a Ne ee

Botanical Society of the British Isles

Patron: Her Majesty Queen Elizabeth the Queen Mother

Applications for membership should be addressed to the Hon. General Secretary,
c/o Department of Botany, The Natural History Museum, Cromwell Road, London,

SW7 5BD, from whom copies of the Society’s Prospectus may be obtained.

Officers for 1996-97

President, Mr D. A. Pearman

President-elect, Mrs M. Briggs

Vice-Presidents, Mr A. O. Chater, Mr R. G. Ellis, Dr R. J. Gornall, Mr C. D. Preston
Honorary General Secretary, Mr R. G. Ellis

Honorary Treasurer, Mr M. Walpole

Editors of Watsonia

Papers and Notes, J. R. Edmondson, R. R. Mill, E. C. Nelson, B. S. Rushton*
Plant Records, C. D. Preston

Book Reviews, C. D. Preston

Obituaries, B. S. Rushton

“Receiving editor, to whom all MSS should be sent (see inside back cover).

© 1997 Botanical Society of the British Isles
The Society takes no responsibility for the views expressed by authors of Papers,
Notes, Book Reviews or Obituaries.

Watsonia 21: 231-251 (1997) 225

Presidential Address, 1996

DAVID PEARMAN SMITHSON Ay

. WK 25 1997}
~ LIBRARIES

TOWARDS A NEW DEFINITION OF RARE AND SCARCE PLANTS

This address will review progress in recording over the last 40 years, with particular reference to
Rare (found in not more than 15 10-km squares) and Scarce Plants (found in 16-100 10-km squares),
and will suggest that the old definitions for these categories, whilst they might have been perfectly
adequate at the time, are no longer the best we can do. They should be replaced by more precise
measures of distribution and to that end I will examine the merits of tetrads (2 x 2 km squares) and
1-km squares, touch on sites and populations, and suggest what I think is the best of the alternatives.

As you all know recording by 10-km squares really took off with the Atlas of the British Flora
(Perring & Walters 1962), for which fieldwork began in the early 1950s. Before this, maps showing
distribution by vice-counties had been used, but it was the invention of the National Grid, and its
general appearance on maps after the Second World War that gave the necessary impetus.

The first attempt to cover ‘“‘Rare Plants” in Britain appeared in 1977 (Perring & Farrell 1977, 2nd
ed. 1983), using the quite arbitrary (but perfectly acceptable in terms of a percentage of our native
flora) definition of Rare as being in 15 10-km squares or fewer. Since then, the then Nature
Conservancy Council (NCC) enshrined this definition in its Sites of Special Scientific Interest
guidelines (Nature Conservancy Council 1989), and that for the less rare, now called Scarce, as
being in 16-100 10-km squares. This last was equally arbitrary but, again, logical in that it covered
another 20% of the native plants, and 100 squares is a nice round figure, and so on.

It is now 1996; 10-km grid squares and recording have been around for 40 years. The 1962 Atlas
has spawned a mass of county Floras on a2 X 2 kmsquare and 5 Xx 5 km square basis, and a few area
Floras have used 1 x 1 km squares. A mass of more detailed information is now available — on a2 X
2 km (tetrad), 1 x 1 km (4 figure grid reference) and even, dare I say, on a site basis (6 figure grid
reference), yet we continue to display, categorise and judge Britain’s Rare and Scarce plants on the
coarse basis of 10 x 10 km squares. On the other hand this is the appropriate approach for an overall
view of our whole flora, as in our own new Atlas 2000. Here we need coverage, for the whole of the
British Isles, of over 3000 species. We also need to know which may be increasing or declining,
because even after the Scarce Species project we still have no idea what is happening to species that
may be declining but which fell beyond its parameters. The Monitoring Scheme (Rich & Woodruff
1990), undertaken during 1987-8, has offered plenty of thoughts here. Also, since BSBI’s
pioneering 1962 Atlas many other groups, from birds to woodlice, have been covered at the 10-km
range and thus we need a modern 10-km square Atlas as a baseline. But Iam not remotely content
to rest the Red data book and Scarce plants on this basis and I wish to expand on this.

The Scarce plants atlas (Stewart et al. 1994) covered 325 species. With not too much effort (mainly
trying to ensure that the grid references would be fine enough where we had site details) a further 73
maps were produced showing the number of tetrads in which particular species were recorded
within each 10-km square. Of course there are anomalies and imperfections here, in that:—

® some upland species may not be well enough recorded (but I think it probably works now for
England and much of Wales and lowland Scotland);

@ a tetrad may contain many records or just one;

@ only numerals up to 9 (out of 25 in each 10-km square) fit elegantly on to a map. Extra numerals
had to be shown by notes, although it would be possible to use more symbols;

e@ and, of course, as you will hear, the data we have are much less than adequate in terms of
accurate grid references;

226 D. PEARMAN

FiGurE 1. The distribution of Myosotis stolonifera in Britain at 10-km scale (left hand page) and tetrad
distribution within 10-km squares (right hand page) (from Stewart, Pearman & Preston 1994).

227

PRESIDENTIAL ADDRESS

228 D. PEARMAN

Ficure 2. The distribution of Asplenium septentrionale in Britain at 10-km scale (left hand page) and tetrad
distribution within 10-km squares (right hand page) (from Stewart, Pearman & Preston 1994).

229

PRESIDENTIAL ADDRESS

230 D. PEARMAN

but, as any map is a simplification, the extra information that a tetrad map shows definitely
outweighs the effort and the caveats. The centres of distribution begin to emerge and the vulnerable
areas can be identified.

I never thought I would be enthusing over tetrads, since I feel in County Flora terms and
conservation terms that they are a disaster (but no side-tracking at this point) — but I think that
tetrad maps for rarer species are a major step forward (and a use for all that information carefully
gathered and never used). Of course, they are only a first step forward and for the rarest species
much finer scales should be aimed for.

Two examples of 10-km square and tetrad distribution maps are given for Myosotis stolonifera*
(Fig. 1) and Asplenium septentrionale (Fig. 2). The tetrad map gives a far more relevant picture and
it can be seen that Asplenium septentrionale could be a very rare plant.

Examples of the coarse picture shown by 10-km squares can be produced from looking at every
County ‘‘tetrad” Flora. In Dorset, I spent eight years looking for sedges, and mapping them on a 1-
km square basis, and also searching for good examples of habitats for the “Sites of Nature
Conservation Interest’? scheme run by the Doryset Wildlife Trust and Dorset Environmental
Records Centre. The map at the back of Sedges and their allies in Dorset (Pearman 1994) shows the 1
km square coverage and demonstrates that I went to the majority of 1-km squares in the county that
were not wall-to-wall arable or towns. As one of the handful produced on a 1-km square basis, I
consider the coverage to be good, as it allows meaningful comparisons at the three scales, and shows
what is hidden by using a 10-km square scale (Table 1).

TABLE 1. SEDGES IN DORSET - THE NUMBER OF SQUARES RECORDED SINCE 1980 AT
DIFFERENT SCALES AND FREQUENCY RATIO

10-km Frequency ratio
squares Tetrads 1-km squares _(tetrads/10-km squares)
Carex acutiformis 30 ileyil ZA2 5-03
C. divulsa ah 86 107 2-77
C. riparia 30 106 149 3-95
C. remota 34 259 460 7-62

Thus tetrads and 1-km squares can give a totally different picture from that shown by a 10-km
square atlas. When I showed Arthur Chater my first draft of this address he criticised me for
choosing tetrads when he assumed much finer information was available. This is definitely not the
case. There are still quite a few records in the Red data book (RDB) database without six figure grid
references, and a substantial proportion of the Scarce plants database, not just for those species that
might be common locally, like Phyteuma orbiculare on the Sussex downs or alpine species such as
Carex saxatilis, but for whole counties that only collect records on a tetrad basis. An area of 2 X 2km
is far too coarse for rare and scarce species. Of course I appreciate that tetrad or 1-km square
recording may be satisfied with one site or population per tetrad or 1-km square (even with a six
figure grid reference) and ignore all the others.

Just out of interest, and to demonstrate the scale of the problem, Table 2 indicates the numbers of

TABLE 2. ACCURACY OF POST-1970 RECORDS IN THE SCARCE PLANTS DATABASE
(BRC, MONKS WOOD)

Scale of recording No. of records
6 fig. grid reference (i.e. 100 m square) 26000
4 fig. grid reference (i.e 1-km square) 11500
Tetrad only 5500
10-km square only 5500
Total 48500

*Nomenclature follows Stace (1991).

PRESIDENTIAL ADDRESS 231

post-1970 records of different degrees of accuracy, held on the Scarce plants database at the
Biological Records Centre, Monks Wood. 46% do not have a six figure grid reference, and this is
after 40 years of recording!

The different roles of Recording and Mapping should be made clear.

Recording (and monitoring) of rare and scarce plants should always be on as fine a basis as is
practical, at least to a six figure grid reference, i.e. to within 100 metres, and attempts should be
made to define a site. This is a problem for many reasons — the spreading through rhizomes or

SORBUS LANCASTRIENSIS

FicurE 3. The distribution of Sorbus lancastrienis in Lancashire: upper left — site distribution; upper right— 1-km

square distribution; bottom left — tetrad distribution; and bottom right — 10-km square distribution (prepared by
hGuGe Rich):

232 D. PEARMAN

PRESIDENTIAL ADDRESS 233

PEAT a (ate ee ;

(
\

aes

Ficure 4. The distribution of Lotus angustissimus in Cornwall and Devon: facing page, upper — 1-km square
distribution; facing page, bottom — tetrad distribution; and this page — 10-km square distribution. O pre-1970;
@ 1970-1995. (Prepared by M. Wigginton, JNCC.)

stolons of so many perennials, the fragmentation of a locally common species over an area, and the
spread of the site over several 1-km squares. Progress needs to be made in settling some arbitrary
definitions, and perhaps “moveable” 1-km squares might help. It can be seen from the map of
Sorbus lancastriensis (Fig. 3), that defining a site is not an easy task, although perhaps easier with a
tree species.

For Mapping there is inevitably a simplification, and from the map of S. lancastriensis, I think 1-
km squares or tetrads are quite satisfactory. It is feasible to map most of the current RDB plants at a
1-km square scale, but only some of the Scarce plants given the caveats in Table 2. At the site scale,
it might only be possible to prepare maps of sites for a handful of RDB plants and I doubt whether
any Scarce plants could yet be done.

Two examples of maps at the different scales have been prepared, Sorbus lancastriensis (Fig. 3)
and Lotus angustissimus (Fig. 4). In fact for Lotus angustissimus one might go further and say that in
the years 1993 and 1994:

@ three 1-km squares for Pentire have seven sites and 200-300 plants;
@ 14 1-km squares for South Cornwall have 15 sites and 600 plants;
@ seven 1-km squares from Prawle to Start Point have 13 sites and 6500 plants;

and all the other 1-km squares have few sites and tiny populations. I am not certain how you map
this, and how you take into account the effects of burning and clearing, and good and bad seasons.

Again Simon Leach would argue that the extra effort in looking for every site of Lotus
angustissimus has put it over the national 10-km square RDB threshold, and possibly over what Iam
going to suggest as my tetrad threshold too, although it depends on which cut-off date you use! A
difficult problem, but as I shall say later, one that we are close (well, fairly close) to overcoming.

These two sets of maps (Figs 3 & 4) show a valuable picture and I hope my successor in 2000 will

234 D. PEARMAN

be standing here similarly demonstrating that Asplenium septentrionale really is restricted to only 40
sites in ‘x’ 1-km squares and ‘y’ 2-km squares whereas, say, Carex rariflora, little known to many of
us, but an absolute gem to the aficionado, is wall-to-wall in the east Grampian mountains with huge
sites with many thousands of plants over its restricted area. For the moment, then, tetrad maps are
as fine as it is possible to go to obtain a meaningful comparison between species.

Chris Preston and the Biological Records Centre have kindly produced tetrad totals for all the
Scarce species, and Martin Wigginton has done the same for all the RDB species. Although the
details are available I have restricted the tables to only the rarest scarce and the commonest RDB
species. All figures have been adjusted to post-1970 records and figures, and only the lowest tetrad
numbers for Scarce species, and the highest numbers for RDB species are shown. Those from the
RDB are inevitably slightly better recorded and the commoner Scarce species may not have been so
conscientiously recorded. But the point I am trying to make is that the tetrads must be a more
effective way of defining a rare plant, and the order of frequency is substantially different from that
shown on a 10-km square sequence. The table of Scarce species found in low numbers of tetrads
(Table 3) shows that many really are infrequent in each 10-km square, with frequency ratios ranging
from 1-20 to 2-50.

On the other hand the table of Scarce species found in high numbers of tetrads (Table 4)
demonstrates that some species that are found in low numbers of 10-km squares are much more
frequent in tetrad terms.

In comparison, however, when one converts RDB species into tetrads rather than 10-km squares
(Table 5) it will be seen that quite a few are more common than many of the so-called ‘Scarce’
species.

It is relevant at this stage to consider the 73 Scarce species which were also mapped as tetrads (Fig.
5). Some species have been omitted in the lower left section of the graph for clarity only. I have
looked at the effect of under-recording, or not fine enough recording, on all species in the tables and
this graph, and adjusted where I can. I do not think that lack of tetrad recording in upland areas will
do other than slightly blur the picture, because these are better-recorded species.

At this point one could bring together the strands illustrated to date, and say:

@ a small number of 10-km squares but high tetrad numbers indicates a local or restricted
distribution, but not uncommon where it occurs, e.g. Carex humilis found in 28 10-km squares
and 124 tetrads; whereas

@ a small number of 10-km squares, but low tetrad numbers indicates scattered, possibly
widespread even, but now with isolated sites and populations, e.g. Asplenium septentrionale
and Vulpia unilateralis. These are the vulnerable species that we should be really concerned
about.

Potentially interesting are those in quite a few 10-km squares, but with low tetrad numbers that
one might not have thought of as vulnerable, e.g. Silene gallica, Thelypteris palustris and Pilularia
globulifera.

There has been much work done, particularly in the bird world, on these ratios of tetrads to 10-km
squares, and the resulting ratio has been usefully called a “frequency index”’.

Le Duc, Hill & Sparkes (1992) use this concept and the New atlas of breeding birds in Britain and
Ireland (Gibbons et al. 1993) makes great use of ‘‘frequency”’ maps (p. 457 et seq.). The “‘frequency
index”’ could be a valuable conservation tool, provided, of course, that the basic data are correct, or
even that the basic data are available! I asked David Gibbons, one of the joint editors of the Bird
Atlas, as to whether they had used low frequency ratios as a conservation tool. He replied that there
was no need to, as they had population counts of all the rare and important species — totally different
from the botanical experience. But for the common species, we can already use Monitoring Scheme
data to plot similar maps and hope to use this technique in the new Atlas 2000.

I do not know whether it is relevant that for these 73 Scarce Species (Fig. 5), the average number
of tetrads per 10-km square was 2-2 — i.e. there were records from only 2-2 out of a possible 25
tetrads for these species. I repeated the exercise for the 30 most common (in tetrad terms) of the
RDB species, and found, to my surprise, that the average was 2:16. When all the Scarce species were
included the average was 1-95, but all the caveats about locally common, coastal and upland species
need to be taken into account.

PRESIDENTIAL ADDRESS 235

TABLE 3. “SCARCE” SPECIES FOUND IN LOW NUMBERS OF TETRADS - 1970 ONWARDS

TOTALS
10-km Frequency ratio
Tetrads squares (tetrads/10-km squares) Comments

Corynephorus canescens 17 12 1-41

Veronica spicata 20 16 1:25

Carex vulpina 22 12 1-83

Luzula arcuata 22 12 1-83

Calamagrostis stricta 22 15 1-46

Galium pumilum 23 Ue) 1-21

Cyperus longus 23 20 115 Picture confused by planting

Vulpia unilateralis 24 20 1-20

Cystopteris montana 25 15 1-67

Melampyrum sylvaticum 26 21 1-23

Alchemilla glomerulans 27 24 1-12 Certainly under-recorded

Pulsatilla vulgaris 28 19 1-47

Galium parisiense 29 19 153

Mentha pulegium 29 20 1-45

Ribes alpinum je) 20 1-45

Lathyrus palustris 29 22 1-32

Chenopodium chenopodiodes 30 12 2:50 Probably in at least 32 tetrads

Spiranthes romanzoffiana 30 18 1-66

Phleum alpinum 30 21 1-43

Circaea alpina 30 24 1-25

Ophioglossum azoricum Si 24 1-25 Probably in >13 more 10-km
squares and >15 tetrads

Asplenium septentrionale 31 a7, F-5

Najas flexilis 31 15 2-07

Illecebrum verticillatum 32 16 2:00

Elatine hydropiper 32 19 1-68

Sagina saginoides 32 20 1-60

Alopecurus borealis a2 24 1-33

Crepis mollis a2 28 1-14 Probably in at least 43 tetrads
giving ratio of 1-54

Carex appropinquata 32 Zi 1-52

Marrubium vulgare 32 19 1-68

Ranunculus tripatitus 33 19 L73

Juncus filiformis 33 20 1-65

Atriplex longipes 34 21 1-26

Daphne mezereum 34 24 1-41

Ajuga chamaepitys 35 20 1-75

Salix reticulata 35 ig 2-06

Allium schoenoprasum 36 17 212 Probably in at least 39 tetrads
giving ratio of 2-29

Sorbus porrigentiformis 36 21 1-71

Linum perenne 36 24 1-50

Dianthus armeria 38 36 1-06 Probably in 33 tetrads, 31 10-km
squares, giving ratio of 1-06

Silene conica 39 22 1-77

Adiantum capillus-veneris os) 26 1-50

Frankenia laevis 39 28 39

Juncus biglumis 40 23 1-74

Draba norvegica 40 24 1-67

Polypogon monspeliensis 41 21 1-95

Melampyrum cristatum 4] 23 1-78

Linnaea borealis 4] 32 1-28

Carex ericetorum 42 26 1-61

Actaea spicata 42 22 £90

236

Arabis glabra

Salix arbuscula
Veronica alpina
Juncus castaneus
Sorbus devoniensis
Helianthemum canum
Carex maritima

Nuphar pumila
Juncus balticus
Juncus alpinus
Trifolium occidentale
Peucedanum palustre
Minuartia sedoides
Impatiens noli-tangere
Cerastium arcticum
Vicia bithynica
Sonchus palustris
Fallopia dumetorum
Cerastium cerastioides

Tetrads

D. PEARMAN

TABLE 3. (continued)

10-km
squares

Frequency ratio
(tetrads/10-km squares)

35
1-82

Comments

15 tetrads are 1970-1979 only

Probably in at least 47 tetrads,"
giving ratio of 1-15

TABLE 4. A SELECTION OF “SCARCE” SPECIES FOUND IN HIGH NUMBERS OF TETRADS AND
LOW NUMBERS OF 10-KM SQUARES - 1970 ONWARDS TOTALS

Cardamine bulbifera
Carex digitata
Carex humilis

Dryopteris submontana

Gentianella germanica
Orchis purpurea

Ornithogalum pyrenaicum

Primula elatior
Pulmonaria longifolia
Wolffia arrhiza

Tetrads

10-km
squares

Frequency ratio
(tetrads/10-km squares)

4-26
2-88
4-43
Salts)
3-10
2°85
4-13
3-41
4-15
3-28

At this stage in 1996, I do not think there is sufficient information to draw conclusions from these
ratios, but I really do believe it is a subject for future exploration. To me it seems extraordinary low.
I would have thought we could acquire 1-km square data fairly easily for the rarer species, and then
move to six figure grid references. But why are the figures so consistently low?

These figures might be put into context by looking at any County “‘tetrad”’ Flora to see what sort
of frequencies one finds by including all plants. Perhaps Daisy and Dandelion and Nettle would tend
towards the maximum of 25 and no doubt Jack Oliver from Wiltshire has done some work on this,

PRESIDENTIAL ADDRESS 237

TABLE 5. RED DATA BOOK SPECIES FOUND IN HIGH NUMBERS OF TETRADS -— 1970
ONWARDS TOTALS

10-km Frequency ratio
Tetrads squares (tetrads/10-km squares)
Fumaria occidentalis 84 27 3-11
Scrophularia scorodonia ll 29 2-66
Gastridium ventricosum 52 26 2-00
Erica ciliaris 49 15 3-27
Poa infirma 44 26 1-69
Genista pilosa 4] 13 3-15
Ophrys sphegodes 36 14 PST
Salvia pratensis 36 24 oe 50)
Cirsium tuberosum 35 14 2:50
Lotus angustissimus 35 ZS 1-40
Phleum phleoides 35 12 eID
Potamogeton nodosus a3 11 3-00
Bunium bulbocastanum 32 12 2°67
Carex rariflora 32 15 218
Gentiana verna 30 4 7:50
Lithospermum purpurocaeruleum 30 15 2-00
Muscari neglectum 30 23 1-30
Silene otites 30 8 a= 75
Carex ornithopoda 28 11 2-95
Cynodon dactylon 28 Zl 1-33
Bartsia alpina 26 13 2-00
Erica vagans 26 5 5-20
Euphrasia vigursii 26 15 t-73
Kobresia simpliciuscula a) 12 : 2-08
Orobanche purpurea pis) 17 1-47
Physospermum cornubiense v2) 9 218
ib
Carex hum
Carda bul
Pulmo lon
45 Prim far
Strat alo Clino cal
Oo Wolff arr as
Ss enti @voP sub Y
o 3 4 ie ig aeen mar Minua ver |
oO Helia can peep i
Cc ; Hyper und
2 pe ae tba pul Gentipna. Poten neu Horde eur
(on Illec ver — Rhync fys Alope bul
O25 ene Oenan sil Melit mel
Ee Paty eines cha moe Seco Cicut vir Viola lac Pilul glo
Pulsa chem mot bituren nae Sil ae co
“hae ea bor 3 Dacty tra ksplevs Thely pal
Asple sep |
1 |
Figure 5. The relationship
between the frequency ratio
(number of tetrads per 10-
0 : . km square) and number of
0 50 40 60 30 400 10-km squares for Scarce

Species. (Based on post-
Number of 10km squares se ie .

238 D. PEARMAN

but I would like to know what is the average. Because it is a small manageable suite I have produced
a chart for Dorset Cyperaceae (Fig. 6), with the only caveat that no sedge is a really ubiquitous
plant. The average frequency ratio here is 4-16 tetrads/10-km square. Interestingly enough, on a 1-
km square: 10-km square basis the average only increases to 4-45. It would be useful to know if there
was a predictable relationship between tetrads and 1-km square frequency ratios.

20 = Lee

15

No. of species
o

123 45 67 8 9 1011 12 13 14 15
Frequency ratio

Figure 6. The number of species occurring in each band of the frequency ratios (tetrads per 10-km square) for
Cyperaceae in Dorset.

