salt. The statement is false, what Hustedt had is a littoral marine species which is not even remotely like spenceri type specimens either morphologically or ecologically. The correct procedure, then: 1) do not continue to refer to unverifiable data for an undefined potpourri of several different organisms dustbinned under a superfluous and illegitimate label. 2) verify the organism's identity with reference to type material 3) verify the ecology of that species and enjoy the beauty of a solved puzzle.