I agree with Brian the comprehensive text and annotated images are really useful! I reiterate an earlier statement, your images of taxonomic detail are worth inclusion in any published document.
From: Richard Carter
Sent: Fri, August 17, 2012 7:17:49 PM
Subject: [diatom_forum] More on the Genus Tabellaria (Long!)
In working out what might have happened when I gave Charles two different identifications for the image he posted, I decided to finally tackle the genus Tabellaria head on, something I have been avoiding for a couple of years now. I've always had problems distinguishing between the taxa in the genus, so this morning I sat down with Krammer & Lange-Bertalot (and my trusty German dictionary!), and read the entire section on Tabellaria. Then I turned to two very nice slides that exhibit large populations of the genus, two species on each slide, had some long looks, and made some quick snapshots. The two slides are from Pike Bay, Lake Vermilion, Minnesota (a strew by Klaus Kemp), and Kilgore West, Spiddal, County Galway, Ireland (a splendid gift from Leszek Wolnik). I've added the snapshots to "Dick's Diatoms", and this is what I think they reveal:
1. Tabellaria flocculosa (Roth) Kützing. This taxon varies considerably in size and valve outline. Like all Tabellaria species, it has a swollen center and swollen ends. Unlike the other taxa, the central swelling tends to be broader than the ends (although they are sometimes pretty similar). Also like all Tabellaria, it has girdle bands of two types: septate and non-septate, the latter much less frequent (one per frustule?). Non-septate girdle bands are nondescript, and with no useful taxonomic characters that I know of. They are always open; i.e., there is a narrow opening in the band at one of the ends. The septate girdle bands are very useful in the identification of Tabellaria species, and should always be looked for. In T. flocculosa these bands are always closed at the non-septate end, i.e. without an opening in the "unfilled" end of the band. The end opposite to the normal septum may have a small "rudimentary septum", as illustrated in one of the new images. I have made no actual counts, but I would estimate that perhaps 10% of the septate girdle bands exhibit a rudimentary septum. In very short valves the sides between the swollen regions are simply concave, and the ends appear rather angular. In longer valves the sides gradually converge to the swollen ends, where they meet the rather drop-shaped ends. Often the sides between the swollen regions will appear slightly undulate. The rimoportula in this taxon is somewhat variable in placement: it may be right on the transapical axis, right in the middle of the swollen center, but is more frequently displaced toward the margin of the swollen center. It is small, but can be observed without too much difficulty in most cases. The mounting medium is important here; I think Zrax is preferable.
2. The slide from Kilroe West also has a smaller population of T. quadriseptata Knudson. In many ways, this one is rather easy to identify. T. flocculosa has small marginal spines on the valve face, which I find quite difficult to see (although Krammer & Lange-Bertalot say they are visible in LM); in T. quadriseptata these marginal spines are considerably longer and thicker, making them very easy to spot. The valve outline is distinct from that of T. flocculosa: the sides between the swollen regions are approximately parallel rather than convergent, without undulation or constriction below the swollen ends. The latter are quite angular-looking, with characteristic flattened ends. The rimoportula is large and easy to see, tends to look rather round, and is strongly displaced from the transapical axis, at least as far as the margin of the swollen center, and often completely outside of it. The septate girdle bands are always closed, like those of T. flocculosa, but I have seen no examples with rudimentary septa. (Krammer & Lange-Bertalot make no mention of this; Patrick & Reimer say they may be present or absent.)
3. The slide from Pike Bay has large populations of both T. flocculosa and T. fenestrata (Lyngbye) Kützing. The latter always has open septate girdle bands, which never exhibit rudimentary septa. Often the opening at the end of the girdle band "springs" quite wide, so that the girdle band resembles a hairpin. But often the opening at the end is extremely narrow, and the opposing ends may touch -- making it hard to see the narrow break. Valve outline is rather similar to that of T. flocculosa, and individual valves may be difficult to place. On the whole, though, T. fenestrata has a moderate constriction just below the ends, which makes them look more capitate than those of T. flocculosa. Undulation along the converging sides between the swollen areas is generally more apparent. The rimoportula is generally located on the transapical axis, or very close to it. It is large, transversely elongate, and very easy to see.
These taxa are slightly different in terms of their ecological preferences, too. T. flocculosa seems to be the most tolerant, and apparently has a wider distribution. T. fenestrata prefers water that is circumneutral to alkaline, while T. quadriseptata is an acidophil, with preference for water that is low in dissolved electrolytes.
A fourth species is worth mentioning, although I have seen no specimens. T. ventricosa Kützing is similar to T. flocculosa, but broader in the center, with a lower aspect ratio. The rimoportula positioning is unique within the genus: there is typically a rimoportula at either pole of the valve, not near the swollen center. In some specimens there may be two rimoportulae at one pole, with one at the other.
I hope this minilecture will assist those who wish to ID Tabellaria specimens in their own collections.