Looking then at the tables (Tables 3, 4 & 5), there are some very interesting observations to be
made. Many of the rarer Scarce species are really uncommon in areas they occur — a frequency ratio
of under 1-5 —i.e. found in less than 1-5 tetrads (out of a possible 25) in each 10-km square, less than
6%. Most of the less rare RDB are much commoner — look at Erica ciliaris, Potamogeton nodosus
and especially at Gentiana verna. Apart from the fact that it is very pretty and photogenic, why are
we even considering putting conservation resources into this plant, which is six times more common
where it occurs than Asplenium septentrionale? The further argument that they are both widespread

PRESIDENTIAL ADDRESS 239

and frequent in Europe, is outside the scope of this address, but is something that JNCC, by
adopting the new IUCN guidelines on Red data book plants, is beginning to address. I suppose
cynics would quite fairly argue that “public appeal” must be brought into the equation, and that is
why birds and dormice always win and Gentiana verna will always beat Asplenium septentrionale!
But that is for the politicians in JNCC and the three country agencies, not for me!

Table 6 shows Scarce plants, which have been reliably recorded, that have the lowest frequency
ratios. I would be very happy to use conservation resources on these species. Cyperus longus was on
this list, but I have left it out as itis now much more common as a garden escape than a native. From
personal experience they are locally rare, they are certainly scattered and they are more worthy in
my mind of conservation effort than many RDB species. I appreciate that this begs another
question. I am being simplistic in saying that *x’ species should be protected and ‘y’ species should
not. The reality might be that all are protected in the parts of their range where they are most
vulnerable or looked after in their core areas because they are core areas — another reason for
acquiring more precise data. Some of these, particularly those lower in the list, may well be under-
recorded, but again, there are some very interesting observations to be made about species

TABLE 6. “SCARCE” SPECIES WITH LOW FREQUENCY RATIOS - 1970 ONWARDS TOTALS
(based on Stewart, Pearman & Preston 1994)

10-km Frequency ratio
Tetrads squares (tetrads/10-km squares)
Dianthus armeria 33 31 1-06
Festuca arenaria 55 52 1-06
Juncus balticus to 39 1-13
Asplenium septentrionale Sih 27 15
Carex maritima 47 4] 1-15
Corallorrhiza trifida 76 65 1-17
Vulpia unilateralis 24 20 1-20
Galium pumilum 23 19 1-21
Centaurea cyanus 156 127 1°23
Melampyrum sylvaticum 26 21 1-23
Ophioglossum azoricum 46 37 1-24
Circaea alpina 30 24 1-25
Veronica spicata 20 16 1-25
Atriplex longipes 34 as | 1-26
Linnaea borealis 4] 32 1-28
Zostera marina 99 77 1-28
Deschampsia setacea 7a 55 1-29
Equisetum pratense 117 89 £31
Pyrola media 111 85 1-31
Arabis glabra 4] Si 1-32
Lathyrus palustris 29 22 1-32
Alopecurus borealis 32 24 1-33
Limosella aquatica 73 55 1:33
Orobanche rapum-genistae 129 7, 1-35
Ulmus plotii 44 33 1-33
Hammarbya paludosa 127 95 1-34
Lycopodium annotinum 105 78 1-34
Pyrola rotundifolia (all) 91 68 1-34
Dianthus deltoides 104 T1. 135
Thelypteris palustris 115 85 1-35
Mertensia maritima 136 100 1-36
Sorbus rupicola 76 56 1-36
Isoetes echinospora 108 ) 1-37

Torilis arvensis 112 82 1-37
Scandix pecten-veneris 182 131 1-3
Silene gallica 82 57 1-4

240 D. PEARMAN

seemingly widespread but apparently at very low frequency. I will return at the end to suggest how
we might use this frequency ratio in conservation terms.

I am going to commit a major crime at this stage, and change the subject, before coming back for a
final assault. The reason for this is that the more I delved into the Scarce plants database, the more
avenues I found that could be explored using this discrete set of data, and it seemed a shame to miss
this opportunity to share them with you.

One avenue I explored came from a table in Scarce plants in Britain (Stewart et al. 1994, Table 3)
which listed the totals of Scarce plants that have ever been found in each vice-county, and compared
them with the numbers recorded for that project, i.e. after 1970. I found it difficult to assimilate 112
vice-counties, and I thought it would be useful to interpose a third date, 1930, as that was the date
for current records in the 1962 Adlas of the British Flora (Perring & Walters 1962). I have grouped
the vice-counties into regions (the same as used by the Ecological Flora database in York), and
shown the losses up to 1930, between 1930 and 1970, and after 1970 (Table 7). I have then shown
them as a percentage of species lost, and alternatively, as a variation from the norm, because that
shows the regional differences much better. There are, at least, a couple of caveats, as I have shown,
but I think the figures are very interesting. In this case we are talking about an absolute loss of
species in the entire vice-county, so, although the picture is perforce crude, it shows that we have lost
almost a third of all the individual Scarce Species recorded in each of the vice-counties. The date
classes allow you to see whether the losses have occurred recently or before 1930 and the variations
from the norms show when was the worst in each region. More work could be done by looking at all
the species listed by County Floras as becoming extinct — most Floras, even perfunctory tetrad
Floras, contain at least this sop to the past.

This leads neatly on to the second part of my address; should we redefine 1-15 10-km squares and
16-100 10-km squares for the RDB and Scarce categories, in the light of 40 years of recording? I
believe we should, and there are several ways of doing so.

Firstly, there is the concern of under-recording; Simon Leach elegantly expressed his concerns in
a note in BSBI news last year (Leach 1995) where he said that greater recording effort only meant
that Scarce Species were found in more squares, occasionally leading to tipping over the 100 10-km
square barrier into being “not Scarce”!

Tim Rich in his Monitoring Scheme report (Rich & Woodruff 1990), estimated that the 1962 Adlas
possibly under-recorded by a factor of 50%. Certainly, in preparing the Scarce plants atlas, we felt
really concerned about species that showed a decline between 1962 and 1992, despite all the extra
recording. I must say at this point that I have grave reservations about using statistics to extrapolate
the distribution of rare plants, which often occur in restricted niches or have a very uneven
distribution in the country as a whole. Tim Rich, using Monitoring Scheme data, has attempted to
do this. Although my statistical knowledge was rudimentary, and a long time ago, and my
admiration for him as being one cf the few original thinkers on plant distribution knows no bounds, I
think he is mistaken on this point. The Monitoring Scheme is excellent for evenly distributed and/or
common plants, and is increasingly used nationally for work on these, but I do not think it can be
used for Scarce and RDB plants or those with an uneven distribution.

Secondly, there is the concern that 10-km square recording fails to show historical losses.
Information on this is very rare indeed. Because of the late arrival of the National Grid, and because
intensive recording only started in the 1960s, we have few or no baselines with which to compare.

One of the only exceptions to this is again in Dorset, where the late Prof. Good recorded species
from 7500 stands of vegetation in the 1930s. He marked these stands on a set of six inch Ordnance
Survey maps, which made it possible to add the later National Grid. Andy Byfield and I thus were
able to revisit more than 400 of his heathland sites, where he had recorded 41 heathland plants,
which we chose because of their interest to an informed botanist looking at heaths. They included 18
Scarce plants and four RDB plants, so fitted nicely into my other work.

I doubt if Good went to all his sites on a random basis, but that he gravitated to the better sites,
and he was actually looking at representative examples of habitats rather than all possible sites of
particular species. For example, he looked at over 1000 heathland sites, including 30 on Hartland
Moor alone, but still would not have looked everywhere.

The full results are at last being published (see Byfield & Pearman, 1994 for preliminary findings),
but I use Lycopodiella inundata as an example here just to show how limited is 10-km square data.
His data and our researches since are summarised in Table 8. This shows that an atlas produced in

PRESIDENTIAL ADDRESS

24]

TABLE 7. LOSS OF SCARCE SPECIES PER GROUPS OF VICE-COUNTIES (I.E. SUM OF TOTALS OF
SPECIES*PER V:C.)

S. Wales

N. Wales
NW

NE

S. Scotland
E. Highlands
W. Highland
N. Scotland

NE

S. Scotland
E. Highland
W. Highland
N. Scotland

Notes

1. E.W. and N. Scotland very good even with poorer 1970+ recording;

V.c. numbers

1-6,9-11
13-21

7-8 ,12,22—24 ,30,32-34,36-38

25-29 ,31,53-54,61
39-40 55-58
35,4146

47-52
59-60,69-71

104-112
Totals

V.c. numbers

1—6,9-11

13-21

7-8 1222-24 ,30,32-34 36-38
25-29 ,31,53-54,61
39-40 55-58

35 ,41—46

47-52
59-60,69-71
62-67

68,72-85

86-96

97-103

104-112

Average

Loss
All up to
records 1930
741 81
772 114
786 145
671 112
284 74
338 34
295 47
302 Si
384 87
436 115
612 87
289 22
417 20
6327 989
% Losses
Upto 1930-
1930 1970
10-9 13-9
14-8 20-0
18-5 20-8
16-7 16-2
26-1 22:5
10-1 19-5
15-9 15-6
16-9 1-5
22-7 21-6
26:4 13-9
14-3 11-5
7-6 14-2
4-8 17-3
15-6 17-2

Records
1930+

1970+

24-8
34-8
39:33
32:9
48-6
29-6
31:5
34-4
44.3
40-3
25-8
21:8
227i

32-8

2. Some of the pre-1930 figures might be skewed by poor early recording, etc.

Loss
1930—
1970

Records Loss
1970+ ever
Spiyl 184
503 269
477 309
450 221
146 138
238 100
202 93
198 104
214 170
260 176
454 158
226 63
325 92
4250 2077

Variation from norm

Up to
1930

1930-
1970

1970+

+8-0
—2-0

TABLE 8. LYCOPODIELLA INUNDATA IN DORSET USING PROFESSOR GOOD’S DATA

Good B/P Scarce Now
(1932-38) (1990-92) (1970-92) (1991-94)
No. of sites 48 6 28 20
No. of tetrads 34 5 26 25
No. of 10-km squares 10 3 10 6

242 D. PEARMAN

1939 would have shown ten extant 10-km squares, and that the Scarce Atlas showed ten extant 10-
km squares. But within that:

a. we only refound it in 3 out of 10 of his squares (we did not, of course, look as carefully for new
sites as we did for old);

b. we only refound it in 6 out of 48 of his sites;

c. the scarce total of 10-km squares was the same, but the number of sites had declined from 48 to
28; and

d. the tetrad total still under-represented the known loss at site level (only down from 34 to 26).

There are at least three caveats:

a. Ihave not, of course, covered populations. It is possible (although not true in this case) that 99%
of the Dorset population present in 1932-1938 is still present in 1994. I think, despite the
difficulties, populations must be assessed. The broad bands that Dick David used for his Carex
counts — A 1-20, B 21-100, C hundreds, D thousands are probably fine for most purposes;

b. some plants have better “mobility” than others. Limosella springs to mind. Again, I do not think
this applies to Lycopodiella, but ‘‘mobility” must be borne in mind when considering plant
trends; and

c. itis along time ago! I have not mentioned this, but by using the 1970 baseline, which is probably
the only practical one in a country of disparate habitats and spread of recorders — we are seriously
out-of-date, and that there were another 15 years of agricultural improvement and heathland
dereliction to go before any lessening of pressure on plant sites and populations.

So, in 1900, L. inundata was widespread and 10-km square mappings would have been adequate
to show its decline. By the 1930s it had a very localised distribution in Britain, so it came into the
category of localised species whose decline can be easily seen on a tetrad scale but scarcely on a 10-
km scale. In addition the 10-km picture is inadequate because L. inundata has a localised
distribution, but tends to be fairly common in areas it does occur in.

Many of us have felt that the national 10-km square picture, at least for the Scarce and Rare plants
where we can prove it, increasingly represents a shroud — the dots are still there but the number of
tetrads and populations inside each dot are diminishing. As I have said above, we have little or no
comparative data. Good’s Dorset data show this neatly, and, of course, I could do it for many other
species from Dorset. Here the loss in the number of 10-km squares has barely started to show, years
after a finer resolution would have shown the same. If other areas had similar data available for
comparison then I am sure we would see the same pattern.

In September, 1995 JNCC adopted radically new guidelines for RDB plants based on IUCN
criteria. These new guidelines are much stricter and more quantitative than the old and, I feel, are a
major step forward in conservation terms. It is time-consuming to calculate and apply, but is exactly
on the right lines. It will cover about 60% of the plants in the ‘“‘old”” RDB (Perring & Farrell, 1983)
and will be an adjunct to what I am proposing as it will deal with only the most threatened plants.

To summarise my points then, I suppose there are three alternatives in defining RDB and Scarce
plants:—

1. to accept that the 1962 Atlas considerably under-recorded and therefore double the Scarce Plant
limit to 200 10-km squares;

2. to move to a tetrad basis so RDB species are those occurring in 1-50 tetrads and Scarce those in
51-250 tetrads; and

3. to decide, as did those who originally chose 1-15 10-km squares for RDB and 16-100 10-km
squares for Scarce species, that the rarest 20% in tetrad terms of our 1500 native species are
RDB, and the next 20% are Scarce. There is an interesting recent book on rarity (Gaston 1994)
with many thoughts on this, which suggests 25%, bu I feel that is academic at this stage.

This last alternative would have the extra advantages that individual species could not be
reassessed without assessing all, and therefore there would be greater stability between resurveys,
and perhaps a greater chance of catching declining species, as they are ranked in percentage terms.

There might of course be a political point here. Our excellent environmenta! masters might object
to a fixed percentage of plants always being protected, always needing funding and blocking nice

PRESIDENTIAL ADDRESS 243

new roads! However the day when rare plants are doing anything other than retreating is some way
off and thus I feel a percentage figure is quite realistic at this stage.

I think that all these approaches — 10-km square, tetrad, 1-km square, site, post-1970, post-1987
etc., suffer from another drawback. They are numerical criteria based on a point in time. I am sure
one needs a more subjective approach based on decline and threat, or a more objective approach
based, say, on population size. But populations, as I have said earlier, are extremely difficult to
count, and more difficult to map. The only readily available information now might be these
frequency ratios that I mentioned earlier, which seem to me a relatively crude but valuable extra
tool to go alongside and raw 10-km square or tetrad information. Something is needed to balance
rarity with frequency, as in the two extreme examples I showed earlier of Gentiana verna and
Asplenium septentrionale (Table 6).

The authors of the last edition of the Red data book (Perring & Farrell 1983) attempted something
along these lines using “‘threat’’ categories, but it was much too subjective. We need a “‘rarity index’”’
to express the fact that there is no need to worry about G. verna (unless they build a much bigger
reservoir in Teesdale!!). Their RDB system awarded points for perceived threats, and to some
extent the new IUCN guidelines do the same. I have made a first attempt at a new system, aimed
primarily at Scarce Species, and those RDB species which are not covered by the new IUCN
guidelines, falling into the Lower Risk (LR) category. I have dropped points used by Perring &
Farrell (1983) for Attractiveness, Remoteness and Accessibility, since I do not perceive these to be
threats today. I am totally convinced that today plants are lost almost entirely by ignorance and
neglect; ignorance of their existence and neglect of their habitat. Apart from a handful of orchids
and a couple of ferns I think that all references to threats of collecting should be routinely excised
from any publication. I am also deeply sceptical about the success in protecting species in habitats
protected by conservation agencies — note that I am saying species in their habitats rather than
habitats. Our work in Dorset has shown little difference in protection whether the site is a NNR, a
SSSI or has no protection, the species are still lost in large numbers (Table 9). The key to protection
is management, management and management.

TABLE 9. MAINTENANCE OF PLANT SPECIES DIVERSITY ON PROTECTED AND UN-
PROTECTED HEATHLAND STANDS IN DORSET

Number of indicator species recorded in 1991-1993
as % of indicator species recorded by Prof. Good
(1931-1938)

SSSIs overall 57%
Reserves (NNR, RSPB, DWT) 50%
Other extant sites 35%
Destroyed sites — now Forestry 13%

now Agriculture T%

My first attempt uses the frequency ratios I have been describing, and a figure for the rate of
decline. To arrive at a figure for the decline, I have expressed post-1970 10-km squares as a
percentage of post-1930 squares. I have had to use 10-km records as there are no historical figures of
a finer resolution, but in time we should be able to improve on this. Thus Corynephorus canescens is
found, post-1970, in 71% of the squares it was recorded in post-1930. Ranunculus tripartitus, on the
other hand, is only found in 39% of its previous squares, and thus has declined more. I have then
multiplied this percentage by the frequency ratios I have described earlier to give a Threat Index
(Table 10). The lower the Threat Index figure, the greater the threat. I totally appreciate this is a
first attempt, and that there is an element of rearranging the deckchairs in producing revised lists of
rarity when the actual plants continue to vanish. But it is no use pretending that budget cuts do not
exist, and whilst going for wider and more worthwhile goals we must improve the data we have
available now.

I never thought I would be quite so keen on the division of NCC into country agencies, but we
now have Chris Sydes in Scottish Natural Heritage and Andy Jones in Countryside Council for
Wales actively trying to provide this information and who knows, English Nature might think about

244 D. PEARMAN

TABLE 10. A NEW ‘THREAT INDEX’ FOR A SELECTION OF SPECIES

1970+: 1930+ Frequency
10-km squares % Ratio Index

Scarce

Corynephorus canescens IP BY, til 1-41 100
Veronica spicata 16:17 94 1-25 118
Carex vulpina 22 Dy/ 1-83 104
Luzula arcuata 12:19 63 1-83 115
Galium pumilum £9:43 44 1-21 53
Chenopodium chenopodioides 12:23 a2 2:50 130
Dianthus armeria 36:83 43 1-06 46
Asplenium septentrionale 29235 Vd, ikl) 89
Linum perenne 24:31 Tal 1-50 116
Cystopteris montana 15:18 83 1-67 139
Orchis ustulata 66:134 49 leg) 88
Pulsatilla vulgaris 19:40 48 1-47 70
Illecebrum verticillatum 16:19 84 2-00 168
Elatine hydropiper Iva 90 1-68 152
Ranunculus tripartitus 19:48 39 1°73 68
Carex humilis 28:30 93 4.42 412
Commoner RDB

Fumaria occidentalis Le yailys 88 3-11 274
Gastridium ventricosum 25:41 61 2-00 122
Erica ciliaris 14:16. 87 3:27 286
Genista pilosa 13:15 87 3-15 BIR)
Ophrys sphegodes 15:26 =i Zid 148
Lotus angustissimus 24:34 71 1-40 99
Carex rariflora tere 100 2°13 213
Gentiana verna 4:6 67 7:50 500
Erica vagans De 7 5-20 371

it before too long. If I was rewarded for every time I have been approached for information on
Gentianella anglica in the past two years, I would be in the mountains in Turkey instead of here.
Yes, it has declined in 10-km square totals, and yes it has retreated westwards, and yes it is an
endemic (collect £200 and pass go) but 1994 surveys showed 3 million in the Isle of Wight and half a
million in Dorset. We currently have no way of knowing if this is good or bad and whether resources
should be put into this plant at the expense of others. Another favourite is Dianthus armeria (Table
11) which is suddenly on everybody’s threat list. | am not certain why this is so, because I cannot
trace careful investigative work on it. But for once somebody’s hunch is right (Martin Wigginton
says it is mine — but age dims memory and in fact I think it was Ro FitzGerald who suggested it to
English Nature and me) and the figures are certainly dire —- remember it had the lowest ratio in Table
9. But having gone so far, why haven’t we up-to-date (post-1990) information, and as it is an easy
matter to count an annual, why are there no population data?

I suspect that is another of my failings, that I’m a plodder, one who accumulates information,
rather than somebody with vision (and poetic licence) who is sure Dianthus armeria or Gentianella
anglica, and others are declining and need lots of Species Action plans. But these examples only

TABLE 11. RECORDS OF DIANTHUS ARMERIA

Recording scheme Years — 10-km squares Tetrads 1-km squares _ Sites

1962 Atlas 1930+ 40

Scarce plants atlas (adjusted) 1970+ 31 33 33 34
1970-79 only 10
1980-89 only 14

19905, 10

PRESIDENTIAL ADDRESS 245

demonstrate that we already have or can complete more precise data within a very few years and
that it is essential that we do so.
So in conclusion, I feel I must nail my colours to the mast, and I am recommending:—

1. areal campaign by JNCC and its allies to get complete 1-km square, six figure grid references and
site data, and where feasible populations too, for all Rare and Scarce Plants;

2. that we re-define Rare and Scarce Plants on a tetrad basis, with the rarest 20% as RDB, and the
next 20% as Scarce; and

3. that we weight these figures by a frequency ratio of tetrads/10-km squares, and endeavour to
move to 1-km squares/10-km squares within five years. We should also explore the feasibility of
combining these ratios with a “‘decline” rating as set out in Table 10.

The BSBI and I would be delighted to do the job.

ACKNOWLEDGMENTS

I received great assistance from Chris Preston and the staff at the Biological Records Centre at ITE
Monks Wood. Martin Wigginton at JNCC provided much information on RDB plants and help with
the maps and Alan Morton kindly allowed me to use his DMap programme. Richard Stillman and
Alison Stewart gave assistance on species frequency and Dorset Cyperaceae statistics. Tim Rich
allowed me to use the information on Sorbus lancastriensis and Simon Leach provided valuable help
on an early draft. Arthur Chater has commented on several drafts and has been a major source of
advice. Chris Sydes has also been of major help, especially in tempering theory with reality! Andy
Byfield was of much assistance on the concept of threat ratios. Nevertheless all the statistics and
interpretations are mine and my responsibility.

REFERENCES

ByFIELD, A. J. & PEARMAN, D. (1994). Dorset’s disappearing heathland flora. Unpublished report for R.S.P.B.

Gaston, K. J. (1994). Rarity. Chapman & Hall, London.

Gipzons, D. W., Reip, J. B. & CHAPMAN, R. A., eds (1993). The new atlas of breeding birds in Britain and
Treland: 1988-1991. T. & A. D. Poyser, London.

Leacu, S. J. (1995). The new list of ‘Nationally Scarce’ plants. BSBI news 6: 22-23.

Le Duc, M. G., Hitt, M. O. & Sparkes, T. H. (1992). A method for predicting the probability of species
occurrence using data from systematic surveys. Watsonia 19: 97-105.

NATURE CONSERVANCY COUNCIL (1989). Guidelines for the selection of biological SSSIs. Nature Conservancy
Council, Peterborough.

PEARMAN, D. (1994). Sedges and their allies in Dorset. Dorset Environmental Records Centre, Dorchester.

PERRING, F. H. & FARRELL, L. (1977). British red data books: I — Vascular plants. Society for the Promotion of
Nature Conservation, Lincoln.

