292 



MARTIN BORRILL 



that features such as leaf-shape are of limited use in identification. For instance, an 

 abruptly contracted, mucronate leaf-apex is a characteristic, though inconstant, feature 

 of G. declinata. This morphological intergrading could be explained by the formation 

 of interspecific hybrids, especially if associated with hybrid swarms produced by back- 

 crossing. Church (1949) noted a large range of variation in lemma shape and length 

 in the North American G. septentrionalis (from 4 mm. long with truncate apices, to 5-5 mm. 

 with acute apices) and drew attention to the possibility of hybridisation with G. fluitans, 

 followed by back-crossing. 



An equally large range of variation in lemma length occurs in both the British 

 tetraploids (Table 1). However, any suggestion that forms of G. fluitans with short 

 lemmas are due to the introduction of genes from G. plicata following hybridisation, 

 can be ruled out, since the naturally occurring hybrid between G. fluitans and G. plicata 

 (G. X pedicellata) is very highly pollen-sterile, and the anthers are in any case non-dehiscent. 

 In addition, the existence of naturally occurring triploid hybrids between G. declinata 

 and G. plicata or G. fluitans has not been confirmed cytologically. These three species 

 would therefore appear to be free from confusion due to hybridisation. The most useful 

 morphological criteria for the identification of British species of section Glyceria are 

 shown in Table 1, from which it appears that G. declinata and G. plicata are similar 

 in many respects such as lemma and anther length. The toothing of the lemma apex 

 in G. declinata is not by itself a satisfactory character, since this intergrades into the 

 markedly three-lobed condition seen in some specimens of G. plicata. 



Epidermal Characters 



The species can be clearly separated by using the epidermal characteristics of the 

 sheath nerves towards the blade, and the type of grain, oblong-elliptic in G. declinata 

 and obovate in G. plicata. Such qualitative epidermal characters can be diagnostic for 

 taxa when they are mutually exclusive; for instance Church (1949) was able to distinguish 

 between Glyceria with smooth-walled cells, and Puccinellia and Torreyochloa with ripple- 

 walled cells. 



The structure of the epidermal cells of the nerve of the leaf-sheath is completely 

 different in each species of the G. fluitans complex. 



These cells are characteristically swollen in G. declinata and each cell, at the end 

 towards the blade, bears a very small lateral tooth. By contrast, in G. plicata many of 

 the cells bear a conspicuous lateral tooth; these render the surface distinctly scabrid. 

 These types are quite distinct in appearance, and in G. plicata the sheaths are rough 

 to the touch, the teeth being visible under a X 10 hand lens. G. fluitans has sheaths with 

 rounded smooth cells. The sheaths of both G. declinata and G. fluitans are smooth 

 to the touch, and a low-power binocular microscope of about X 50 is required in order to 

 distinguish clearly between these types. 



Pollen 



Pollen diameter is sometimes useful in discriminating between species with other- 

 wise similar phenotypes (Goodspeed and Bradley, 1942). Measurements of pollen size 

 and counts of good pollen were therefore made using a modification of the cotton- blue 

 staining technique described by Stebbins & Valencia (1946). Only darkly stained grains 

 were counted as fertile. 



Three mature, undehisced anthers were taken from a basal floret of each gathering. 

 In order to avoid selecting particular grains all those present in three eyepiece-fields were 

 measured. No data were obtained for pollen diameter in G. X pedicillata since this pollen 

 is mostly shrunken and irregular in shape (See Plate 19). 



