5. COMPARISON WITH EARLIER INVESTIGATIONS
When comparing the constant species that occur in both the Dutch lake district of NW‑Overijssel (Coesel, 1981) and in the area investigated by the present author, only part of the 'desmid‑ communities' appear to be related. Only the eu‑mesotrophic communities of this investigation can be compared with the Dutch material. Cluster B‑5 of Coesel shows most agreement with our cluster 3‑C (26 species are in common, 18 are not). . B‑3 is related to cluster 4‑C (23 species are in common, 14 are not). Of the A‑series of Coesel, A‑3 shows some agreement with, again, cluster 3‑C (13 species are in common, 18 are not). B‑5 and B‑3 represent submerged fen‑localities. pH‑values of B‑5 are about the same as in cluster 3‑C, however conductivities are much higher. pH values of B‑3 are about the same as in cluster 4‑C and again conductivities are much higher. The degree of trophy of B‑5 is higher than that of B‑3. Most other communities distinguished by Coesel represent habitats with a higher degree of trophy, which are not found in the Andmyran‑resrve. A number of the constant species in the Dutch B‑5 and B‑3 are absent on Andøya and vice versa. Dutch species are Closterium moniliferum, Cosmarium hornavanense, C.rectangulare, C.crenulatum, C.subreinschii, C.quadratulum, C.goniodes, C.paragranatoides, Euastrum bipapillatum, Micrasterias crux‑melitensis. Norwegian species of cluster 3‑C and 4‑C are Closterium striolatum, Cosmarium subundulatum, C. pseudopyramidatum, C. pyramidatum, C. rectangulare var. croasdaleae, C. perforatum, Euastrum boldtii, Staurastrum teliferum . In general the constant species in common seem to indicate the same degree of trophy.
In the Belgian Ardennes, Symoens (1957) discerns, amongst others, a Penium silvae‑nigrae ‑Cosmarium obliquum assembly. This semiathmophytic community can be compared with our cluster 1‑A, which has a comparable ecology. All Belgian species occur in cluster 1‑A, which, however, is more species‑rich. Symoens (l.c.) Micrasterias truncata‑Vanheurckia crassinerva assembly is related to our cluster 3‑C. This assembly shows a saprobic character and is favoured by drought and bovine manure. Also in cluster 3‑C drought is an important factor. Nutrient contents in this cluster are fluctuating.
Peterfi (1974) investigated desmids in the Romanian Carpathians. His oligotrophic species group shows agreement with our cluster 1‑ B, and his mesotrophic species group is most related to cluster 3‑ C. The Romanian species that occur in both ha~ bitats, show the same trophic preferences in the Andmyran reserve.
A comparison with the study of Flensburg & Malmer (1970) in South Sweden, shows agreement between their ('richest') Sphagnum plumulosum‑group and our eu‑mesotrophic cluster 3‑C. The ('rich') Menyanthes trifoliata‑group is most related to the mesotrophic cluster 4‑B, the ('poorer') Narthecium ossifragum‑ and Eriophorum angustifolium groups can be compared with the oligo‑mesotrophic cluster 2‑A, but also with the mesotrophic cluster 3‑B. Finally, the ('poorest') Eriophorum vaginatum‑group is most related to the oligotrophic cluster 1‑B.
6. CONLUSIONS AND DISCUSSION
Although the algal and chemical samples are snapshots in time and space, it was possible to discern clusters with significant ecological meaning. Although regional differences play an important role, many communities discerned can be compared with communities in a large geographical area and therefore these communities have a more general value. The ecology of the groups and clusters is summarized in table 3. The major factors which determined the composition of the desmid‑communities appeared to be the rich‑poor gradient and the wet‑dry gradient, which are also the main factors for the macrophytic vegetation. The species‑richest communities develop in the poor‑mesotrophic to rich‑mesotrophic habitats. A permanent presence of a waterbody of sufficient volume increases the number of constant species. Species‑poorness occurs at both ends of the rich‑poor gradient and can be intensified by drought. The trophic levels distinguished in the 'desmid‑vegetations' run parallel with the trophic levels of the macrophytic vegetations, though not absolutely (table 4). One trophic level of the microphytic vegetation can cover different macrophytic levels and vice versa. One reason may be the more 'statistic composition' of the microphytic communities, and another reason for this could be the more rapid response of desmids to changes in the environment. A refinement of the mesotrophic level distuinguished in the macrophytic vegetation was necessary to describe the differences between certain desmid‑clusters. These refined mesotrophic levels cannot be seen in the macrophytic vegetation.
This investigation, like the investigation of Buys (1986), affirms the value of the Andmyran, part of which is protected now. With its large variety of vegetation types, especially the knowledge of mesotrophic desmid communities could be extended.