PERRING, F. H. & FARRELL, L. (1983). British red data books: 1 — Vascular plants, 2nd ed. Royal Society for
Nature Conservation, Lincoln.

PERRING, F. H. & WALTERS, S. M. eds (1962). Atlas of the British flora. T. Nelson & Sons Ltd, London.

Ricu, T. C. G. & WooprurFF, E. R. (1990). The BSBI monitoring scheme 1987-88. Nature Conservancy Council
CSD Report No. 1265.

Stace, C. A. (1991). New Flora of the British Isles. Cambridge University Press, Cambridge.

STEWART, A., PEARMAN, D. A. & Preston, C. D. eds (1994). Scarce plants in Britain. Joint Nature Conservation
Committee, Peterborough.

Watsonia 21: 253-263 (1997) 247

Gynodioecy in British populations of Eriophorum vaginatum L.
(Cyperaceae)

D. P. STEVENS and T. H. BLACKSTOCK

Countryside Council for Wales, Ffordd Penrhos, Bangor, Gwynedd, LL57 2LO

ABSTRACT

Gynodioecy is shown to be well-established and widespread in British populations of Eriophorum vaginatum L.
Flowering and/or fruiting spikes were sampled from widely distributed populations and examined for the
presence of aborted stamens. The proportion of male-sterile (functionally female) plants varied from 7% to
82%, with an overall frequency of 42%. Sexually intermediate (partially male-sterile) inflorescences appear to
be extremely rare in Britain. Repeated observations at two localities suggest that female tussocks commence
flowering slightly later than hermaphrodite tussocks. A preliminary assessment of seed set in three populations
revealed considerable variation within and between populations, with no consistent differences in fecundity
between sexual morphs. Pollination trials indicated that E. vaginatum is self-compatible; opportunities exist for
geitonogamous (between flower) self-pollination, both between and within spikes on the same tussock. The
findings provide baseline data for further studies on the evolution of gynodioecy in E. vaginatum; it is
provisionally concluded that gynodioecy in this species is a stable form, with no clear evidence of a trend towards
dioecy.

KeEyworbs: sexual dimorphism, male sterility, sex ratios, seed set, self-compatibility, tussock tundra.

INTRODUCTION

Eriophorum vaginatum L. (Hair’s-tail Cottongrass) is one of the major components of ombrotro-
phic mire vegetation in the northern hemisphere, and is a common constituent of tussock tundra
communities in the arctic zone. Commensurate with its ecological importance, several studies of its
population biology have been carried out, primarily in North American tundra sites. These have
shown that tussock longevity may exceed 100 years (Polozova 1970; Mark et al. 1985), that
inflorescence and seed production show annual variation but can be prolific (Polozova 1970; Wein
1973; Chester & Shaver 1982; Gartner et al. 1986; Shaver et al. 1986), that a substantial seed bank
may develop in tundra soils (e.g. McGraw 1980; Gartner et al. 1983), and that successful seedling
establishment is particularly associated with micro-sites disturbed by frost heave, fire and other
agencies (e.g. Wein & Bliss 1973; Gartner et al. 1986). The importance of sexual reproduction in
maintaining natural populations has also been widely emphasised. Polozova (1970) concluded that
reproduction in Russian populations occurs only by seed, and Forrest (1971) found no evidence for
“budding” of young plants from old tussocks in blanket bog vegetation in northern England.
Similarly, Gartner et al. (1986) found that recruitment from seed in Alaskan tundra was probably
sufficient for natural replacement of mature plants. Nevertheless, tillers of E. vaginatum form
adventitious roots (Goodman & Perkins 1968) and are capable of independent growth when
detached from parent plants, so that the possibility of vegetative reproduction cannot be ruled out
completely.

There have been few attempts to characterise the breeding system of E. vaginatum. Florets are
borne in terminal solitary spikes, and there may be few to many spikes per individual tussock,
although a proportion of plants in the population is often non-flowering. Each floret contains three
stamens and a carpel with a single ovule. Like many other members of the Cyperaceae, E.
vaginatum is wind-pollinated and protogynous, but we have traced no studies of compatibility
relationships. Its sexual system is described as hermaphrodite in all major European and British
Floras we have consulted, but in a previous note we recorded the occurrence of male-sterile
(functionally female) plants, and consequently of gynodioecy, in a population from Borth Bog

248 D. P. STEVENS AND T. H. BLACKSTOCK

(Cors Fochno) in Wales (Stevens & Blackstock 1993). Gynodioecy has also been reported by
Raunkiaer in Denmark (Anonymous 1893), and by Polozova (1977) in the Taimyr peninsula in
northern Siberia. Male-sterile plants have vestigial stamens, or staminodes, which are much smaller
than the functional stamens of hermaphrodite plants and are devoid of pollen. In normal stamens,
anthers are usually abscised from their filaments soon after pollen liberation, whereas staminode
anthers fail to dehisce and are retained on short filaments as the nuts begin to ripen. Later in the
fruiting period, after perianth bristles have elongated to form the familiar cotton-like heads,
staminodes and normal filaments are shed, and eventually the spikes disarticulate. Both sexual
forms produce seed, and Polozova (1977) found that females set fewer seeds per spike than
hermaphrodites. No sexually intermediate (partially male-sterile) plants have been reported, but
Polozova noted the presence in small quantity of separate female-sterile (functionally male) plants.

Few data are available on the extent or frequency of gynodioecy in different parts of its world
distribution. Apart from the study by Polozova (1977) at a single site in northern Siberia, there are
no published sex ratio data or sex-specific seed fecundity estimates. The main aims of the present
study were to obtain data on the incidence of male sterility in E. vaginatum populations from a wide
range of localities in Britain, and to compare seed set between hermaphrodite and female plants. In
addition, pollination trials were carried out to examine compatibility relationships.

MATERIALS AND METHODS

SEX RATIOS

Flowering or fruiting spikes of E. vaginatum were sampled from 24 populations in England, Wales
and Scotland in 1991-1993. Individual spikes were collected from separate tussocks separated by at
least one metre, and wherever possible 100 or more plants were sampled. Early in the flowering
season, female inflorescences are noticeably more slender than hermaphrodite spikes, but samples
were collected without conscious bias towards one sexual form or the other. It was assumed that all
flowering tussocks had an equal probability of being sampled, and therefore that inflorescence and
flowering tussock sex ratios are equivalent. No variation in androecial type was observed between
different inflorescences on the same tussock during many hours of field work.

In an ancillary study to test for temporal variation in the frequency of flowering female plants,
populations at two Welsh mire sites were sampled repeatedly during the same flowering season. At
Borth Bog, samples were collected on four separate occasions between mid-March and early June
1992. At Capel Curig, sampling was carried out during three visits between late March and early
June 1993. Both populations had been previously sampled as part of the main study during 1991
when high frequencies of female inflorescences (82% and 56%, respectively) were recorded. At
each site, tussocks were sampled along fixed linear transects.

All sampled inflorescences were sexed under a dissecting microscope, involving careful examin-
ation of at least three, and usually many more, florets in a range of positions from the base to the
apex of each spike. Sexing was most difficult with mature fruiting spikes in which stamen filaments in
hermaphrodites and staminodes in females were being shed, but with experience could be carried
out up to the stage when spikes began to disarticulate. Almost all inflorescences could be
unambiguously determined as hermaphrodite or female (as described in Stevens & Blackstock
1993), but a very few fruiting spikes with mixtures of putative staminodes and remnant elongated
filaments (considered to be indicative of functional stamens) were classified as intermediate.

In the main study, female frequencies were tested for departure from 0-5 and for inter-population
heterogeneity using G-tests (Sokal & Rohlf 1981), and were also related to four geographical and
temporal variables (northing, easting, altitude and date of sample collection) using linear regression
in an attempt to explain the variation observed. In the ancillary study, female frequencies in the
repeat samples from Borth Bog and Capel Curig were analysed using G-tests to test for significant
differences between sampling dates.

SEED SET

Fruiting hermaphrodite and female spikes sampled from each of three British populations (Capel
Curig, Butterburn Flow and Silver Flowe) in June 1993 were scored for seed set. In each case,
fruiting spikes were at an advanced stage of maturation with extensive bristle development but had

GYNODIOECY IN ERIOPHORUM VAGINATUM 249

not yet started to disarticulate. After sexing, spikes were stripped down, and developing nuts were
isolated and examined under a dissecting microscope. Seed viability was assessed by careful
observation of nut size and structure rather than by germination tests; turgid nuts, most of which
had ripe, dark brown testas, were scored as viable. Median numbers of seeds per spike were
calculated for hermaphrodites and females in each population and compared statistically using
Mann-Whitney U-tests.

POLLINATION TRIALS
Rooted portions of five hermaphrodite tussocks, each with several immature inflorescences, were
collected from Capel Curig on 5 March 1992. Plants were grown in a mixture of ericaceous potting
compost and Sphagnum in 16 cm plastic pots. The first flowering spikes appeared after two weeks.
Virgin inflorescences were isolated in porous cellophane bags shortly before stigma lobes in any of
the florets began to protrude beyond their subtending glumes, and were assigned to one of three
treatments: a. “‘crossed’’, in which stigmas were dusted with fresh pollen from a separate plant; b.
“‘selfed”’, in which stigmas were pollinated using another inflorescence belonging to the same plant;
and c. “isolated’’, in which inflorescences remained bagged but were agitated daily to distribute
pollen from dehiscing anthers. Spikes used for pollen donation were also isolated prior to anther
exsertion to exclude the possibility of contamination with wind-blown pollen from other plants.
Fruiting spikes were harvested four to six weeks after pollination, and developing nuts from ten
florets were excised and examined under a dissecting microscope. Swollen ovaries, mostly 2-0-2-2
mm long, containing green turgid embryos, were considered to indicate successful fertilisation;
shirvelled ovaries, usually 0-8—-1-2 mm in length, with shrunken yellowish embryos, were taken to
result from fertilisation failure or early zygote abortion.

One of the hermaphrodite plants was subsequently maintained in isolation and examined for seed
production and viability in June 1995.

RESULTS

SEX RATIOS

Estimated frequencies of female (and sexually intermediate) tussocks in 24 British populations of E.
vaginatum are given in Table 1. Sexual morph ratios are represented diagrammatically in Fig. 1. The
sampled populations are widely distributed, covering 17 Watsonian vice-counties, with an
ecological range which includes low and high altitude blanket bog and lowland raised mire habitats.
Female individuals were present in each population at an estimated overall frequency of 42% (n =
2392), indicating that gynodioecy is a widespread and characteristic feature of E. vaginatum in
Britain. Sexual intermediacy was very rare; only six apparently partially male-sterile spikes were
recorded from three populations (overall frequency 0:2%).

Variation in female frequency between populations was considerable, ranging from 7% to 82%,
and highly significant (G-value for heterogeneity = 511-9, d.f. = 23, P<0.001). There was a
significant excess of female over hermaphrodite plants in five populations from widely scattered
locations in England and Wales, and in eleven others there was no significant difference in the
frequencies of the two sexual forms.

Regression analysis indicated that female frequency was strongly positively associated with date of
sample collection, which explained 58% of the variation between populations (Fig. 2). Regressions of
female frequency on northing (negative) and easting (positive) were also significant. However, after
adjusting female frequency for variation in sampling date, only the regression on northing remained
significant, explaining a further 19% of the inter-population variation in female frequency, and
indicating that southern populations tend to have relatively high proportions of females.

Seasonal differences in flowering phenology between female and hermaphrodite tussocks were
also detected within the Borth Bog and Capel Curig populations (Table 2). In each case, the ratio of
female:hermaphrodite flowering tussocks increased during the main flowering period, which lasted
from mid- or late March, when the first stigmas were exerted, to early May, when the last anthers
had dehisced. After May, there was no further consistent change in female frequency. No instances
of sex change were observed in tussocks of E. vaginatum, and these results suggest that females tend
to come into flower slightly later than hermaphrodites in the same population.

D. P. STEVENS AND T. H. BLACKSTOCK

250

‘JUBSUYTUSIS JOU

SU SORA Xs oMpoOIydeUoy:opeUsy fs] WoO sINJIedop 10] $189} JRONSNRIS ,. ,

“YOY JOM HEL AA (O4TUL posted ALY *OTUU JOUR] “SAT y.

“S'u
“Su
100-0>d
“S°u
100-0>d
100-0>d
100-0>d
100-0>d
100-0>d
C0-0>d
100:0>d
‘Su
‘s'u
10-0>d
10-0>d
100:0>d
‘Sou
‘S'u
‘Sou
100:0>d
‘Su
100:0>d
LSU
[Sal

x%tS9I-D) JO
DOURIYIUBIS

00-0
00-0
00-0
00-0
00-0
00-0
00-0
00-0
00-0
00-0
10-0
00-0
00-0
00-0
v0-0
00-0
00-0
00-0
10-0
00-0
00-0
00-0
00-0
00-0

syoossn}

d}RIPOWI9}UT
yo Aouonboa.y

poveuwns7|

9b-0
cs:
8e:0
9P-0
LI‘0
S10
v1-0
LO:0
LO-0
£9:0
C8°0
tv'0
OS-0
9L-0
C90)
08-0
LS‘0
tr-0
SS‘0)
S70
9S-0
ve:0
bv-0
vv-0

syoossn}
ayPRUd] JO
Aoguanbo.y
poyeunssy

COT
901
COI
66
cOl
OO!
cOl
OOT7
SS
96
801
LOT
cOl
LE
901
901
901
601
OL
L6
IL
OCI
Iv
EG

po1o0os
SONUDISIIOYUI

JO “ON

C661 OuNL ZZ
C661 OUNL TI
C661 API |
7661 OUNL Z]
7661 API TI
Z661 API L
7661 [dy pz
7661 dy ZI

Z661 YOIPIAL SZ
1661 9uNL PZ
1661 ounr [Z
1661 ounr g
1661 unr /
166] 9unfe /
1661 oun Z
1661 ounr Z
1661 9UNL Z
1661 oung Z
L661 ounr Z
1661 9UNL |
1661 API 8Z
1661 ABI 8Z
1661 ACW 1
1661 API I

UOTIAT]OO
ajdures
Jo eq

Wa
Wa
Wa

JeuqeH

(97
O87
097
OCEe
O&
SC
SOP
06!
Ore
SCS
S
DOL
OIL
09]
OP
OOP
O&8¢
OLt
SES
O8
OLZ
OSP
OOS
OO€

(ut)
SpnintTy

£8 LV/XN
9L°99/AN
60°90/f11N
S6 C9/NN
ty OS/HN
S9°¢8/AN
ev 9L/HS
81° 19/UN
97'86/INN
06°8S/XS
16 ¢€9/NS
€L°00/AS
L6'v8/AS
€9°89/NS
€8°SC/AS
OL CO/AS
SO TO/AS
8L°9C/AS
98° 80/AS
[8 c8/f$S
6S OL/HS
by 9L/HS
$6 €8/NS
€9°C6/NS

VIUDIOJOI

pus “S'O

WOALVNIDOVA WOYOHdOlYY AO SNOLLVWTNdOd HSILPUYE 77 WOH
SHdYOW TVWOXHAS AO SHAIONANOAYN GALVWILLSA GNV STIV.LAdG ONITANVS ‘SOLLSIMALOVUVHO LVLIGVH SHILITVOOT ‘1 ATaV

(EL (9°A) IMO ADATIS

(QL °o°A) MOLY WaNqsoiNg
(89 ‘o'A) AyeyeED

(96 '9'A) URMOT]eg

(OL (9°A) ssaip

(Q1] (o°A) ueyoryg

(6p '9°A) gf JUIOUsI)

(LOL (9°) ystueyiuayorypy
(86 ‘9°A) yINUAR YT,

(p ‘o°A) UOWOD UOdWIeYDIO

(Op ‘9'A) Bog YUOg
(p9 ‘0°A) DpISUIOY AA TID

(SQ ‘9°A) [Jay JOUUNYS JeaIH

(Op ‘o°A) 30g UOIRSOL],
(LS “O°A) SOO, WIRTTE HY
(Lg ‘9°A) & SIOO|] YOO]
(6€ ‘0'A) Y SLOOP 4997]
(LS ‘9°A) SIOOJ UDISeA
(LG 9°) Yeo YSIH

(8G ‘o°"A) UOWOD MOpUTT
(6p 9°A) BLIND Jodeg
(6p (9°A) VW IuIoUsIy
(Ly “9°A) URIpRy [904
(Zp °9°A) youl)

(AJUNOD-d9dIA) JUILU 9S

GYNODIOECY IN ERIOPHORUM VAGINATUM 251

Figure 1. Estimated frequencies of male-fertile (hermaphrodite) tussocks (white sectors) and male-sterile
(female) tussocks (black sectors) in 24 British populations of Eriophorum vaginatum.

252 D. P. STEVENS AND T. H. BLACKSTOCK

0.8

0.6

0.4

Frequency of females

0.2

O 4 7 : :
20 40 60 80 100 120
Date of sampling (days after 1 March)

FiGurE 2. Relationship between estimated female frequency and date of sample collection in 24 British
populations of Eriophorum vaginatum. The linear regression equation (Y = 0-007X — 0-178) is highly significant
(P<0-001).

TABLE 2. SEASONAL VARIATION IN THE RELATIVE FLOWERING FREQUENCY OF SEXUAL
MORPHS IN TWO POPULATIONS OF ERIOPHORUM VAGINATUM

Date of sample No. of tussocks Estimated frequency of
Site collection sampled female tussocks
Borth Bog (v.c. 46, Cards.) 16 March 1992 120 0-117
30 March 1992 130 0277-
10 May 1992 100 0-430*
4 June 1992 94 0-553
Capel Curig (v.c. 49, Caerns.) 26 March 1993 o2 0-163
12 May 1993 122 0-402*
8 June 1993 126 0-357

*Female frequency significantly (P<0-05) greater than recorded in previous sample.

SEED SET

Seed output per spike varied considerably, both between spikes within populations and between
populations. Fruiting spikes from Capel Curig contained on average about three times as many
seeds as spikes from Butterburn Flow and Silver Flowe (Fig. 3). A significant difference in seed set
between sexual morphs was detected only at Butterburn Flow, where hermaphrodite spikes
produced an estimated 1-8 times as many seeds as female spikes (Mann-Whitney U-test, U = 4-37,
P<0.001).

Sa “€ (a) “6-6 (A) “€ (dd) ‘5 (O) ‘PE (A) ‘S-P1 CV) :SMmozfoy se ore sonjea ueIpoyy ‘ureig UT suoNe{ndod sa14)
A wor pojduies wnwuisna unsoydougy JO sayids (a[eurdq) d]119}s-o]eu puke (oy1porydeutsoy) o[419J-o[euL UT UONONpod poss Jo suoNNgUIsIp Aouonboly '¢ aun
ayIds Jad speas Jo ‘ON
09 OS OV O€ O¢ Ol 0 09 OS OV Oe Oc OL 0

9|IJ9]S-9]/eW | S|IWaJ-a}ewW
MO|4 WINqie}ing (4) MO} 4 WINqJayNg (J)

O|l49}]S-a}eW itt O|IOJ-a/eW
‘OMO} 4 JEAIIS (Q) ‘OMO|s JAS (D)

GYNODIOECY IN ERIOPHORUM VAGINATUM
Sayids 1O ON

Q|I49}S-a}eW Q[Iaj-ajeuw
‘Hung jaded (q) ‘BUND jadeg (Vv)

254 D. P. STEVENS AND T. H. BLACKSTOCK

TABLE 3. NUMBERS OF DEVELOPING EMBRYOS (OUT OF 10 SCORED) IN INFLORESCENCES
OF ERIOPHORUM VAGINATUM FOLLOWING CONTROLLED POLLINATION TREATMENTS

Maternal parent

Treatment 1 2 3 4 5

Crossed OFS 0 8-550 7
Selfed 720 8

Isolated 0 1 7

POLLINATION TRIALS

Results of the pollination tests are shown in Table 3. Embryo development was recorded in two out
of three selfed spikes of E. vaginatum compared with five out of seven crossed inflorescences.
Asexual seed production seems unlikely in this gynodioecious species, and the results strongly
suggest self-compatibility. Swelling ovaries were also noted in two out of three isolated inflores-
cences, indicating the potential for effective pollination of stigmas with pollen from the same spike.
Gynoecia in some spikes under each of the three treatments failed to develop normally, probably
due to sub-optimal culture conditions.

The hermaphrodite plant subsequently maintained in isolation generated 20 inflorescences in
1995, of which 18 developed ripe nuts. A total of 141 apparently fertile nuts was produced, and
100% germination success was recorded in a sample of 20 seeds sown on to filter paper discs in petri
dishes. These results provide further evidence for self-compatibility in E. vaginatum.

DISCUSSION

DISTRIBUTION AND FREQUENCY OF GYNODIOECY

The occurrence of sexual dimorphism in British populations of Eriophorum vaginatum has been
overlooked until recently. Stevens & Blackstock (1993) recorded the presence of male-sterile,
functionally female, plants in mid Wales, and this extended study has shown that gynodioecy is a
widespread and characteristic feature of the breeding system of E. vaginatum in Britain. According
to Hultén (1962), populations of E. vaginatum in Europe and northern Asia represent E. vaginatum
subsp. vaginatum, and it would be particularly interesting to establish whether gynodioecy also
occurs in New World populations, especially in the ecologically well-studied E. vaginatum subsp.
spissum. Many aspects of the population biology of this latter taxon have been studied recently in
cottongrass tussock communities in North American tundra ecosystems, including its flowering
behaviour and capacity for regeneration from seed (e.g. Gartner et al. 1986), but we are not aware
of any accounts of its sexual system.

SEX RATIOS

High frequencies of female tussocks were found in most populations of EF. vaginatum during the
present study, occasionally exceeding 50%. Female frequencies of this magnitude are very unusual
in gynodioecious species (see Couvet et al. 1990), and it is important to consider possible sources of
bias in sex ratio estimation. Care was taken during field sampling to ensure that inflorescences were
collected from separate tussocks and without conscious bias towards either sexual morph. Non-
flowering tussocks were common in the sampled populations, and there is a possibility that sex ratio
estimates have been influenced by differences between sexual forms in flowering propensity,
although there is no evidence for this. At two sites, Borth Bog and Tregaron Bog, samples were
collected late in the fruiting season when spikes had begun to disarticulate, and it is possible that
earlier completion of fruit maturation in hermaphrodites could have inflated the estimates of female
frequency. In all other populations, samples were collected before fruiting spikes began to break up,
and in many of these females appeared to be present among flowering tussocks at genuinely high
frequency. Low female frequencies were found mostly in populations sampled early in the flowering
season and may have been underestimated because of the late flowering of female tussocks (see
below).

GYNODIOECY IN ERIOPHORUM VAGINATUM 255

Polozova (1977) recorded a female frequency of 17% (n = 483 tussocks) from a population of E.
vaginatum in northern Siberia.

SEED SET

During the present study, seed set in E. vaginatum was found to be mostly low but highly variable,
with median values for three populations ranging from four to 14 seeds per spike. We are not aware
of any previous estimates of seed production in British populations, although Rawes & Hobbs
(1979) noted that few seeds were produced in a population in the north Pennines. High levels of
inter-population variation are also apparent from studies in North America, where average values
of three to 82 seeds per spike have been reported (Wein 1973; Chester & Shaver 1982; Gartner et al.
1986). The genetic and environmental factors controlling seed set in E. vaginatum are poorly
understood, although Gartner et al. (1983) found that spikes harvested early and late in the
flowering season contained fewer seeds than those collected during the intervening period.

A significant difference in seed set was detected between sexual morphs in only one population, in
which hermaphrodite spikes yielded 1-8 times as many seeds as female spikes. Polozova (1977)
found a similar sex difference (male-fertile 30%, male-sterile 14%) in a Siberian population.
Reduced seed output in females was attributed in that case to a lack of opportunities for
geitonogamous (inter-flower) self-pollination, but no studies to confirm this were reported.

A full comparative assessment of seed output from the two morphs would need to take
inflorescence production per tussock into account, as well as variation in annual fecundity at
different life-history phases.

SELF-COMPATIBILITY

The establishment of self-compatibility in FE. vaginatum contributes to characterisation of its
breeding system. The results also accord with Molau (1993) who stated that very few tundra species
are known to be fully self-incompatible.

The extent to which self-fertilisation takes place in natural populations of E. vaginatum must
depend on the opportunities available for self-pollination. According to Hegi (1967), self-
pollination is precluded by temporal separation of male and female function. However, our field
observations suggest that protogyny is at best hemisynchronous (sensu Lloyd & Webb 1986), with
overlap of pollen liberation and stigma receptivity between different inflorescences in the same
tussock. Furthermore, Polozova (1977) found that there was typically an overlap of one to three
days in the female and male stages of flowering within an inflorescence, and evidence for within-
spike self-fertilisation was found during the present study. It is likely therefore that some degree of
self-fertilisation takes place geitonogamously in hermaphrodite plants in the field, although further
studies are required to assess its frequency.

FLOWERING PHENOLOGY IN HERMAPHRODITE AND FEMALE PLANTS
Flowering periods of female and hermaphrodite plants appear to be somewhat asynchronous in at
least some British populations of E. vaginatum. A similar phenomenon was reported for
gynodioecious Gingidia decipiens (Hook. f.) Dawson (Apiaceae) in New Zealand by Webb (1976),
although in this species female plants peak and finish flowering before rather than after
hermaphrodites in the same population. Early female flowering was interpreted as resulting from
selection for fruit maturation before the onset of limiting winter conditions. In many British
populations of E. vaginatum, environmental conditions are probably more important in controlling
the beginning rather than the end of the flowering period, since hard spring frosts cause severe
damage to exposed inflorescences (Goodman 1950), whereas seeds are shed around mid-summer
well before end-of-season limitations become apparent. It is possible that in protogynous
hermaphrodite plants selection against early flowering is countered by competition between pollen
donors for fertilisation of the first available ovules. In male-sterile plants, there can be no such intra-
sexual selection for early flowering, and females may flower as late as possible before pollen
availability becomes limiting due to completion of flowering in hermaphrodite plants.
Interestingly, Polozova (1977) found that in the short summers of northern Siberia female plants
usually finished flowering as pollen production by hermaphrodites reached its peak. The differences
between sexual morphs in flowering period evident in British populations do not therefore appear to
extend to arctic Siberia where unfavourable late-season conditions may impinge more directly.

256 D. P. STEVENS AND T. H. BLACKSTOCK

MAINTENANCE OF GYNODIOECY

There has been considerable recent interest in the selective factors involved in the evolution and
maintenance of gynodioecious polymorphisms. In E. vaginatum, such factors are largely unknown,
as are the stages of the life cycle at which they are expressed. Phenotypic models of gynodioecy with
nuclear genetic control of male sterility predict hermaphrodite-predominance (Lloyd 1975, 1976;
Charlesworth & Charlesworth 1978), even with strong female advantage in seed production, unless
sexual forms differ in their capacity for vegetative reproduction (Stevens & van Damme 1988) or in
adult survival rates (van Damme & van Damme 1986). Differences in seed set per spike between
sexual forms were established for two out of four populations studied by us and Polozova (1977), but
in each case it was hermaphrodites rather than females that yielded the most seeds, pointing to a
possible additional reproductive disadvantage of the latter. Other fitness components have not been
studied. Vegetative reproduction is unlikely to be of major significance in E. vaginatum, but the
possibility of differences between sexes in tussock mortality rate or flowering propensity cannot be
excluded. It is also possible that other assumptions of the nuclear phenotypic models do not apply in
this case. For example, male sterility may be wholly or partly under the control of cytoplasmic
genes, or there could be overdominance for fitness at the sex-determining loci (or closely linked
sites). A range of genetic and demographic effects has been demonstrated for Plantago lanceolata L.
(Ribwort Plantain), which has many similarities with E. vaginatum in its inflorescence structure and
pollination mechanism, by van Damme (1984) and van Damme & van Delden (1982, 1984).
However, E. vaginatum differs from P. lanceolata in being self-compatible, so inbreeding
depression as well as pleiotropic effects of male sterility genes may be involved in the maintenance
of gynodioecy in the cottongrass. On the other hand, differentiation of hermaphrodite and female
plants is more clear cut in FE. vaginatum, and the genetic basis for male sterility may prove to be
much simpler than the complex mechanism found in P. lanceolata.

EVOLUTIONARY STABILITY OF GYNODIOECY

The role of gynodioecy as a transitional stage in the evolution of dioecy from hermaphrodity has
attracted considerable attention since Darwin (1877). Ross (1978) classified gynodioecy into two
types: an unstable form in which hermaphrodites are partially female-sterile, and a stable form
which shows no such tendency towards dioecy. Unstable gynodioecy was considered to be
associated with simple nuclear genetic control of male sterility, and stable gynodioecy with more
complex mechanisms of inheritance, particularly those involving cytoplasmic factors. In E.
vaginatum, there appears to be little evidence for evolution from gynodioecy towards dioecy; a few
fully male plants were reported by Polozova (1977), but they occurred at very low frequency, and we
have found no comparable examples in British populations. It seems likely therefore that
gynodioecy in E. vaginatum is of the stable type.

ACKNOWLEDGMENTS

We are grateful to Fiona Evans, Dr Liz Howe, Drs Jane and Neil Hughes, Dr Simon Lawrence,
Stuart Smith, Marcus Yeo and especially Dr Carrie Rimes for collecting cottongrass inflorescences
for us. We also thank Ian Langford (Scottish Natural Heritage) for guiding us to the Silver Flowe
National Nature Reserve. Dr Jos van Damme kindly provided constructive comments on a draft
manuscript.

‘ REFERENCES

ANONYMOUS (1893). Meddelelser fra den Botaniske Forening i Kj@benhavn. Botanisk tidsskrift 18: 18-29.

CHARLESWORTH, B. & CHARLESWoRTH, D. (1978). A model for the evolution of dioecy and gynodioecy.
American naturalist 112: 975-997.

CHESTER, A. L. & SHAVER, G. R. (1982). Reproductive effort in cotton grass tussock tundra. Holarctic ecology 5:
200-206.

Couvet, D. et al. (1990). Co-evolution between two symbionts: the case of cytoplasmic male-sterility in higher

GYNODIOECY IN ERIOPHORUM VAGINATUM 257

plants, in FUruyAMA, D. & AntTonovics, J., eds. Oxford surveys in evolutionary biology 7: 225-247. Oxford
University Press, Oxford.

Darwin, C. (1877). The different forms of flowers on plants of the same species. J. Murray, London.

Forrest, G. I. (1971). Structure and production of north Pennine blanket bog vegetation. Journal of ecology 59:
453-479.

GartTn_ER, B. L., CHAPIN III, F. S. & SHAvER, G. R. (1983). Demographic patterns of seedling establishment and
growth of native graminoids in an Alaskan tundra disturbance. Journal of applied ecology 20: 965-980.

GarTn_ER, B. L., Cuapin III, F. S. & SHAVER, G. R. (1986). Reproduction of Eriophorum vaginatum by seed in
Alaskan tussock tundra. Journal of ecology 74: 1-18.

GoopmaNn, G. T. (1950). A contribution to the study of the ecology of Eriophorum species in Britain.
Unpublished M.Sc. thesis; University College of Wales, Aberystwyth.

GoopMaN, G. T. & PERKINS, D. F. (1968). The role of mineral nutrients in Eriophorum communities. III.
Growth response to added inorganic elements in two E. vaginatum communities. Journal of ecology 56:
667-685.

HEai1, G. (1967). Ilustrierte Flora von Mitteleuropa. 2, 2nd ed. Paul Parey, Berlin.

Huttén, E. (1962). The circumpolar plants. I. Vascular cryptogams, conifers, monocotyledons. Almquist &
Wiksell, Stockholm.

Lioyp, D. G. (1975). The maintenance of gynodioecy and androdioecy in angiosperms. Genetica 45: 325-339.

Lioyp, D. G. (1976). The transmission of genes via pollen and ovules in gynodioecious angiosperms. Theoretical
population biology 9: 299-316.

Lioyp, D. G. & Wess, C. J. (1986). The avoidance of interference between the presentation of pollen and
stigmas in angiosperms. I. Dichogamy. New Zealand journal of botany 24: 135-162.

McGraw, J. B. (1980). Seed bank size and distribution of seeds in cottongrass tussock tundra, Eagle Creek,
Alaska. Canadian journal of botany 58: 1607-1611

Mark, A. F., FETCHER, N., SHAVER, G. R. & Cuapin III, F. S. (1985). Estimated ages of mature tussocks of
Eriophorum vaginatum along a latitudinal gradient in central Alaska, U.S.A. Arctic and alpine research 17:
1-5.

Mo tau, U. (1993). Relationships between flowering phenology and life history strategies in tundra plants. Arctic
and alpine research 25: 391-402.

PoLozova, T. G. (1970). Specific biological features of Eriophorum vaginatum L. as a hummock-forming plant
(according to the observations in the tundras of the western part of Taimyr peninsula). Botanicheskii
zhurnal 55: 431-442. [In Russian]

Po.Lozova, T. G. (1977). Observations on the flowering of cottongrass Eriophorum vaginatum L. (Cyperaceae)
in western Taimyr. Botanicheskii zhurnal 62: 1338-1340. [In Russian]

Rawes, M. & Hosss, R. (1979). Management of semi-natural blanket bog in the northern Pennines. Journal of
ecology 67: 789-807.

Ross, M. D. (1978). The evolution of gynodioecy and subdioecy. Evolution 32: 174-188.

SHAVER, G. R., FETCHER, N. & CHAPIN III, F. S. (1986). Growth and flowering in Eriophorum vaginatum: annual
and latitudinal variation. Ecology 67: 1524-1535.

SOKAL, R. R. & Rou_r, F. J. (1981). Biometry, 2nd ed. W. H. Freeman, New York.

STEVENS, D. P. & BLacxstock, T. H. (1993). Sexual dimorphism in Eriophorum vaginatum L. (Cyperaceae).
Watsonia 19: 190-192.

STEVENS, D. P. & VAN DAMME, J. M. M. (1988). The evolution and maintenance of gynodioecy in sexually and
vegetatively reproducing plants. Heredity 61: 329-337.

VAN Damme, J. M. M. (1984). Gynodioecy in Plantago lanceolata L. III. Sexual reproduction and the
maintenance of male steriles. Heredity 52: 77-93.

VAN DamME, J. M. M. & VAN DAMME, R. (1986). On the maintenance of gynodioecy: Lewis’ result extended.
Journal of theoretical biology 121: 339-350.

VAN DamME, J. M. M. & VAN DELDEN, W. (1982). Gynodioecy in Plantago lanceolata L. 1. Polymorphism for
plasmon type. Heredity 49: 303-318.

VAN DamMME, J. M. M. & VAN DELDEN, W. (1984). Gynodioecy in Plantago lanceolata L. IV. Fitness components
of sex types in different life cycle stages. Evolution 38: 1326-1336.

Wess, C. J. (1976). Flowering periods in the gynodioecious species Gingidia decipiens (Umbelliferae). New
Zealand journal of botany 14: 207-210.

WEIN, R. W. (1973). Biological Flora of the British Isles: Eriophorum vaginatum L. Journal of ecology 61: 601—
615.

WEIN, R. W. & Buss, L. C. (1973). Changes in arctic Eriophorum tussock communities following fire. Ecology
54: 845-852.

(Accepted July 1996)

OO _—

=

45% .
ey >»

o>: “ein

7 ,

ae

5° Leta nh peice een
coer we sore ete eer origins ln eanre Bie <page
rs eS) ye eeeeey S “at SaaNs ry Rieter ee
22 Serta OE aie Seritt aedee ,
peta ibe ge en AES: Tf eh [tS ae ei iv a

eo *

Wirt OVS ash

! Seabeetinlines na noteaatat Sie tae ;
‘ inlore wr ae Se 42 Seed Fid?

=e are es 3

oN é i 4 1%; Le + 14 ane Pate sabe’ * ak u 7 i ap ty i] Peas
: 7 = Foy -— o > 7"
pepe se tee Oye ve Rey a Fy pbs its et} aes a
7 a : Sos £ behest ies ao, A a) sai th fel %: :
- = — i= * 4 AO - Fi
in - 3 1 j ye aN est “oy. tad Ey ar = Vel peel ven
Boge eetA (oat os Seth oe
: a Sut anal at _.2 ee ee Bee
by LT ip asia! 5 Ale © gy eA ae ae
= ~ y i yl - 7
+ J 2 PL os
eS vi? ; AA) -
: why fe
A ' ©. pa} o¢)
r i en) ay
4 (i Pr aes bd
ia . t
| ve
} } 7. “ys ul
, ili ete coe
‘ , ’ 18
{ on t ~ Mek 4
4 = Fi mt og
iz 5 cr Weds out
of be : nF a5
: +s —sit pe.
{ ‘ J y i
- . e 7 Treat
7 7 =f
a Spe i
; £221
= =.
; eye iru. Are “ *otst DAES ch
ps oe
; ; : ich EGR GaSe eat) ,
‘ y — _
j ae iio [-eemaye ee, iL: te ciond iw Et ;
: Neen 1a a sort We
; five (a Asis = pie? we nti ty 1S ered fash si sae
on : Doh al r
=i: Ve CUE J RRWIO teapaph Oley ai eee ei dosiebla
77 a = “3 oe | 4 * Pate || ss. Jon
: aa
uke ae » ale EW _ oO many Fediorid ihe ,
2 oa
: a ee
n> 1 e i cies ae aun a A siete aes i.
Ss re) =
= F = : y i
a 8 Tid, : ae
i ; rm aus
any,
= eo ee

Watsonia 21: 265-276 (1997) 259

A morphometric study of the woodrushes Luzula campestris
(L.) DC., L. multiflora subsp. multiflora (Ehrh.) Lej. and
L. congesta (Thuill.) Lej. (Juncaceae) in the British Isles

J. V. ARMSTRONG

Division of Environmental & Evolutionary Biology, Institute of Biomedical & Life Sciences,
Graham Kerr Building, University of Glasgow, University Avenue, Glasgow, G12 8OQ

ABSTRACT

A morphometric study of the three widespread taxa, Luzula campestris (L.) DC., L. multiflora subsp. multiflora
(Ehrh.) Lej. and L. congesta (Thuill.) Lej. (section Luzula; Juncaceae) has revealed all three taxa to be distinct,
each characterised by a unique character constellation and ecological ‘preference’. Whilst overlap occurs in
some of the character ranges, identification can still be afforded by using a combination of floral and/or fruit
characters. A table-‘of critical values is provided for each taxon for the identification of problematic specimens.

KEyYworDs: taxonomy, ecology, diagnostic characters.

INTRODUCTION

The genus Luzula DC. (Juncaceae) comprises over 100 species (Novikov 1990) which are widely
distributed throughout much of the Northern Hemisphere (in particular Europe and Central Asia),
with a few species extending into the Southern Hemisphere and the Tropics. It has been
traditionally divided into the three subgenera, Luzula (syn. Gymnodes (Griseb.) Buchenau),
Pterodes (Griseb.) Buchenau, and Anthelae (Griseb.) Buchenau on the basis of inflorescence, seed
and leaf tip characters. Novikov (1990) further divided subgenus Luzula, the largest of the three
subgenera, into two sections and nine subsections. Section Luzula is characterised by the possession
of subumbellate inflorescences, with pedunculate or subsessile clusters on straight erect branches at
anthesis (Kirschner 1991). It is usually regarded as a taxonomic complex containing a range of
karyotypes (polyploids and agmatoploids), in which recognition has been variously afforded to the
discrete entities at the ranks of species, subspecies or variety dependent on the taxonomic authority
(see Fernald 1945; Voss 1966; Kirschner 1980, 1990; Kirschner & Kyisa 1979). Buchenau (1890,
1906) regarded the section as a single polymorphic species, with varietal and subspecific status being
applied to polyploid members. In contrast, Kirschner (1991) considered the section to contain up to
50 species, with a typically high proportion of continental endemics, of which 15 occur in Europe.
Chrtek & Kiisa (1980) recognised four species and three subspecies for their account in Flora
Europaea.

In Britain and Ireland, the section is represented by the three closely related and widespread taxa
Luzula campestris (L.) DC. (2n = 12), L. multiflora subsp. multiflora (Ehrh.) Lej. (2n = 36) and L.
congesta (Thuill.) Lej. (2n = 48), in addition to the localised L. pallidula Kirschner, restricted to
fenlands in Huntingdonshire (v.c. 31) (Stace 1991). Difficulties have been encountered mainly with
the discrimination of L. campestris and L. multiflora subsp. multiflora, as a result of the considerable
overlap in the character ranges of the two taxa (Wigginton & Graham 1981). In addition, there has
been some dispute over the ranking of L. congesta in Britain and Ireland. Whilst this taxon has
traditionally been treated as a subspecies of L. multiflora in many British Floras (Clapham, Tutin &
Moore 1987; Stace 1991) and in Flora Europaea (Chrtek & Kfisa 1980), a specific status has been
argued for by Nordenskidld (1956), Buchanan (1960) and most recently by Kirschner (1990) and
Kirschner & Rich (1996). In an attempt to address these difficulties, a taxonomic reappraisal of the
three taxa was undertaken, the results of which are reported in this paper. The objectives of the

260 J. V. ARMSTRONG
study were to determine the discreteness or otherwise of the three taxa; to identify the characters of
taxonomic value; and to appraise the ranking of the three taxa.

MATERIALS AND METHODS

Callections of single plants were made from all three taxa from a variety of locations throughout
Cumbria (v.cc. 69, 70) and West Lancashire (v.c. 60). These were subsequently sorted into three
groups on the basis of gross morphology using the gestalt method. The groups so formed were then
used as the source material for a re-appraisal of all characters traditionally used in Floras and for the
investigation of new diagnostic characters (see Table 1). Floral and fruit character (see Fig. 1)
measurements, with the exception of pollen, were based on a sample size of 150 individuals from
each of the three taxa, whereas vegetative characters were assessed on 25—40 individuals. This bias
reflected the reported importance of floral and fruit characters in the discrimination of the three taxa
(Wigginton & Graham 1981; Kay 1993), and the use of fruit characters to test the validity of the
groupings (see below). All floral and fruit characters were measured using a monocular microscope
at x60, unless otherwise stated. Pollen dimensions and stomatal dimensions were taken from 40
individuals per taxon, measured at a magnification of x450. For vegetative characters, 40
individuals were scored per group, with the exception of leaf thickness (25 individuals). Leaf width

TABLE 1. CHARACTERISTICS OF LUZULA CAMPESTRIS, L. MULTIFLORA SUBSP.
MULTIFLORA AND L. CONGESTA (RANGE, MEAN AND STANDARD DEVIATION)

Character

Habit
Height-complete (cm)
Height-culm (cm)
Leaf
Width (maximum) (cm)

Thickness (mm) (at 50% leaf length)

Colour
Tip
Guard cell length (mm)
Inflorescence
Form
Floral clusters
Number
Arrangement
Number of flowers within
Flowers
Anther (mm)
Anther length:filament length ratio
Pollen grain diameter (um)
Perianth segment
Length (mm)
Width (mm)
Bract length (cm)
Fruit
Capsule valve length (mm)
Seed
Length (1) (mm)
Breadth (b) (mm)
1/b ratio
Caruncle (mm)
Size (1 Xb)
Weight (mg)

L. campestris

Loosely tufted/rosette
6:4-25-6, 14-94+4-78
6-0-23-4, 15-144-51

2:0-4-5, 2:840-57
0-16-0-20, 0-18+0-01
Light green, shiny
Truncate

44-0-63-0, 52-9+0-01

Loosely cymose

2-7, 3:6+1-23
1 sessile, 2-7 pedunculate
3-11, 5-8+1-62

1-13—3-23, 1-81+0-03
2:5-9-4, 5-141-56
33-3-44-4, 39-2+3-00

2:7-4-1, 3-140-22
0-76-1-32, 0-97+0-11
0-6-3-0, 1-6+0-55

2-6-3-6, 2-9+0-20

0-79-1-26, 1-:04+0-09
0-74-1-10, 0-90+0-07
1-0-1-3, 1-18+0-07

0-21-0-63, 0-45+0-08
0-66-1-31, 0-9340-13
0-43-0-99, 0-67+0-13

L. multiflora
subsp. multiflora

Densely tufted
18-7-52-2, 31-647-89
15-1-46-5, 29:747-40

3-0-7-0, 4-340-85
0:20-0:28, 0-22+0-02
Dark green, dull
Rounded

66-0-96-0, 77-4£0-01

Subumbellate

3-17, 5-7+1-99
Pedunculate/subsessile
6-18, 9-342-24

0-66-1-58, 1-16+0-20
1-0-3-5, 1-90-60
29-6-48-1, 38-6+5-20

2:5-3-5, 2:-940-18
0-47-0-82, 0-65+0-08
1-4-6-5, 2:941-30

1-8-2-6, 2:2+0-20

0-79-1-32, 1-03+0-11
0-50-0-87, 0-73+0-07
1-2-2-6, 1-44+0-22

0-16-0-74, 0-41+0-01
0-52-1-24, 0-81+0-01
0-18-0-56, 0-33+0-01

L. congesta

Loose-densely tufted
20-8-85-0, 35-2+10-76
20-2-84-0, 36-9414-89

Bright green
Rounded

Congested

1-4 congested
Subsessile/sessile
28-99, 52-4

1-52-1-67, 1-60+0-11
1-48-2-7, 2:05+0-29
27-8-58-3, 40-43+5-41

2:5= 330129 20:23
0-50-0-92, 0-72+0-07
1-0+4-7, 2-1+0-91

1-8-2-8, 2-4+0-2

1-0-1-39, 1-21+0-08
0-60-1:03, 0-86+0-09
1-3-1-9, 1-48+0-13
0-21-0-71, 0-4+0-08
0-68-1-40, 1-00+0-15
0-44-1-20, 0-66+0-12

LOUZOLAIN DHE-BRITISHMISLES 261

Stamen Seed Capsule

Ficure 1. Some of the characters measured in the morphometric analysis of Luzula for which precise description
is required. Stamen: al, anther length; fl, filament length. Seed: sl, seed length; sb, seed breadth; crl, caruncle
length. Capsule: cl, capsule length.

measurements were made at three points along the total length (25%, 50% and 75% from the base),
and leaf thickness at half-way along the leaf using a screw gauge micrometer. Measurements of
stomatal dimensions were made from silicone impressions taken half-way along the leaf from
material grown in the glasshouses at the University of Lancaster. In all cases the maximum
dimensions of a character were used and measured to the accuracy indicated in Table 1, using a
calibrated eye-piece or a ruler.

To verify the validity of the groups determined ‘by eye’, a canonical discriminate analysis was first
performed using the fruit characters of seed length, seed breadth, seed weight, seed caruncle length
and capsule length, and perianth width. In canonical discriminate analysis (CDA) combinations of
the original variables are sought which maximise the ratio of between group to within group
variation, such that functions of the original variables can be used to discriminate between the
groups. The analysis was performed using the statistical package SAS in which all squared distances
between group means are Mahalanobis D* statistics (D*(;,=(x; — x;) cov” '(x; — x).

On validation of the three groups, an analysis of variance was performed using a Tukey LSD
(least significant difference) test on all the quantitative data. For selected quantitative characters
(see below) ratios were also produced (i.e. anther length/filament length ratio, seed length/breath
ratio and fruit index (= capsule length X perianth width)) and critical value tables constructed. The
latter involved first dividing the total range of the character for all three taxa into 100 size classes.
For example, if the character range varied from 1-80 mm to 3-60 mm, as for capsule segment length,
the difference, 1-80 was divided by 100 to give 0-018. This value was then added to the smallest value
(1-80) to give size class 01 a value of 1-800—1-818, and then successively to the culminating total to
give size classes 02, 1-819-1-836, class 03, 1-837—1-872, class 04, 1-873-1-890 . . . class 100, 3-583—
3-600. The means for each taxon were then replaced by their appropriate class number, which in the
above case was 63 for L. campestris (k = 2-934), 22 for L. multiflora subsp. multiflora (k = 2-196)
and 32 for L. congesta (k = 2-376) (see Table 2). Plus and minus signs were then allocated to
characters to describe their relationship with respect to one another. The scoring method involved
assigning a plus value to those characters which had a larger class value for L. campestris, and a
negative value to those characters which showed the opposite tendency (a negative value was

262 J. V. ARMSTRONG

TABLE 2. CLASS NUMBERS AND CRITICAL VALUES FOR LUZULA CAMPESTRIS,
L. MULTIFLORA SUBSP. MULTIFLORA AND L. CONGESTA
(SEE TEXT FOR DETAILS OF CALCULATIONS)

Taxa
L. multiflora
Stage of development Character L. campestris subsp. multiflora L. congesta
Anthesis Anther length/filament length ratio +49 +10 +05
Anther length +46 seit) +07
Perianth width +60 oe. a29
Leaf width =i5 45 =32
Summed totals 140 6 12.
Fruiting Anther length/filament length ratio
(dehisced) +52 +14 i2
Perianth width +60 +22 +29
Capsule length +63 +22 3,
Seed length +41 +39 +70
Seed breadth +66 +39 +50
Seed length/breadth ratio =iq ee =30
Seed weight +48 +16 +47
Leaf width =15 —45 3
Summed totals 304 80 166

assigned to both L. multiflora subsp. multiflora (-22) and L. congesta (—32) in the example). This
procedure was repeated for a selected number of characters (selected for their large differences in
mean expression and low levels of variation about the mean) and the character class numbers then
combined to give a summed critical value for each group/taxon for a particular combination of
characters (see results). The characters selected included perianth segment width, anther length and
anther length/filament length ratio, capsule length, seed length, seed breadth, seed length/breadth
ratio, seed size, seed weight, in addition to leaf width and leaf thickness. The rationale for this
method is that whilst it may not be possible to use a single character to identify all samples
encountered, this aim can be achieved through the use of several characters combined into a class
number total (Whitehead 1954). This method was then tested on ‘typical’ and ‘problematic’
herbarium material not used in the analysis so far.

RESULTS

The gestalt method resulted in the production of three groups coincidental with the three taxa under
investigation, plus a few specimens in which placement was not immediately obvious. The three
groups could be clearly differentiated from one another and were characterised by unique character
combinations.

The canonical discriminate analysis also supported the existence of group structure and found
highly significant differences between the three groups (Table 3). The first canonical variate axis
(CVI) accounted for 83% of the variation and the second canonical variate axis (CV2) for 17%.
CV1 was able to discriminate all three taxa, but optimally L. campestris from L. multiflora subsp.
multiflora and L. congesta, with seed breadth, seed weight, capsule length and perianth width being
the most discriminatory characters (Fig. 2). CV2 primarily distinguished L. congesta from the other
two, with seed length contributing most to the discrimination afforded. All characters, with the
exception of seed caruncle length were found to have high discriminatory values. The significance of
these group differences was further supported by multivariate analysis, for example Wilks’ Lambda
0-044, F-ratio = 278-16, d.f. = 12,884, p<0-001.

The differentiation of these groups has been further supported by a Tukey LSD test on floral, fruit
and foliar material (Table 4).-From the scatter plots, it was also evident that combinations of fruit

LUZULA IN THE BRITISH ISLES 263

TABLE 3. TOTAL CANONICAL STRUCTURE FOR LUZULA CAMPESTRIS, L. MULTIFLORA
SUBSP. MULTIFLORA AND L. CONGESTA

Canonical axis 1 Canonical axis 2
Canonical correlation 0-941 0-786
Approximate standard error 0-005 0-181
Eigenvalue 7-729 1-161
Proportion oi total variance 0-827 0-172
Total Structure
Seed length 0-142 0-853
Seed breadth 0-682 0-273
Seed weight 0-743 0-531
Seed caruncle length 0-163 —0-190
Capsule length 0-867 —0-246
Perianth width 0-852 —0-307
Class means on canonical variables
L. campesiris 3-479 —0-823
L. multifiora subsp. multifiora —3-301 —0-965
L. congesta =—O-179 1-788
5 my a
| | OL. campestris
| @ L. multifiora
+ + L. congesta
+ a
3 arora
a ra tae +
i Ss ae ul ane as O
2- ee ae
| . ata aoe rn ead oT
: wy ta 28 0
ae : = + - e)
= | . Oe tus ae + ofr00 &
= 0- Soeiee i SSP 205 8
@ te ° e 0.900% D0 oO
se) a 2 © rs @ Ts & COS S
O ° e%& e@ ) 9G 00 o O °
oe 9 F te o a © 8. om
Pn Ss Sap 2 0
-2- ae @ ee ou Oe” oS
| Siac alia tack Pokoe
| 6 ® O Q, © 9
I a ny
| ae =
-3- °2 e ©.6 eo

fs

“7 i): i= Flees 27 ge Bas ee” eed | es ey ee on: Ss | i |
Can 1

Figure 2. Plot of the first two canonical variate axes on seed characters of 150 individuals each from Luzula
campestris, L. multifiora subsp. multiflora and L. congesta.

264

J. V. ARMSTRONG

TABLE 4. CHARACTERS OF VALUE IN THE DISCRIMINATION OF LUZULA CAMPESTRIS,
L. MULTIFLORA SUBSP. MULTIFLORA AND L. CONGESTA
All the following characters are significantly different (p<0-05) between the appropriate pairs of taxa; the most

L. campestris/L. multiflora

Anther length*
Anther/filament ratio*
Capsule length*

Guard cell length*

Leaf apex*

Leaf pubescence

Leaf thickness

Leaf width

No. of flowers in cluster
No. of floral clusters
Overall height

Perianth segment width
Rhizomes*

Seed breadth

Seed length/breadth ratio*
Seed weight*

1550+
1450
1350
1250
1150
1050
950
850
750
650
550
450

Fruit index

diagnostic characters are indicated by an asterisk.

L. campestris/L. congesta

L. multiflora/L. congesta

Anther length* Capsule length
Anther length/filament length ratio* Inflorescence*
Capsule length Overall height
Height Seed breadth*
Inflorescence* Seed length
Overall height Seed size
Perianth segment width Seed weight*
Rhizomes*
Seed length
Seed length/breadth ratio*
Seed size
O
0, get lle
. fo)
a Mo Ores 2
° f, °B 40 0
°o 90 6 O°
oe O
Co 6.9
ao. OF ae
Ase 00 @moO
ome’ -
+ a +
teth+
+t es 4t
+t +
+ HF + + +
Eth, Tea i ~
fF age :
Foy gat ce ++
ce outage gale
++ ote
Mie os ae

Seed weight

Ficure 3. Bivariate scatter plot of fruit index (capsule length x perianth width) against seed weight (mg) for
Luzula campestris, L. multiflora subsp. multiflora and L. congesta.

LUZULA IN THE BRITISH ISLES 265

characters (and one perianth character) clearly separated the three taxa. Specimens of L. campestris
were separated out from L. multiflora subsp. multiflora and L. congesta by seed length/breadth ratio
against capsule length, seed length/breadth ratio against perianth width, fruit index against seed
breadth, fruit index against seed weight, seed length/breadth ratio against seed weight and fruit
index against seed length. Character combinations separating L. multiflora subsp. multiflora and L.
congesta included seed length/breadth ratio against perianth width and seed length against seed
weight. Characters providing good separation of all three taxa included seed length/breadth against
seed weight, fruit index against seed length and fruit index against seed weight (Fig. 3). In many
plots, a few of the outlier specimens were observed, but most specimens fell clearly within one of the
three taxa.

Species polygons were constructed and also supported the value of such character combinations
and the distinctness of the three taxa (Fig. 4). The characters of highest taxonomic potential for
discriminating L. campestris from L. multiflora subsp. multiflora and L. congesta are fruit index,
seed length/breadth ratio, anther length/filament length ratio and, to a lesser extent, capsule
segment length and perianth width. Those of value for differentiating between L. multiflora subsp.
multifiora and L. congesta are seed weight and seed length. Whilst the latter two taxa show overall
similarity in their character constellations, L. congesta has intermediate ranges for some of its
characters and some character states which are indistinguishable from L. campestris.

Character variation was found to be least in the characters of seed length, perianth width and seed
size, greater in seed weight, anther length (particularly for L. campestris), capsule length and seed
breadth (particularly so for L. congesta) and greatest in seed caruncle length. The latter proved to
be highly environmentally plastic. L. campestris was found to be more variable than L. multiflora
subsp. multiflora or L. congesta in a number of characters, including stamen dimensions and
structure of the inflorescence, but not in leaf width. This situation contrasts with the findings of
Nordenskidld (1956).

By the appropriate selection of characters, summed totals were also obtained for character
combinations which maximally differentiated the three taxa. This enabled all test specimens to be
assigned to one of the three taxa based on the ‘nearness’ of the summed class numbers to one of the
critical values for each taxon. The discriminatory value of the fruiting stage was found to be higher
than that of anthesis, enabling fruiting specimens to be more readily and confidently assigned (see
Table 2). Some specimens were more readily assignable than others; the difficult ones possessed
intermediate dimensions or apparently anomalous character combinations. The character table
statistics are presented in Appendix 1.

KEY TO TAXA

1 Plants loosely tufted or more usually forming a rosette, rhizomatous, leaf apex truncate;
inflorescence loosely cymose, nodding, up to 26 cm long, more usually 15-20 cm; anther length/
filament length ratio >2-5, seed globular with a length:breadth ratio <1-3 ...... L. campestris

1 Plants densely tufted, non-rhizomatous, leaf apex rounded; inflorescence subumbellate or
congested, >20 cm long, and more usually 30—40 (—85) cm; anther length/filament length ratio
<2-5; seeds oblong-ovoid with a length:breadth ratio usually >1:3 0.0... eee 2

2 Inflorescence congested, up to 85 cm long; seed breadth >0-86(—1-4) mm; seed weight usually
BUG MIAC Mire mince ve sciu iat Yb ui do onl. cortaa ute Maid Lolcales «okra Hastie ty ole se saesid eds eee cindy ob eiotelenici’ L. congesta
2 Inflorescence subumbellate, up to 60 cm long; seed breadth <0-86 mm; seed weight
AU) OITA Day eictaeevae tent alli, OU uae AD 4h a MN chehe al Get ULE cae «cid amechs wb alte a's L. multiflora subsp. multiflora

Of the other related taxa, L. pallidula is distinguished from L. multiflora subsp. multiflora by the
pale-yellow brown colour of its perianth segments, their smaller size (1-5—2-5 mm) (Stace 1991), the
unequal length of the inner and outer segments and the densely papillose peduncles (Kirschner &
Rich 1996). L. multiflora subsp. hibernica is distinguished by some of its peduncles being recurved,
whilst others are flexuous and the leaves usually less than 3 mm wide (Kirschner & Rich 1996). The
reported difference in capsule segment length for subsp. hibernica (1-9-2:2 mm) overlaps slightly
with that of subsp. multiflora ((1-8-) 2:2 + 0-20 (-2:6)).

266 J. V. ARMSTRONG

Key Diagram

Anther: filament ratio

Seed length <-" : ee .Capsule segment length
e —) .
Ww
z >
% os
Seed L:B ratio L - +Fruit index
“9g

Seed breadth ~ __.-’ Perianth segment width

Seed weight

L. campestris L. multiflora

L. congesta

Ficure 4. Character polygons of Luzula campestris, L. multiflora subsp. multiflora and L. congesta with key
diagram showing the characters measured.

LUZULA IN THE BRITISH ISLES 267
SPECIES DESCRIPTIONS

L. campestris (L.) DC. in Lam. & DC., Fl. Fr., 3rd ed. 3: 161 (1805). Field Woodrush. I/lustrations:
Butcher, A new illustrated British flora Wl: 673 (1961); Fitch & Smith, [lustrations of the British flora:
262 (1901); Ross-Craig, Drawings of British plants 30: pl. 30 (1973).

A compact, rosette-habit perennial with shortly creeping rhizomes. Height in flower 5—23 cm. Basal
leaves 2-5 mm wide, 0-15—0-20 mm thick, linear with small truncate swelling at apex, light yellow
green, shiny, thinly covered with long colourless hairs which are especially concentrated towards
base. Inflorescence loosely cymose or rarely slightly congested, usually nodding, consisting of a
panicle of one sessile and 2-7 pedunculate, spherical-obovate clusters of 3-12 flowers, each
subtended by 2 or 3 bracteoles. Branches of inflorescence + curved, reflexed in fruit. Bracts 1-3
mm, shorter to longer than the inflorescence; leaf-like. Perianth segments 3-1-3-6 x (0-8-) 1-0-1-3
mm, ovate-lanceolate, chestnut-reddish-brown with pale margins, subequal or the inner somewhat
shorter. Anthers 1-0-3-2 mm in length, 3-9 x as long as filaments. Style longer than ovary, filiform.
Capsule 2-6—2-9 (—3-6) mm, cylindrical to obovoid, apiculate, dark reddish-brown, glabrous, shiny,
shorter or as long as perianth segments. Seed 0-7—1-3 x 0-7—1-1 mm, nearly globose, 0-4-1-0 mg in
weight, with an ovate-triangular, flattened or trigonous basal appendage (caruncle) up to half as
long as seed. 2n = 12.

Common in grassland, meadows, dry pastures and sand dunes, usually on shallow soils of low
organic content, though known to persist in relatively productive artificial habitats such as lawns (see
Grime et al. 1988). Tolerant of a range of soil pH (4-9), although mostly found on acidic soils. Has a
wide altitudinal range up to 1010 m (Grime et al. 1988). Survives optimally under grazing regimes.

L. campestris is distinguished from L. multiflora subsp. multiflora and L. congesta by its rosulate
habit, presence of rhizomes, the truncate apical swelling of the leaves, an anther length/filament
length ratio of 3-9, style longer than the ovary, capsule length greater than 2-6 mm, and globose
shaped seeds.

L. multiflora subsp. multiflora (Ehrh.) Lej., Fl. Spa. 1: 169 (1811). Heath Woodrush. J/lustrations:
Butcher, A new illustrated British flora I: 674 (1961); Ross-Craig, Drawings of British plants 390: pl.
31 (1973).

Erect, densely tufted perennial with no rhizomes. Height in flower (15—) 30-55 cm. Basal leaves 34
(-7) mm wide, 0-20—0-28 mm thick, linear with small rounded swelling at apex, dark green, dull,
sparsely covered with long colourless hairs which are concentrated towards the base. Inflorescence
subumbellate with 1 (—3) subsessile and (3—) 6-13 (16) pedunculate, ovate to elongate clusters of
(6—) 9-15 (-18) flowers, each subtended by 3 bracteoles. Branches of infloresence usually erect in
fruit, up to 4 cm long. Bracts (2—) 3-7 mm, shorter to longer than the inflorescence, leaf-like.
Perianth segments 2-5—3-2 (—3-6) X 0-5—0-8 mm, broadly lanceolate, reddish-brown to pale brown
with hyaline margins, subequal. Anthers 0-7—1-6 mm in length, 1-3 X as long as filaments. Style as
long as ovary, filiform. Capsule 1-8-2-2 (—2:8) mm, obovoid to globose, apiculate, pale brown,
glabrous, shiny, usually shorter than perianth segments. Seed 0-8-1-3 x 0-5—0-9 mm, oblong-ovoid,
about twice as long as wide, 0-2—0-4 mg in weight, with ovate-triangular, flattened or trigonous basal
appendage up to half as long as seed. 2n = 36.

Common in moist woods and moors mainly on acidic to neutral soils (pH range 4-7), of high
organic and water content. Has a low altitudinal range, ascending up to 210 m.

L. multiflora subsp. multiflora is characterised by its tufted habit, absence of rhizomes, the
rounded apical swelling of the leaf, subumbellate inflorescence with bract shorter than inflores-
cence, anther length/filament length ratio usually less than 2, seed oblong-ovoid. It is distinguished
from L. campestris as described above and from L. congesta by its stalked infloresence, its perianth
segments being about the same length as the capsule values, and the smaller size of its seeds (in
terms of breadth and weight).

The description given here applies to the hexaploid cytotype of L. multiflora subsp. multiflora.
Tetraploid cytotypes are more commonly associated with alpine habitats. However, recently,
Kirschner & Rich (1996) have reported a tetraploid from Ireland, designated subsp. hibernica.

L. congesta (Thuill.) Lej., FJ. Spa. 1: 169 (1811). Congested Woodrush. —
Erect, densely tufted or single-stemmed stiff tillered perennial with no rhizomes. Height in flower

268 J. V. ARMSTRONG

(20—) 35-84 cm. Basal leaves 3-6 mm wide, linear with a small rounded swelling at apex, bright
green, sparsely covered with long colourless hairs. Inflorescence congested, rounded or lobed, often
more than 1 cm in diameter, of 1-4 congested subsessile or sessile clusters of (28—) 50-100 flowers.
Bracts (1—) 2-5 mm, longer than inflorescence, leaf-like. Perianth segments 2:5-3-5 x 0-5-1-:0 mm,
broadly lanceolate, reddish-brown, with broad hyaline margins, subequal. Anthers 1-5-1-7 mm,
1-5-3 X as long as filaments. Capsule 1-8-2-8 mm, obovoid to globose, apiculate, reddish brown,
glabrous, shiny, shorter than perianth segments. Seed 1-1-1-4 x 0-6-1-0 mm, oblong/ovoid, about
twice as long as wide, 0-4-1-2 mg, with ovate-triangular, flattened or trigonous basal appendage up
to half as long as seed. 2n = 48.

Common on peaty soils in moist meadows, moors and bogs, with a broader tolerance range than
L. multiflora subsp. multiflora and L. campestris, though intermediate in its requirements for pH (5—
8) and organic content. Found in habitats associated with L. campestris i.e. the wetter areas of sheep
grazed turf (cf. Kirschner 1991), where it is usually smaller and less tufted, and also together with L.
multiflora subsp. multiflora. Has a broad altitudinal range, up to 500 m.

L. congesta is characterised by its large size up to 84 cm, congested inflorescence with a much
longer bract, an anther length/filament length ratio of 1-5—2-0 (-3-0), perianth segments much
longer than capsule values, the latter less than 2-8 mm long, and oblong/ovoid seeds with a mean
length/breadth ratio of 1-5, and a mean seed weight of 0-66 mg.

DISCUSSION

TAXON STATUS

From the results obtained, it is evident that the taxa are well discriminated (supported by Wilk’s
Lambda) and that these can be separated on the basis of a few selected floral or fruit characters
(Tukey LSD tests). Furthermore, the nature of the variation is such that whilst L. congesta exhibits
character patterns most similar to L. multiflora subsp. multiflora, it is well circumscribed
morphologically and has morphological characteristics and ecological preferences (see descriptions)
which are frequently intermediate between L. campestris and L. multiflora subsp. multiflora,
together with some that clearly place it with L. campestris (e.g. anther length, seed length, seed
weight). On the basis of this pattern of variation it is maintained that the designation of L. congesta
to the specific rank be supported, an opinion shared by Kirschner (1990).

CHARACTER ASSESSMENT

L. campestris/L. multiflora subsp. multiflora.

Characters traditionally used to discriminate between L. campestris and L. multiflora subsp.
multiflora, include habit, height, number of floral clusters in the inflorescence, the nodding or
straight inflorescence branches, relative length of the anther to the filament, the relative length of
the style to the ovary, the relative length of the perianth to the capsule, seed shape and the presence
of rhizomes (see Chrtek & Kiisa 1980; Wigginton & Graham 1981; Clapham, Tutin & Moore 1987;
Stace 1991). Only some of these characters were found to be consistently reliable, whilst the others
were of use only at their extreme values. For material at anthesis, anther length and anther length/
filament length ratio were found to be the most diagnostic, whilst at the fruiting stage, capsule length
and seed shape were the most discriminatory. However, there is a small amount of overlap in both
anther length and capsule length. The latter is usually expressed in Floras as a relative measure. The
nodding tendency of the inflorescences of L. campestris is also a good diagnostic character. Of the
vegetative characters, the morphology of the leaf apex (truncate versus rounded) and the presence
of rhizomes in L. campestris have proved diagnostic, although the former needs to be assessed by
comparison with standards of the two species to reduce subjectivity.

Of the other potential discriminating characters, these have proved to be of value only at the
extreme values of each taxon. For example, the number of floral clusters and number of flowers
within a cluster showed much more overlap than formerly acknowledged. The former character was
only useful for specimens with eight or more floral clusters and twelve or more flowers within a
cluster, in which case they belong to L. multiflora subsp. multiflora. However, specimens with
stamen and fruit characters of L. multiflora subsp. multiflora, have been found with as few as four
floral clusters. Similarly, height is often cited, with L. multiflora subsp. multiflora and L. congesta

LUZULA IN THE BRITISH ISLES 269

given as being taller than L. campestris. Whilst specimens over 25 cm are usually either L. multiflora
subsp. multiflora or L. congesta, height is environmentally plastic with considerable overlap
occurring between the taxa. Similar tendencies (i.e. greatest values associated with L. multiflora
subsp. multiflora) are also found for leaf width, leaf pubescence and leaf thickness.

In addition to the above characters, three new characters exhibited high diagnostic potential,
being constant and with little or no overlap in their character ranges. These are capsule width, seed
weight and guard cell length (see Tables 1 & 4), of which the latter character is correlated with the
ploidy level. The first two characters are readily quantifiable on both field and herbarium material
and their use is highly recommended. Guard cell length is less useful since the guard cells shrink to
about half their fresh size on drying.

L. campestris/L. congesta

Of the characters usually cited for discriminating between L. campestris and L. congesta, the best
were found to be anther length/filament length ratio and seed length/breadth ratio, the latter also
being recognised by Wigginton & Graham (1981). The other characters of value usually showed
some overlap in their ranges amongst the two species. L. campestris and L. congesta are most
commonly distinguished on the basis of inflorescence shape, the former having at least two
pedunculate clusters, the latter a much larger congested head. However, it should be noted that L.
campestris is variable in this character, and may have either a lax or rarely a small and slightly
congested inflorescence.

L. multiflora subsp. multiflora/L. congesta

In Britain, L. multiflora subsp. multiflora and L. congesta have traditionally been distinguished on
the characters of inflorescence and relative capsule length (Clapham, Tutin & Moore 1987; Stace
1991). Kirschner (1990) also found the character of seed shape to be useful, and Kay (1993) cited
anther length/filament length ratio and caruncle length as discriminatory. All these characters have
been found to be useful (with the exception of caruncle length), with inflorescence type, seed
breadth and seed weight being the best, the latter formerly unreported. Capsule length is also
useful, although there is some overlap in character range between the taxa. Perianth length proved
to be of little diagnostic value when used on its own, but it is useful as a relative measure.

Using the above diagnostic characters, little difficulty should be encountered in the identification
of the majority of specimens. However, where a specimen is not easily identifiable, resort should be
made to the summed critical values. Previously unidentifiable specimens (at either the flowering or
fruiting stage) can successfully be assigned to one of the three taxa through the combined use of
several characters, using the procedure outlined. The procedure will be particularly advantageous
to those with little or no knowledge of the genus, and further offers the possibility of calculating the
probability of correct identification (see Whitehead 1954). In addition, if information is absent for
particular diagnostic characters, a re-summing of the class number can be made to take account of
this absence. However, the omission of many characters is likely to reduce the size of the numerical
gaps between the taxa, making identification more prone to error.

CONCLUSIONS

Considerable variation was found in the majority of the morphological characters previously used in
discrimination such that their value in taxon diagnosis and identification was sometimes limited.
Despite this, it was still possible to identify 18 characters which, when used in various combinations,
allow the consistent separation of the taxa. Such characters expressed the greatest interspecific:
intraspecific variance ratios. Floral and fruit characters proved to be the most reliable, particularly
the dimensions of the stamens, seeds and capsules and perianth width. It was found that no one
character, either primary or derived, consistently allowed unequivocal identification of all
specimens. In cases of doubt or difficulty, it is recommended that resort should be made to the
summed critical value tables. It is further argued that L. congesta is distinctive in its morphology and
ecology and, as such, warrants specific status. ;

270 J. V. ARMSTRONG
ACKNOWLEDGMENTS

The work was prepared for publication whilst the author was at the University of Glasgow, on
NERC grant GT/17/93/3. I am indebted to Dr G. H. Halliday (University of Lancaster) for his
advice and encouragement throughout, and Dr R. R. Mill (Royal Botanic Garden, Edinburgh) for
reading and commenting on the manuscript. Thanks are also due to Dr K. G. Johnson (University
of Glasgow), Dr A. Malloch (University of Lancaster), Miss S. L. Armstrong and the anonymous
reviewer.

REFERENCES

BucHANaN, J. (1960). Luzula campestris (L.) DC. Proceedings of the Linnean Society of London 171: 126-128.

BUCHENAU, F. (1890). Monographia Juncacearum. Botanische Jahrbiicher (Leipzig) 12: 1-495.

BUCHENAU, F. (1906). Juncaceae, in ENGLER, A., ed., Das Pflanzenreich 25: 1-284.

CueteEK, J. & KRisa, B. (1980). Luzula DC., in Tutin, T. G. etal., eds., Flora Europaea 5: 111-116. Cambridge
University Press, Cambridge.

CLAPHAM, A. R., Tutin, T. G. & Moore, D. M. (1987). Flora of the British Isles, 3rd ed. Cambridge University
Press, Cambridge.

FERNALD, M. L. (1945). Notes on Eastern American Luzula. Rhodora 47: 265-271.

GriME, J. P., Hopcson, J. E. & Hunt R. (1988). Comparative plant ecology. Unwin & Hyman, London.

Kay, G. M. (1993). Novel concept of Luzula multiflora and L. congesta. B.S.B.I. news 63: 14.

KIRSCHNER, J. (1980). A new species of the Luzula campestris-multiflora complex in central Europe. Folia
geobotanica et phytotaxonomica 14: 431-435.

KIRSCHNER, J. (1990). Luzula multiflora and allied species (Juncaceae): a nomenclatural study. Taxon 39: 106—
114.

KIRSCHNER, J. (1991). An account of natural hybridisation within Luzula sect. Luzula (Juncaceae) in Europe.
Preslia 63: 81-112.

KIRSCHNER, J. & KRisa, B. (1979). Notes on the taxonomy and cytology of the genus Luzula in West Caucasus,
Czechoslovakia. Preslia 51: 333-339.

KIRSCHNER, J. & Ricu, T. C. G. (1996). Luzula multiflora subsp. hibernica, a new tetraploid taxon of Luzula
sect. Luzula (Juncaceae) from Ireland. Watsonia 21: 89-97.

NORDENSKIOLD, H. (1956). Cyto-taxonomical studies in the genus Luzula. II. bisects On experiments in the
campestris-multifiora complex. Hereditas 42: 7-73.

Novikov, V. S. (1990). Konspekt sistemy roda Luzula DC. (Juncaceae). [An outline of the system of the genus
Luzula.| Byulleten Moskovskogo Obsnchzestva Ispytateley Prirody, Moskva 95: 63-70.

Stace, C. A. (1991). New Flora of the British Isles. Cambridge University Press, Cambridge.

Voss, E. G. (1966). Nomenclatural notes on monocots. Rhodora 68: 435-463.

WHITEHEAD, F. H. (1954). An example of taxonomic discrimination by biometric methods. New phytologist 53:
496-510.

WIGGINTON, M. J. & GRAHAM, G. C. (1981). A guide to the identification of some difficult vascular plant species.
Nature Conservancy Council, London.

(Accepted August 1996)

APPENDIX 1 — CLASS NUMBER STATISTICS USED IN THE CONSTRUCTION OF CRITICAL VALUES

‘000 100 = 9-380 increments = 0-083
-510 100 = 4-580 increments = 0-041
‘658 100 = 3-235 increments = 0-026
-800 100 = 3-600 increments = 0-018

Anther length/filament length ratio 1
0
0
1
2:00 100=7-00 increments = 0-05
0
0
1
0

Anther length/filament length ratio (dehisced)
Anther length (mm)

Capsule length (mm)

Leaf width (max.) (cm)

Perianth width (mm)

Seed breadth (mm)

Seed length/breadth ratio

Seed weight (mg)

‘474 100 = 1-315 increments = 0-008
-500 100 = 1-105 increments = 0-006
‘000 100 = 2-620 increments = 0-016
-180 100 = 1-200 increments = 0-010

Soooococ.co >
eee Ue Ne I eal is att

Watsonia 21: 277-281 (1997) 57]

Calluna vulgaris (L.) Hull dieback in Orkney, Scotland

A. H. KIRKPATRICK
Department of Environmental Science, University of Stirling, Stirling, FK9 4LA
and
A. J. MACDONALD

Scottish Natural Heritage, 2 Anderson Place, Edinburgh, EH6 5NP

ABSTRACT

In recent years there have been occasional sightings of extensive Calluna vulgaris (L.) Hull dieback in
communities of high nature conservation interest in Orkney (v.c. 111). The results of a systematic survey of dead
stands are reported and the ability of Calluna to recover investigated.

KEyworbs: heather dieback, Winter Moth, heather regeneration, conservation, moorland management.

INTRODUCTION

The Calluna vulgaris (L.) Hull (Ericaceae) (Ling, Heather) dominated communities of Orkney (v.c.
111) are of outstanding natural heritage interest providing breeding areas for Red-Throated Diver,
Merlin, Golden Plover and Short-Eared Owl. All are listed in the E. U. Directive on the conservation
of wild birds and the first two species are also protected by Schedule 1 of the Wildlife and
Countryside Act, 1981. In Orkney areas of tall herb vegetation are found on open moorland yet in
the rest of Scotland this vegetation is largely restricted to cliff ledges. In recent years regular
muirburning has not been a feature of Orkney’s moors and there are extensive areas of tall
unmanaged Calluna which provide nest sites for one of the highest concentrations of breeding Hen
Harriers in Britain (Batten et al. 1990). Thus the maintenance of areas of older heather is an
important factor in retaining the natural heritage interest of the moors. Within the last decade there
have been reports of extensive dieback in such stands ranging from one to 20 ha in size (Picozzi
1981). In the summer of 1993 the localities of all known outbreaks were mapped as part of the
Scottish Natural Heritage funded project ‘Moorland Audit and Management in the Northern
Isles”. This included some areas previously identified by staff of Scottish Natural Heritage and the
Royal Society for the Protection of Birds.

METHODS

Data from the Scottish Office Environment Department/Macaulay Land Use Research Institute
project ‘““The Land Cover of Scotland” were used to define the present extent of moorland on
Orkney and a systematic search, using binoculars, was carried out during July and August 1993 on
the islands of Mainland, Hoy, Rousay and Eday. Areas of dieback were highly visible from some
distance and were confirmed by closer inspection. Locations were marked on 1:25,000 O.S. maps
and transferred to a Geographic Information System (GIS). The distribution of dieback (Fig. 1) was
compared with vegetation types from the Land Cover of Scotland data. Picozzi (1981) noted patches
of heather dieback on Orkney and concluded that damage was caused by very high numbers of the
larvae of Winter Moth (Operophtera brumata) and Dotted Border (Agriopis marginaria). As
damage by Lepidopteran larvae can only be confirmed at a restricted period of time only sites at
which larvae have been found were marked as such in Fig. 1 though all stands with dieback were

DAZ A. H. KIRKPATRICK AND A. J. MACDONALD

Location of dieback

Lepidoptera larvae
confirmed as cause
of dieback

" South
Ronaldsay

FicureE 1. The distribution of stands of Calluna vulgaris dieback on Orkney (v.c. 111).

searched by walking through them looking for the larvae. The ability of Calluna to recover was
subsequently investigated at a site at Queenamidda, West Mainland (HY/373.213) where dieback
had occurred in 1980.

Using a random walk 100 8 cm X 8 cm quadrats were placed in the defoliated stand which was
around 0-1 ha and the frequency of the following types of regeneration noted:

a. regrowth from near the tip of branches which had not been lethally damaged;

b. regeneration from the stem base such as occurs after muirburning (Miller 1980); and

c. new growth from “‘layering’’ stems buried in the litter layer (Beijerinck 1940; Scandrett &
Gimingham 1989; MacDonald et al. 1995).

RESULTS

The locations of all known areas of dieback are shown in Fig. 1 and their relationship with
communities derived from the Land Cover Data is summarised in Tables 1 & 2. Dieback had been
recorded throughout the spectrum of dry to wet Calluna dominated communities though none had
occurred in heather/grassland mosaics.

CALLUNA VULGARIS DIEBACK IN ORKNEY, SCOTLAND

ZS

TABLE 1. OUTBREAKS OF CALLUNA VULGARIS DIEBACK ON ORKNEY IN RELATION TO THE
LAND COVER OF SCOTTISH MOORLAND COMMUNITIES (MONTANE VEGETATION AND
INDUSTRIAL PEAT WORKINGS HAVE NOT BEEN INCLUDED)

Land Cover Codes

Vegetation

Number of stands
with Calluna dieback

110,114 Dry heather moor 3
130,134 Undifferentiated heather moor 4
120 Wet heather moor 1
180,182 Blanket bog 4
186 Areas of peat workings 3
370 Mosaic of undifferentiated heather moor and peat workings 0
657 Mosaic of peat workings and blanket bog’ 0
301,326,437 ,496,656 Heather/grassland mosaics 0
938 Blanket bog/grassland mosaics 0
1135 Mosaic of wet heather moor and point water 0
184 Blanket bog with dubh lochans 2
— Boundary between undifferentiated heather moor and 2
blanket bog
Tai 1

Boundary between dry and undifferentiated heather moor

TABLE 2. A SELECTION INDEX OF STANDS OF DEAD CALLUNA IN RELATION TO THE LAND

COVER OF SCOTLAND COMMUNITIES

Vegetation

Dry heather moor

Undifferentiated heather moor

Wet heather moor

Blanket bog

Areas of peat workings

Mosaic of undifferentiated
heather moor and peat workings

Mosaic of peat workings and
blanket bog

Heather/grassland mosaics

Blanket bog/grassland mosaics

Mosaic of wet heather moor and
point water

Blanket bog with dubh lochans

Observed
no. of
stands with Expected no.
Calluna of stands if Selection
Area dieback random Index
(ha) (O) selection (E) (O-E/ZO)
1539 3-5 0-96 +0-14
13164 5-5 8-24 —0-15
1120 1 0-7 +0-02
4897 5 3-1 +0-11
5630 3 3-5 —0-03
66 0 0-04 -

83 0 0-04 ~
1823 0 1-08 —0-06
5 0 0-002 -

205 0 0-1 _
211 0 0-1 -

(The occurrences on the boundaries of two types in Table 1 have been incorporated into this analysis by splitting

equally between them.)

The frequencies of the different forms of regeneration at Queenamidda are shown in Fig. 2. Out
of the total 100 quadrats some form of regeneration was taking place in each one. Regeneration
from near the tip of branches was recorded in 39% of quadrats, from the stem base in 35% of
quadrats and from layering stems in 36% of quadrats. Regeneration was only from the tip of
branches in 21% of quadrats, only from the stem base in 26% of quadrats and only from layering

stems in 34%.

274 A. H. KIRKPATRICK AND A. J. MACDONALD

ay)
—

FREQUENCY

PER QUADRAT
SiN £2 o a of 0) =

1 2 3 4 5 6

No. OF BRANCHES SHOWING REGROWTH
FROM NEAR THE TIP

b)

FREQUENCY
PER QUADRAT
Shak Oo oOo % 5

1 2 3 4 5
No. OF REGENERATING STEM BASES

y)
—
FREQUENCY
PER QUADRAT
[o) ine) _ OD fee) Ps) rs) rs

1 2 3 4 5
No. OF LAYERING STEMS SHOWING REGROWTH

b
ie)

dq)

WAS RECORDED
ie) wo
oO [e)

_.
oO

TOTAL OUT OF 100 QUADRATS IN

WHICH THIS FORM OF REGENERATION
WAS RECORDED
= ho Ww cS
[o) je) [o) oO fo)
i)
—
TOTAL OUT OF 100 QUADRATS IN

WHICH ONLY ONE FORM OF REGENERATION

fo)

Tip of Stem Layering

Tip of Stem Layering
branches __ bases stems

branches bases stems

FIGURE 2. Frequency of regeneration of Calluna after defoliation by Lepidoptera from: a. regrowth near the tip;
b. the stem base; c. layering stems; and d. the total number of quadrats in which a particular form of regeneration
was recorded with e. total quadrats in which only one form of regeneration was recorded for 100 8 cm X 8 cm
quadrats at Queenamidda, Orkney.

CALLUNA VULGARIS DIEBACK IN ORKNEY, SCOTLAND pH

N

DISCUSSION AND CONCLUSIONS

Current data on Calluna dieback indicates that it can occur in both heather moorland and blanket
bog though none was recorded in heather/grass mosaics of more heavily grazed areas.

Within recent years damage to Calluna stands by Winter Moth has been recorded from a range of
sites in Scotland and recovery is often poor especially if the Calluna is old or there is high grazing
pressure (Hartley & Kerslake 1995). The Orkney sites however contrast sharply with mainland
Scotland where outbreaks have been found in Calluna — Vaccinium myrtillus moorland under heavy
grazing pressure. In Orkney a wider range of Calluna communities have suffered from dieback and
grazing has been light. With the importance of these areas for breeding Hen Harriers these
outbreaks of dieback are a cause for concern.

The results in Fig. 2 indicate that recovery of Calluna cover proceeds very slowly so that if sheep
numbers were increased on such sites conversion to grassland of lower conservation interest is likely
to occur.

ACKNOWLEDGMENTS

The authors would like to thank the Orkney staff of Scottish Natural Heritage and the Royal Society
for the Protection of Birds for their co-operation and help; Louise Scott, Stirling University for
assistance in the field and John McArthur, Chief GIS Technician, Stirling. This work was funded by

Scottish Natural Heritage as part of the project ‘Moorland Audit and Management in the Northern
Isles’? (040/93/UPB).

REFERENCES

BaTIEN, L. A., Brissy, C. J., CLEMENT, P., Ettiotr, G. D. & Porter, R. F. (1990). Red data birds in Britain.
T.& A.D. Poyser, London.

BEWERINCK, W. (1940). Calluna. A monograph on the Scotch Heather N. V. Noord-Hollandsche Uitgevers
Maatschappij, Amsterdam.

Hartcey, S. E. & KEeRSLAKE, J. E. (1995). Winter moth outbreaks. Unpublished report to Scottish Natural
Heritage.

MacDonaLp, A. J., KIRKPATRICK, A. H., HESTER , A. H. & SyDEs, C. (1995). Regeneration by natural layering
of heather (Calluna vulgaris): frequency and characteristic in upland Britain. Journal of applied ecology 32:
85-99.

MILLER, G. R. (1980). The burning of heather moorland for Red Grouse. Bulletin d’ecologie 11: 725-733.

Picozzi, N. (1981). Common gull predation of Winter Moth larvae. Bird study 28: 68-69.

SCANDRETT, E. & GIMINGHAM, C. H. (1989). Vegetative regeneration by layering in Calluna vulgaris (L.) Hull.
Transactions of the Botanical Society of Edinburgh 45: 323-334.

(Accepted July 1996)

ub Sedo engi

now gaoelins
aes FIG Ca 7

oat blucweratta eon i.

igfhiie + HS oc) NOUIGET OF) 1Or
age LVF ches bloit ont arsomazalese

eA ‘hip ahs ¥ Lye 4 agains ea ders ans 2. a_i! :

: . Dae lal oa
pe ge aise Lunt ibs
ud. ai ae) Lhd suet hes isa et
shag SME paNE oD Ose oP 4S, elite radiate
A pid Sanaa) pHa ‘a Wheat. pha.
2 — ? Pe Wasco.
By ith EE aa es lays egies a aE EMRE Ae.

Watsonia 21: 283-298 (1997) 207

Notes

JUNCUS CAPITATUS WEIGEL (JUNCACEAE) REDISCOVERED NEAR ITS ORIGINAL
LOCALITY IN ANGLESEY (v.c. 52)

The dwarf rush, Juncus capitatus Weigel, is an annual with a predominantly southern distribution in
Europe. It is known from the Channel Isles, the Isles of Scilly, the Lizard peninsula and a few other
localities in Cornwall; the only other British records are from Anglesey where it is rare and elusive
and recorded only from a few places in or near sand-dune systems on the south-west coast.

The first record of J. capitatus in Anglesey was made by Bolton King in August 1918. It was
reported from near Rhosneigr “‘in good quantity over a limited area, but assuredly native, growing
with the usual damp heath vegetation” (Report of the Botanical Exchange Club 5: 402, 1918).
Specimens were sent by King to G. C. Druce and in an accompanying letter (dated 13 August 1918),
held with the material in OXF, he gave details of the locality at “‘the extreme northern edge of
Towyn-Trewan, close to the S. corner of the encircling wall round a rocky mound called Carnau
(?)”. A few years later, Druce was sent a further Anglesey specimen of J. capitatus (in OXF) by
Lady Kathleen Stanley who wrote on 25 June 1925 that “there was quite a lot of it at Tywyn Trewan
quite close to Rhosneigr’’. Although precise details are not available, the latter locality is unlikely to
be the same as King’s and is probably to the south-east in another part of the extensive sand-dune
and dune hinterland habitat at Tywyn Trewan..

In the 1940s, Tywyn Trewan was extensively modified by the construction of RAF Valley airfield,
and as this encroached up to the wall of the Carnau mound (Roberts 1982) it had been assumed that
J. capitatus was extinct at this locality. The site has been searched on various occasions in recent
years, but always without success. It was thus a welcome surprise when J. capitatus was found
nearby by one of us (T.H.B.) while recording bryophytes on 31 March 1995.

About 10-20 individuals were present on the sides of two adjacent small hollows within an old
vegetated blow-out, in a stand of dune heath situated to the south-west of Carnau farmhouse,
immediately to the west of the airfield perimeter fence. The deeper of the two hollows had a shallow
depth of water, but the other was above the water table. The J. capitatus plants were in patches of
very moist sand in a narrow zone above the base of the hollows, with only a very sparse cover of
associated species, including Carex flacca Schreb., Erica cinerea L., Lotus corniculatus L., Luzula
sp., Pedicularis (seedling), Sagina sp. and Salix repens L. The Calluna-Cladonia heath in the
surrounding blow-out was interspersed with open bryophyte-covered patches, but no further plants
of J. capitatus were detected in this or in other parts of the heath.

All the J. capitatus plants at this locality were already post-mature, brown and with few remains of
basal leaves by the end of March 1995; the inflorescences were still intact, but the capsules had all
dehisced and were empty. A few shoots were collected and have been deposited in NMW.

Elsewhere in Anglesey, there are records from two other dune systems. It was observed on a field
excursion of the Botanical Exchange Club in June 1937 from “‘S.W. side of Newborough Warren’”’
(Report of the Botanical Exchange Club 11: 49, 1937); no further details are available, and J.
capitatus has not been recorded again from this locality where its habitat may have been destroyed
by the establishment of a conifer plantation. It has also been reported from Tywyn Aberffraw by J.
G. Duckett and J. N. B. Milton, and a specimen (now in herb. A. J. Byfield) was gathered in August
1983 “‘on damp sand’ with two of the liverwort specialities of the Aberffraw dune system,
Petalophyllum ralfsii (Wils.) Nees & Gott. and Southbya tophacea (Spruce) Spruce. More recent
searches at this site by R. H. Roberts and others have been unsuccessful.

J. capitatus has thus been recorded only very sporadically and locally in Anglesey from three sand
dune systems. Viable seed of this species can remain dormant for at least 29 years (Coombe 1987).
and its reappearance near Carnau demonstrates that it can persist unnoticed for over 50 years.
Londo & van Leeuwen (1974) reported the appearance of J. capitatus at a new locality within a dune
system in the Waddendistrict of The Netherlands following the creation (by excavation) of

278 NOTES

depressions which were part-flooded in winter. A curious feature of the Anglesey records is their
seasonal range; J. capitatus (presumably always with inflorescences) has been observed in March,
June and August, suggesting variable phenological behaviour.

ACKNOWLEDGMENTS

We are very grateful to R. H. Roberts for details about Anglesey localities of J. capitatus, as well as
for comments on this note in draft. Thanks are also due to A. J. Byfield, Dr D. E. Coombe,
Professor J. G. Duckett and Gwynn Ellis for providing helpful information.

REFERENCES

CoomBE, D. E. (1987). Longevity of spores of some terrestrial Jsoetes species. B.S.3B.I. news 45: 38.

Lonpo, G. & VAN LEEUWEN, C. G. (1974). Het gedrag van Juncus mutabilis en J. capitatus in enkele
duingebieden van het Waddendistrict. Gorteria 7: 42-46.

Roserts, R. H. (1982). The flowering plants and ferns of Anglesey. National Museum of Wales, Cardiff.

T. H. BLACKsTOCK

Countryside Council for Wales, Ffordd Penrhos, Bangor, Gwynedd, LLS57 2LQ

R. A. JONES

Countryside Council for Wales, Ladywell House, Park Street, Newtown, Powys, SY15 1RD

RUBUS PANNOSUS P. J. MUELLER & WIRTGEN (ROSACEAE) IN BRITAIN

Watson (1958) described Rubus gravetii as ‘‘very rare; [v.c.] 15 (Bigbury to Chartham Hatch), [v.c.]
18, (Epping Forest, High Beach). Belgium, W. Germany.” No search has yet been made for the
plant in v.c. 15, E. Kent, but during the summer of 1995 A.L.B. studied intensively the Rubi of the
whole northern half of Epping Forest, S. Essex (v.c. 18), on behalf of the Superintendent of Epping
Forest (Corporation of London). The plant was already known to him from visits to the Forest in
1987 and subsequently. It had been determined by A.N. as “‘false gravetii’’ with the proviso that it
had not been compared critically with the Continental European plant since Watson’s time and
could possibly be the same. Watson’s Bigbury-Chartham (v.c. 15) specimen (SLBI) had, however,
been compared with the type of gravetii by Newton and E. S. Edees in 1978, when its identity was
rejected.

Interestingly, the “false gravetii’’ was met with only once in the northern part of the study area,
whereas the closely related Rubus fuscus Weihe was frequent and occurred at nine of the study sites.
In the southern and south-western part of the study area R. fuscus was only found once, but the
‘false graveti’’ was abundant and often dominant in eleven sites. At this time the plant was accepted
as being “‘probably a local endemic.”’

During August 1995, whilst revisiting the tetrads covered by the Norfolk B.S.B.I. recording
weekend in June to record the Rubi, A.L.B. discovered “‘false gravetii’” to be frequent on the
Sheringham Park estate at NGR TG/1.4.

As it had now been established that the plant was not an Epping Forest endemic, a specimen was
sent to Professor H. E. Weber (University of Osnabriick, Germany) with the request that if possible
he compare it with the Continental R. gravetii. In his reply, he stated that he and Dr G. Matzke-
Hajek (of Alfter, Germany) had independently examined the specimen and both had determined it
without hesitation as Rubus pannosus P. J. Mueller & Wirtgen: however he wished to see
photocopies of additional herbarium specimens. Among the photocopies sent to him were copies of
specimens from both the Epping Forest and Sheringham Park populations. All of these were
accepted by Weber as falling within the range of possible intraspecific variation.

After some discussion, A.N. offered to borrow specimens of R. gravetii and R. pannosus from
MANCH for comparison. The sheets examined were as follows:

Rubi praesertim Gallici exsiccati (1895) no. 39, coll. F. Gravet (an isolectotype of Rubus gravetii).

NOTES 279

Wirtgen Herb. Rubor. Rhen. ed. 2, fasc. 2 no. 77, Glandulosi Muell.: ““Bopparder Walde hinter
Waldesch bei Coblenz’’, 17. vii. 1859 (the holotype of R. pannosus).

G. Braun Herb. Rub. Germ. no. 134, coll. [Th.] Braeucker, ‘‘in der Rheinprovinz” (R. eifelensis
Wirtgen).

Wirtgen Herb. Rubor. Rhen. ed. 1, fasc. 4, no. 94 (R. eifelensis).

The last-named sheet above was laid aside as being distinctly different from the other three, which
were subjected to intense scrutiny. Syntypes of R. pannosus all belonging to set 77 had been sent
from BM and BR and were included in the comparisons. It was concluded that, despite minor
differences, chiefly with regard to leaf shape, all three belonged to the same species and that they
matched the Epping Forest plant well. Thus, while Watson was correct in equating the Epping
Forest plant with that distributed by Gravet in 1895, the name Rubus pannosus P. J. Mueller &
Wirtgen has priority. R. pannosus should therefore be admitted to the British list with a known
distribution of NGR: TL/4.9 and TL/4.0, Epping Forest (v.c. 18); TG/0.3, Holt Lowes (1972, herb.
A.L.B.); and TG/1.4, Sheringham Park (both v.c. 27, E. Norfolk).

We have also studied the defective holotype specimen (BR) and other specimens of Rubus
cinerascens Weihe ex Le}j. sent from Belgium by H. Vannerom, which he considers to be conspecific
with R. pannosus and R. gravetii, but this identity in our opinion cannot be sustained.

Rubus pannosus may be recognised by its villous stem, deep reddish-black in colour, the hairs
covering dense, short, blackish glands. The prickles are short-based, straight, somewhat slanting,
fairly many, and coloured like the stem. The leaves are digitate, softly pilose, especially on the veins
beneath, which are pectinately hairy. The terminal leaflet is ovate or elliptical, acuminate, the base
somewhat cordate or emarginate. The panicle is pyramidal, usually round-topped, the middle and
lower branches long and spreading, the pedicels exactly patent when well developed, armed and
coloured like the stem. The white-edged sepals may be loosely reflexed or patent, sometimes long
pointed. The petals are narrowly obovate, mid-pink; the filaments are pink and a little longer than
the deep red styles.

R. fuscus Weihe differs in the stem being not quite so densely hairy or glandular, but with fairly
numerous pricklets. The leaves are more abruptly and shortly acuminate, and the margins less
evenly serrate. The panicle of R. fuscus is longer and more symmetrically pyramidal than R.
pannosus, and the petals in the Epping Forest population of the former are typically several shades
darker than R. pannosus, often almost red.

ACKNOWLEDGMENTS

The loan of specimens from BR, BM and MANCH is gratefully acknowledged.

REFERENCE

Watson, W. C. R. (1958). Handbook of the Rubi of Great Britain and Ireland. Cambridge University Press,
Cambridge.

A. L. BULL

Hillcrest, East Tuddenham, Dereham, Norfolk, NR20 3JJ

A. NEWTON

10 The Fairways, Leamington Spa, Warwickshire, CV32 6PR

THE STATUS OF PULICARIA VULGARIS GAERTNER IN BRITAIN IN 1995

In 1990 the Hampshire Wildlife Trust undertook a national census of Pulicaria vulgaris Gaertner,
Small Fleabane (Asteraceae). All surviving populations were identified and surveyed. The
population was estimated at 10,000 plants. Of all the populations in Britain 98% were found in S.
Hants (v.c. 11), the majority in the New Forest with outliers in the Avon Valley. The remaining

280 NOTES

populations were found in the Thames Basin of N. Hants (v.c. 12) and Surrey (v.c. 17). A summary
of the 1990 survey results was published in Watsonia 18: 405-406.

Since 1990 there have been resurveys of selected populations by Wildlife Trust volunteers and a
collation of population studies pre-dating 1990. These studies, together with informal observations
of various sites, suggested that the plant may still be in decline in various localities. This general
concern prompted the 1995 survey.

The 1995 survey was undertaken by Lady Rosemary FitzGerald, assisted by Geoffrey Field,
Elizabeth Young, John Ounsted and Joyce Smith. The 1995 survey adopted the same methodology
as the 1990 survey. Small Fleabane was refound at all of the sites identified in 1990. The plant was
not refound in any sites whose populations were considered extinct by 1990, nor was it reported
from entirely new sites. Since 1990 additional populations have been found within the immediate
vicinity of known sites. Some of these extentions to populations identified were significant,
particularly those associated with farmyards and enclosed lands in S. Hants.

The British population of Small Fleabane in 1995 was estimated at some 28,000 plants. The
distribution of plants by vice-county was found to be very similar to the 1990 survey, with 98% of the
population in the Hampshire Basin, S. Hants. (v.c. 11), with 87% in the New Forest and 11% in the
Avon Valley. The Thames Basin supported 2% of the population, 1-5% in N. Hants. (v.c. 12) and
0.5% in Surrey (v.c. 17).

The 1995 populations were compared to the 1990 survey and a partial 1985 Nature Conservancy
Council survey, together with other records for the sites. Some well recorded sites have been
recorded up to 13 times over the last three decades.

Since the onset of detailed recording in 1985, only one area had entirely lost its population of
Small Fleabane in Britain. This was a population formerly known from English Nature’s Ashford
Hill National Nature Reserve in N. Hants. As a result of the 1995 survey remedial works to revive
the population have been adopted.

The population size in Britain as a whole has decreased by an order of magnitude om over
100,000 plants in 1985. This decline had occurred by 1990 and was not considered to be significant in
conservation terms (Chatters 1991). The national decline is predominantly accounted for by
changes in a single sub-site which still supports a healthy dynamic population.

Since 1990 four areas have experienced increases in their Small Fleabane population which are
considered to be significant. In two sites the popuiations increased by one order of magnitude and in
two other sites by two orders of magnitude. No sites have declined by one or more orders of
magnitude.

Whereas Small Fleabane is usually associated with grazed village greens and heathland edges, the
1995 study found it in enclosed farm lands and within non-intensively managed farm-yards. Historic
ecological studies revealed that these were former commons and greens or closely associated with
the same. The farmyard habitats have not been described previously and pose unusual challenges to
ensure continuity of management to conserve the population.

This survey would not have been possible without the financial assistance of English Nature’s
Species Recovery Programme and the Guinness Trust. We are also very grateful for the assistance
of the many landowners and managers on whose private property Small Fleabane grows.

REFERENCES

CHATTERS, C. (1991). The status of Pulicaria vulgaris Gaertner in Britain in 1990. Watsonia 18: 405-406.
FirzGERALD, R. et al. (1996). The status and conservation of Small Fleabane in the United Kingdom 1995.
Unpublished report to Hampshire Wildlife Trust, Eastleigh, Hampshire.

R. FitzGERALD

Beggars Roost, Lilstock, Nr Bridgewater, Somerset, TA5 ISU

G. D. FIELD

37 Milton Grove, New Milton, Hampshire, BH25 6HB

C. CHATTERS

Hampshire Wiidlife Trust, 8 Romsey Rd, Eastleigh, Hampshire, SO50 9AL

NOTES 281
A WOLLEY-DOD LETTER OF HISTORICAL INTEREST

The following letter from A. H. Wolley-Dod to the Rev. H. J. Riddelsdell was found in one of the
folders of the BM British Rosa collection. It is reproduced here exactly as Wolley-Dod typed it,
apart from the correction of a few obvious typing errors.
Glenrinnes,
Tadworth, Surrey.
18th Oct. 1920
Dear Mr. Riddelsdell,

I have not been many days over your roses, but that is not much indication of the number of hours
spent. I have done little else since they came. My motives were partly selfish, because I had just
started on a parcel for Dingler when they arrived. I fear I shan’t get much out of him, and that not for
some months. He finds British Roses very difficult and grudges the time they take.

As you anticipate, my criticism of your plants in general is that you have collected too
indiscriminately. It is easy to go out and cut a specimen from every bush you see, but it is quite
another matter to name them. Personally I pass over 50% of the bushes I see as unnameable, and
only collect those which are in good condition and which present some marked feature. This of
course is shirking my responsibilities but the difficulties are greater than I can face. Another result of
indiscriminate collecting is the number of bad specimens; I mean those which are badly grown or in
some way abnormal. In so difficult a genus only the best grown material gives a critic a chance, that
is the stuff must be characteristic of the bush it is cut from, and that bush should be a well grown one.
Someone once said that the average herbarium is a collection of freaks, meaning that botanists have
a tendency to gather unusual looking specimens, just to show how species vary. That is all very well
in general where the species and varieties are more or less easily recognisable, but it will not do in
Rosa. .

Another minor point, which concerns your own herbarium more than my determinations is the
awkward shape of some of the specimens owing to your having cut pieces with the old stem not in the
same plane as the flowering branches. Old stem is often desirable, since the prickles on the flowering
shoots are often not characteristic of the specimen, but you can almost always get these of a less
awkward shape. Some of your worst bits have not even prickles on them so that they are useless
encumbrances.

One moye suggestion is that you try and arrange your gatherings in Groups. No doubt you will
make mistakes. I do so myself, but it is very helpful to have all of one affinity together, if only to
facilitate reference to my own herbarium. Moreover the chances of my giving the same name to
different varieties or conversely, is reduced.

I hope you won’t mind these criticisms, but I hope they may help you as well as me.

I have got quite a lot of N.C.R.s for Gloster, but very few for Oxon as Druce has skinned that
V.C. fairly closely. I am surprised at the absence or rarity of several of what I had regarded as our
commoner forms e.g. viridicata, stenocarpa, adscita, andegavensis, urbica, trichoneura, jactata,
Gabrielis, (the last 3 I think totally absent, but all very near urbica and semiglabra). Typical
dumetorum is also absent, but I believe that to be rare. All Deseglisei, not a very common Group,
also do not appear nor do Villosae (except one abnormal plant) and Rubiginosae. I am not surprised
at seeing no Glaucae or Coriifoliae, as they thin out greatly as you go S.E., still Gloster is hardly
S.E., nor is Oxon.

Do not be surprised at my nomenclature. I think in the last lot I did for you I followed that I had
adopted in my last paper, but Iam contemplating changes, not so much in names as in combinations,
e.g. I shall probably drop R. lutetiana as a type, i.e. species name, and go back to R. canina. I have
therefore in the main given you the names in the form their authors wrote them.

I am not sure that I am not wrong in attempting to name more than above 30% of what I see. The
more I see of the genus the more new combinations of characters appear. Doubtless many of these
have names, but they are unknown to me. I have at least 100 names from Sudre and Dingler which I
do not use, since the specimens seem to me to be too near to other better known ones, or are
probably those of individual bushes which I should not adopt unless they presented very striking
features. I am getting stronger in the opinion that Roses are classified on wrong lines. They are much
over-split, and the Subgroups, and even some Groups are based upon purely artificial characters, so
that plants which have great natural affinities get widely separated. I believe we may have species

282 NOTES

which may be glabrous or hairy, uniserrate or biserrate, glandular or not, with variations in the fruit,
sepals and style. So you may say we have if I adopt R. caninae as a large aggregate species, but I
mean its varieties should be associated on different grounds, abolishing most of its Subgroups,
though how to do it beats me. If only I could find someone in England who knows roses well enough
to discuss these matters, I should feel happier. As it is I fear I shall develop into a dictator, with no
one able, if I may say so, to contradict me. The result may be disastrous. Still former rhodologists
have differed so widely in their treatment of the genus that I may perhaps be allowed to be original,
but I hope not comic or worse.
Yours sincerely
(signed) A. H. Wolley-Dod.

The main interest of this letter arises from the fact that it was written at a time more or less half
way between the publications of Wolley-Dod’s major papers: The British roses (1908 and 1910) and
his Revision of the British roses (1930-31). It is plain from this letter that he was dissatisfied with the
state of affairs at that time, if not actually bewildered by it. He came to distrust Sudre’s
determinations, and based his nomenclature on that of R. Keller, though realising that British
material could not necessarily conform with the large numbers of named varieties and forms derived
from Continental specimens. E. B. Bishop, on the other hand, working concurrently with Wolley-
Dod, tried to follow Keller exactly, with the result that he grossly over-collected, as his large
herbarium collection in BM shows: there are literally hundreds of specimens of R. canina from the
small areas around Godalming in Surrey and near Ailsworth in v.c. 32 (Northants). This is one of
the things which Wolley-Dod was warning Riddelsdell against in the letter here reproduced. When
Wolley-Dod published his Revision of the British roses he reduced the species to a workable
number, dividing some of the species into “‘groups” which though not strictly valid taxonomically, at
least provided pigeon holes into which the majority of the roses could be placed. The varieties and
forms which he retained were less satisfactory, as Wolley-Dod himself (1936) came to realise.
Though Wolley-Dod’s system can be criticized in the light of modern taxonomic opinion, he did
produce some order out of chaos, and provided a workable system which served British rhodologists
for 60 years.

It is a pity that Wolley-Dod would not admit the possibility of extensive hybridization in Rosa; he
would only allow a rose to be a hybrid if it showed complete or partial sterility. Had he allowed for
this, with his intimate knowledge of the British roses he could have arrived at a system more
conformable with modern taxonomic opinion. Also it appears from his writings that he was aware of
the researches of Blackburn & Harrison (1921) into the peculiar reproductive behaviour of the
caninoid roses, but did not seem to realise their significance. It was not until 1975 that R. Melville,
by the publication of his account of British Rosa hybrids, provided a basis for research (Melville
1975; Graham & Primavesi 1990). This made possible a complete revision of the British roses, as
described in Stace’s New Flora (1991) and Roses of Great Britain and Ireland (Graham & Primavesi
1993). Further research is still required, especially for the informal groups of R. canina, but at least
it can be said that taxonomy of British Rosa is now placed on a firm and sound basis.

ACKNOWLEDGMENT

The author is grateful to the Trustees of the Natural History Museum, London for permission to
publish this letter.

‘REFERENCES

BLACKBURN, K. B. & Harrison, J. W. H. (1921). The status of the British Rose forms as determined by their
cytological behaviour. Annals of botany 35: 159-188.

GRAHAM, G. G. & PRIMAVESI, A. L. (1990). Notes on some Rosa taxa recorded as occurring in the British Isles.
Watsonia 18: 119-124.

GraHaM, G. G. & PrimAVEsI, A. L. (1993). Roses of Great Britain and Ireland. Botanical Society of the British
Isles, London.

NOTES 283

MELVILLE, R. (1975). Rosa L. in Stace, C. A., ed. Hybridization and the flora of the British Isles, pp. 212-227.
Academic Press, London.

Stace, C. A. (1991). New Flora of the British Isles. Cambridge University Press, Cambridge.

Wo ttey-Dop, A. H. (1908). The subsection eu-caninae of the genus Rosa. Journal of botany, London 46,
Supplement: 1-110.

WotteEy-Dop, A. H. (1910). The British roses (excluding eu-caninae). Journal of botany, London 48,
Supplement: 1-141.

WoLLEY-Dop, A. H. (1930-31). A revision of the British roses. Journal of botany, London 68 & 69,
Supplements.

Wo tteEy-Dop, A. H. (1936). Some rose notes. Report of the Botanical Society and Exchange Club of the British
Isles 11: 68-81.

A. L. PRIMAVESI
Ratcliffe College, Fosse Way, Ratcliffe on the Wreake, Leicester, LE7 4SG

ALIEN PLANTS AT FOYNES PORT, CO. LIMERICK (V.C. H8), 1988-1994

Despite their obvious potential as sites of plant introductions, there have been no published
botanical surveys of ports in Ireland prior to the survey of alien plants made at Dublin Port for 1988—
1994 (Reynolds 1996a). This note and list of established and casual alien plants at Foynes Port, Co.
Limerick (v.c. H8), has been written to complement the list for Dublin Port.

Foynes Port (grid ref. R/25.51), on the Shannon estuary about 30 km west of Limerick city, is the
largest public port in the west of Ireland, and the only west coast port with a substantial trade in
animal feed. It has a land area of 52 ha and the annual total throughput was approximately 1 million
tonnes in 1988, increasing to 1-5 million tonnes in 1994, with imports making up 88% of the total
throughput in 1994 (M. V. O’Brien, pers. comm. 1996). Animal feed, including grain, made up 25%
of imports in 1988, increasing to 43% in 1994. Over the study period, about 60% of the animal feed
was imported from New Orleans, and the next most common sources were Indonesia and Malaysia.
Some feed also came from China and more rarely Argentina. Other imports were coal, fertilizers,
petrol, fuel oils and molasses. Although the port had been botanized regularly since 1977, there was
no conspicuous influx of alien plants before 1988 (Reynolds 1990).

In 1988, there was a luxuriant growth of alien plants at Foynes Port, with many rare and new
species, and the probable source was as seeds with the imported animal feed. Animal feed was
normally unloaded by grab into lorries, stored in nearby warehouses and then transported from the
port by road, so there was a certain amount of spillage at each stage. Open land at Foynes Port is
largely unpaved, unlike Dublin Port, where there is little unpaved ground near the docks. At the
eastern edge of Foynes Port, there is an area of newly reclaimed land, which has also been used as an
unofficial dump since 1994.

Established and casual alien plants have been a botanical feature of Foynes Port since 1988, anda
list of all those species found in the port area from 1988 to 1994 inclusive is given in Table 1. Detailed
records for some of these have already been published (Reynolds 1990, 1992, 1993, 1994, 1996b).The
alien species which occurred in the greatest numbers over the seven years were Amaranthus retro-
flexus and Thlaspi arvense while Erucastrum gallicum, Setaria viridis and Erysimum cheiranthoides
were much less frequent. Many other species were found only in small numbers, such as Amaranthus
albus, A. hybridus, Chenopodium spp., Echinochloa crusgalli, Lepidium spp., Matricaria recutita,
Sinapis alba and Sisymbrium loeselii. On the newly reclaimed area in 1994, Amaranthus retroflexus,
Chenopodium ficifolium and Erucastrum gallicum thrived, intermingling with potatoes (Solanum
tuberosum), tomatoes (Lycopersicon esculentum) and other plants of local domestic or garden
origin. Hordeum distichon has been exported from Foynes, hence plants occur casually.

Compared to the 66 species found at Dublin Port over the same period (including Rapistrum
rugosum 1988-1994, which was inadvertently omitted from the Dublin Port list), 41 species were
found at Foynes Port, with an overlap of 24 species (Table 1). Some plants which were very common
~ at Dublin Port were not found at Foynes, for example, Conyza canadensis, Melilotus officinalis,
Rapistrum rugosum, Senecio squalidus, Sisymbrium orientale and Hordeum murinum. Hirschfeldia
incana has been abundant in parts of Dublin Port since at least the early 1980s (Rich 1988), but it was
only seen for the first time at Foynes in 1992, where it is now well established particularly on stony
ground.

284 NOTES

TABLE 1. ESTABLISHED AND CASUAL ALIEN PLANTS FOUND AT FOYNES PORT, CO.
LIMERICK, FROM 1988 TO 1994, WITH YEAR(S) WHEN FOUND.

Years found

Species at Foynes Port Origin Recorded at Dublin Port
Amaranthus albus 1989, 1990 Gr
Amaranthus hybridus 1990 Gr
Amaranthus retroflexus 1988-1994 Gr D
*Avena fatua 1988-1990, 1992, 1994 D
Avena sativa 1989, 1994 Cult D
Bassia scoparia 1988, 1989 Gr D
Brassica juncea 1993 Gr
* Brassica napus 1988, 1989 Cult
* Buddleja davidii 1995 Cult D
*Camelina sativa s.s. 1988, 1989
*Chaenorhinum minus 1993 D
Chenopodium capitatum 1988
Chenopodium ficifolium 1994
Chenopodium glaucum 1988-1990 D
Chenopodium leptophyllum 1990 Gr
Chenopodium strictum 1988 Gr
Crepis tectorum 1988-1991 Gr
* Echinochloa crusgalli 1990 D
*Epilobium ciliatum 19892.1993 D
Erucastrum gallicum 1988-1994 Gr D
*Erysimum cheiranthoides 1988-1993 D
* Hirschfeldia incana 1992-1994 Gr D
Hordeum distichon 1988, 1989, 1992, 1994 Cult D
Hordeum vulgare 1989 Cult
Lepidium ruderale 1990
Lepidium virginicum 1989 Gr
Lycopersicon esculentum 1993, 1994 Cult D
* Matricaria discoidea 1988-1994 Gr D
* Matricaria recutita 1988, 1989 D
*Phalaris canariensis 1993 Gr D
Pisum sativum 1992 Cult D
* Senecio viscosus 1993 D
Setaria viridis 1988-1993 Gr D
*Sinapis alba 1990 Cult
Sisymbrium loeselii 1988 Gr
Solanum tuberosum 1994 Cult D
*Tanacetum parthenium 1993 Cult
*Thlaspi arvense 1988-1994 D
*Trifolium hybridum 1988 Gr, Cult D
Triticum aestivum, unawned 1989, 1994 Cult D
Tropaeolum majus 1994 Cult

Nomenclature follows Stace (1991).

* = listed in Scannell & Synnott (1987) as an established alien in Ireland.

Gr = known grain alien (Clement & Foster 1994; Ryves, Clement & Foster 1996).
Cult = of garden or agricultural origin.

D = recorded at Dublin Port 1988-1994 (Reynolds 1996a).

Several species which are not considered to be ‘certainly introduced’, i.e. alien, in Ireland
(Scannell & Synnott 1987), but which are almost cosmopolitan (Clapham, Tutin & Moore 1987),
were likely to have been introduced with animal feed at Foynes Port, for example, Descurainia
sophia (1988-1992) and Spergula arvensis (1988-1991). In addition, it should be noted that some
species which are considered native or probably native in Britain and so excluded from Clement &
Foster’s (1994) list of aliens, may have arrived at Foynes Port as grain aliens, the grain being used for

NOTES 285

animal feed. For example, Thlaspi arvense was abundant at the port and was found with other aliens
on roadsides leading away from Foynes; it is otherwise rare in Ireland.

In 1995, despite some use of weedkiller, alien plants were again conspicuous. A new addition was
Malva pusilla, a distinctive mallow with tiny pale flowers, which is unlikely to have been missed in
previous years. As has been described elsewhere (Reynolds 1992), many alien plants found at
Foynes Port have also been found on roadsides in Co. Limerick. Although some, such as
Amaranthus retroflexus, Setaria viridis and Thlaspi arvense, are capable of setting seed successfully
under Irish conditions, presumably there are also fresh introductions of aliens each year with spilt
animal feed, mainly grain and feed nuts. Such plants are particularly noticeable where grass verges
have been scraped back mechanically to expose new soil. So far there is no evidence that any of the
plants recorded at Foynes Port are competing with the native flora or are becoming invasive in
natural habitats.

ACKNOWLEDGMENTS

I would like to thank Murrogh O’Brien for the information about Foynes Port, Julian Reynolds for
helpful comments on the manuscript, and both of them for their continuing interest in my work.

REFERENCES

CiaPHAM, A. R., Tutin, T. G. & Moore, D. M. (1987). Flora of the British Isles, 3rd ed. Cambridge University
Press, Cambridge.

CLEMENT, E. J. & Foster, M. C. (1994). Alien plants of the British Isles. Botanical Society of the British Isles,
London.

REYNOLDS, S. C. P. (1990). Alien plants at Foynes and Dublin ports in 1988. Jrish naturalists’ journal 23: 262-
268.

REYNOLDS, S. C. P. (1992). Distribution of alien and adventive plants at ports and on roadsides in Ireland in
1989. Irish naturalists’ journal 24: 59-65.

REYNOLDS, S. C. P. (1993). Records of alien and adventive plants in Ireland 1990-1992. Jrish naturalists’ journal
24: 339-342.

REYNOLDS, S. C. P. (1994). Records of alien and casual plants in Ireland 1993. Jrish naturalists’ journal 24: 515-
D7:

REYNOLDS, S. C. P. (1996a). Alien plants at ports and in coastal habitats on the east coast of Ireland. Watsonia
21: 53-61.

REYNOLDS, S. C. P. (1996b). Records of casual and alien plants in Ireland. /rish naturalists’ journal 25: 186-189.

Ricu, T. C. G. (1988). Hirschfeldia incana (L.) Lagréze-Fossat present in Ireland. Jrish naturalists’ journal 22:
531-532.

Ryves, T. B., CLEMENT, E. J. & Foster, M. C. (1996). Alien grasses of the British Isles. Botanical Society of the
British Isles, London.

SCANNELL, M. J. P. & Synnotrt, D. M. (1987). Census catalogue of the flora of Ireland, 2nd ed. Stationery Office,
Dublin.

STAcE, C. A. (1991). New Flora of the British Isles. Cambridge University Press, Cambridge.

S. C. P. REYNOLDS
115 Weirview Drive, Stillorgan, Co. Dublin, Ireland

THE PUTATIVE HYBRID BETWEEN TWO TEASELS, DIPSACUS FULLONUM L. AND
D. SATIVUS (L.) HONCK. (DIPSACACEAE) IN DUBLIN (v.c. H21)

Dipsacus fullonum L., Wild Teasel, is native in the British Isles; it is common in south-east Britain
and more local elsewhere. D. sativus (L.) Honck., Fuller’s Teasel, is not native, and its origin is
uncertain; it is grown for fulling (raising the nap on woollen cloth) in Somerset, and is also known as
an escape from cultivation and a bird-seed alien (Stace 1991; Clement & Foster 1994). Topham
(1968) and Ryder (1996) have written interesting accounts of Fuller’s Teasel in Britain. In Ireland,
Colgan (1904), quoting an earlier work published in 1772 which described the “Manured Teasel’ as

286 NOTES

being ‘‘cultivated and dried for the use of the clothiers in Dublin”, then commented that the wild
teasel in Co. Dublin was “‘perhaps but a reversion from the plant once so cultivated”’. D. sativus (as
D. fullonum subsp. sativus) has been recorded as a casual with D. pilosus in Co. Down (y.c. H38) in
1915 (Hackney 1992) and at one site in Co. Dublin (v.c. H21) since 1990 (see below). There are no
specimens of D. sativus in the herbaria at the National Botanic Gardens, Dublin (DBN), Ulster
Museum, Belfast (BEL) or Trinity College, Dublin (TCD).

In the past, the two taxa in question were frequently treated as subspecies of D. fullonum (e.g.
Clapham, Tutin & Warburg 1962), but currently they are recognized as distinct species (Hansen
1976; Kent 1992). Recently, the status of D. sativus as a distinct species has again been questioned
(Ryder 1996). I have not been able to discover any mention of the hybrid between these species in
continental Floras despite most of them recognizing two separate species. However, the hybrid
between D. fullonum and D. laciniatus L. has been found in a garden in England (Campbell 1993),
and it is well known on the continent (C. A. Stace pers. comm., 1996).

In the British Isles, 1am aware of only two possible occurrences of the putative hybrid between D.
fullonum and D. sativus. Firstly, there is a record of D. sativus at Bradger’s Hill in Bedfordshire
“with intermediates with D. fullonum” (Dony 1953). Secondly, in a letter to C. A. Stace dated 27
June 1974, Dr J. T. H. Knight wrote that he had found “‘the putative Dipsacus hybrid near Langport
in Somerset some two miles [3 kms] east of the town in ‘Wagg Drove’ which lies off the Langport-
Wincanton road. It was mid-August [probably 1973]. . .”. Dr Knight was attracted by the tallness
of the plants and “‘certain features of the inflorescence’’. He also added that D. sativus was grown as
a crop some five miles [8 kms] west of Langport. A slide of the putative hybrid taken by Dr Knight,
and seen by me, resembles the plant found in Dublin (described below).

In Co. Dublin, D. fullonum is more common along the coast than inland. It has been abundant on
an area of reclaimed land at the edge of Dublin Bay, known informally as ‘Ringsend Dump’ (Grid
ref. O/19.33) since at least the early 1980s. In July 1990, a few plants of D. sativus, also a biennial,
were found in one patch on a cleared gravel site at Ringsend Dump, with dense stands of D.
fullonum within 100 m. A small number of D. sativus plants have been found in the same place every
year since then. In 1994 a possible intermediate between the two species was seen, but not further

10 cm

FicureE 1. Mature, dried flower heads showing the involucral and receptacular bracts of: A. Dipsacus fullonum,
B. the intermediate, and C. D. sativus, from Ringsend Dump, Dublin.

NOTES 287

10 mm
—

5mm

d B .

FiGure 2. Drawings of the receptacular bracts (above) and seeds of: A. Dipsacus fullonum, B. the intermediate,
and C. D. sativus, from Ringsend Dump, Dublin.

TABLE 1. LENGTHS AND WIDTHS OF MATURE SEEDS OF DIPSACUS FROM RINGSEND DUMP,
DUBLIN (MEAN AND RANGE (IN BRACKETS) FOR 20 SEEDS OF EACH TAXON)

D. fullonum Intermediate D. sativus
Mean length (mm) 3-7 (3-1-4-1) 4-6 (4-0-5S-0) 4-3 (3-8-4-7)
Mean width (mm) 1-1 (0-9-1-8) 1-8 (1-3-2-2) 2-0 (1-7-2-3)

checked. The following year, when the site was visited with the Dublin Naturalists’ Field Club on 9
August 1995, there were two plants of D. sativus and, beside them, one large plant with
conspicuously intermediate characters between it and D. fullonum, particularly noticeable in the
involucral and receptacular bracts, and stem leaves. This plant was 2 m tall with about 30
inflorescences. Presumably the more distant D. fullonum provided the pollen for this cross.

288 NOTES

Further visits were made to the site later in August and in mid-September, when specimens were
taken and seeds collected. Fig. 1 shows whole flower heads, and Fig. 2 shows details for the
receptacular bracts and seeds. Descriptions of some characters of Ringsend Dump plants in fruit are
given below for both parents and for the intermediate. Measurements were made on 20 mature
seeds (Table 1). Voucher specimens have been deposited in DBN.

D. fullonum

Stems with many prickles; stem leaves with sharp prickles on underside of midrib; the longer
involucral bracts spiny, curving upwards, usually as long as or exceeding the inflorescence;
receptacular bracts with long straight flexible spines, conspicuously ciliate; seeds longer than wide,
c. 3-7 X 1-1 mm, four-sided with one longitudinal ridge in the middle of each side, brown with
appressed hairs.

D. sativus

Stems with fewer, shorter, blunter prickles than D. fullonum; stem leaves with no prickles on
underside of midrib; the longer involucral bracts not spiny, spreading, much shorter than the
inflorescence; receptacular bracts stiff, short with strongly recurved tips, inconspicuously shortly
ciliate; seeds longer than wide, c. 4-3 x 2:0 mm, four-sided with one, two or occasionally three
longitudinal ridges in the middle of each side, pale brown with silvery white appressed hairs, more
densely hairy than D. fullonum or the intermediate.

Intermediate

Stems with fewer prickles than D. fullonum; stem leaves with prickles on underside of midrib; the
longer involucral bracts curving upwards, usually shorter than the inflorescence, somewhat spiny;
receptacular bracts stiff, slightly recurved at the tip, longer than in D. sativus, ciliate; seeds longer
than wide, c. 4-6 X 1-8 mm, four-sided with usually one longitudinal ridge in the middle of each side,
occasionally two, dark brown with appressed hairs.

It was noted that the length of the longest involucral bracts and degree of prickliness of the stems
were variable on individual plants at Ringsend Dump. However, the structure of the receptacular
bracts was much more constant for each taxon. Mature, dried seeds of D. fullonum were distinctly
smaller and narrower than those of D. sativus, while the seeds of the intermediate were slightly
larger and darker than those of the latter species. By mid-September, the seeds of all three taxa were
mature and being released from the inflorescences. Seed-set was as good in the intermediate as it
was in both parents. Seeds of the intermediate, planted outdoors in the author’s garden in
November 1995 had germinated by early April 1996, while others planted indoors on 5 April 1996
had germinated eleven days later.

As Dipsacus seeds are heavy and have no adaptation for wind dispersal, many will drop beneath
the parent plants; hence D. fullonum may form dense stands. However, the number of D. sativus
plants at Ringsend Dump has not increased since 1990. At that time, the site was more open than in
August 1995 when it and the intermediate were growing among a dense cover of Agrostis
stolonifera, Dactylis glomerata, Cirsium arvense, Plantago lanceolata, Hypochaeris radicata, etc. In
September 1995, the area had been burnt leaving small patches of open ground. This may allow new
plants to establish themselves more easily.

There are many plants of garden origin at Ringsend Dump, and it is possible that the D. sativus
arrived here as seeds in garden refuse. The 1993 catalogue for Chiltern Seeds listed D. fullonum,
“used by fullers”, as ‘“‘one of the most popular flowers for drying for use as flower arranging
material’, also attractive to bees and butterflies, and easily grown. It is also listed in the 1996
Thompson & Morgan catalogue, with D. sativus given as-a synonym. The picture on the seed
packet, only labelled “Teasel’, is that of D: sativus; however the seeds are not identical with those of
that species collected at Ringsend Dump. The description on the packet says that the cylindrical
heads are evenly covered by hooked spines used for raising the nap on cloth, but then goes on to
describe this teasel as a ‘‘native’’ plant. The picture on the Suttons Seeds “Teasel (Dipsacus
fullonum) packet appears to be of Wild Teasel, and the flower heads are described as large and
spiny, ‘‘a valuable source of nectar for bees and butterflies” or to be cut and dried for winter
decoration. The enclosed seeds are similar in appearance to those of Thompson & Morgan.

NOTES 289

In conclusion, since fertile hybrids may be formed between two recognized species (Stace 1975),
and since two distinct species, D. fullonum and D. sativus, are now recognized, then the plant found
in Dublin with intermediate characters and mature, viable seeds should be considered the putative
hybrid.

ACKNOWLEDGMENTS

I would like to thank Clive Stace for his very helpful and prompt responses to my enquiries about
Dipsacus hybrids and for allowing me to quote from Dr Knight’s letter; John Parnell for bringing
Ryder’s paper to my attention; David Nash for the photograph of the flower heads from which Fig. 1
was prepared; Conor Reynolds for help with Fig. 1; Julian Reynolds for drawing Fig. 2; and David
and Julian for constructive comments on the manuscript.

REFERENCES

CAMPBELL, W. D. (1993). Dipsacus laciniatus. B.S.B.I. news 63: 30-31.

CriaPHAM, A. R., Tutin, T. G. & WarBuRG, E. F. (1962). Flora of the British Isles. Cambridge University Press,
Cambridge.

CLEMENT, E. J. & Foster, M. C. (1994). Alien plants of the British Isles. Botanical Society of the British Isles,
London.

CoLeGan, N. (1904). Flora of the County Dublin. Hodges, Figgis & Co., Dublin.

Dony, J. G. (1953). Flora of Bedfordshire. The Corporation of Luton Museum and Art Gallery, Luton.

HAcKneY, P. (1992). Stewart & Corry’s Flora of the north-east of Ireland, 3rd ed. Institute of Irish Studies, The
Queen’s University of Belfast, Belfast.

Hansen, A. (1976). Dipsacus L. in Tutin, T. G. et al., eds. Flora Europaea 4: 58-59. Cambridge University
Press, Cambridge.

KENT, D. H. (1992). List of vascular plants of the British Isles. Botanical Society of the British Isles, London.

Ryper, M. (1996). Is the Fullers’ Teasel (Dipsacus sativus) really a distinct species? The Linnean 11: 21-27.

Stace, C. A., ed. (1975). Hybridization and the flora of the British Isles. Academic Press, London.

STACE, C. A. (1991). New flora of the British Isles. Cambridge University Press, Cambridge.

TopuaM, P. N. (1968). The Fuller’s Teasel. Proceedings of the Botanical Society of the British Isles 7: 377-381.

S. C. P. REYNOLDS
115 Weirview Drive, Stillorgan, Co. Dublin, Ireland

EARLY GENTIAN (GENTIANELLA ANGLICA (PUGSLEY) E. F. WARB.)
PRESENT IN WALES

Gentianella anglica (Pugsley) E. F. Warb. is a rare endemic protected under Schedule 8 of the
Wildlife and countryside Act, 1981 (as amended). It is also listed in Appendix 1 of the Council of
Europe’s Bern Convention and Annex II of the E.U. Directive on the conservation of habitats and
wild fauna and flora (‘Habitats Directive’).

This short note documents the two records of Early Gentian, Gentianella anglica (Pugsley) E. F.
Warb. for Wales, both from Pembrokeshire (v.c. 45) and determined by TCGR. The plant has
previously been regarded as an English endemic and there are no published records for Wales.

1. Near Tenby, 17 May 1921, J. E. Arnett (OXF). The collection consists of one specimen with a
corolla 18 mm, three internodes and a terminal internode forming 35% of the length of the stem. It
was mounted on a sheet with other material named by G. C. Druce as Gentiana amarella var.
praecox, but the record does not appear to have been published. The herbarium sheet was seen by
N. M. Pritchard during his work at Oxford on Gentianella but he makes no reference to the
specimen (Pritchard 1959).

Arnett’s copy of ‘On the botany of South Pembrokeshire’ (Babington 1863) contains a pencil
annotation to “‘Gentiana amarella var. praecox”’, which is what G. anglica was known as at the time,

290 NOTES

but there are no additional specimens in Arnett’s herbarium at Tenby Museum (TBY; S. V. Baldwin
pers. comm., 1996).

I am aware that Gentianella uliginosa (Willd.) Borner also occurs in the Tenby area. All
specimens seen have been collected flowering from July onwards and differ in morphology.

2. Stackpole National Nature Reserve, sparsely in dry hollow dominated by mosses and lichens, SR/
984.944, 10 June 1994, A. Jones (NMW). The collection of three plants has a mean corolla length of
15 mm, a mean of 2-3 internodes and the terminal internode forms a mean of 63% of the length of
the stem.

Odd ‘G. amarella (L.) Borner’ plants were first shown to me on Stackpole Warren in 1990 by Bob
Haycock. Plants collected for identification on 16 July 1990 mostly had four internodes and a slightly
contracted terminal internode (measurements by S. B. Evans; specimens not seen). These may be
hybrids between G. anglica and G. amarella.

The rediscovery and conservation of the Tenby locality is much to be desired. The plant should be
searched for in short, dry, open, calcareous grassland in May or early June. Populations of G.
anglica elsewhere are known to fluctuate markedly in abundance from year to year due to its
biennial habit, so suitable sites may need to be investigated repeatedly.

ACKNOWLEDGMENTS

I would like to thank S. V. Baldwin for information about the collections at Tenby Museum and for
tracing Arnett’s annotations, Serena Marner for help and access to OXF, and Stephen Evans, Lynne
Farrell, Bob Haycock and Andy Jones for help and information.

REFERENCES

BaBINGTON, C. C. (1863). On the botany of South Pembrokeshire. Journal of botany 1: 258-270.
PRITCHARD, N. M. (1959). Gentianella in Britain. I. G. amarella, G. anglica and G. uliginosa. Watsonia 4: 169-
193. ;

T..C. G. Ricu
Botany Department, National Museum of Wales, Cardiff, CF1 3NP

NOTES ON SOME EARLY SUSSEX BOTANICAL RECORDS BY THOMAS SOCKETT
(1777-1859)

Born into quite a humble family on 20 November 1777, Thomas Sockett was plucked from obscurity
when he was about 14 years old by the poet William Hayley (Povey 1928).The young Sockett was
helping to operate a device for generating static electricity for therapeutic purposes in Weston
Underwood, Buckinghamshire. Hayley, who was interested in scientific gadgets in general and this
machine in particular, encountered the boy while on a visit to fellow poet Cowper; so impressed was
he with his natural intelligence that Hayley took Sockett back to his home in Eartham, West Sussex,
to be preceptor and companion to his son. The arrangement evidently proved satisfactory and in
1794 Sockett assisted Hayley with the transcription of the autobiography of Edward Gibbon which
was being prepared for publication by Lord Sheffield, and he was described as having a good
education and being able to read Latin and French.

In 1795 he became preceptor to Lord Egremont’s eldest son, and in 1797 was tutor at Petworth
House. In 1806 he went to Exeter College, Oxford (presumably at Lord Egremont’s expense) and
was ordained in 1808. He graduated in 1810 and became Rector first of Tadcaster and then of North
Scarle, Lincolnshire and of Duncton and Petworth in Sussex. He resided mainly at Petworth, and
died on 17 March 1859. There is a monument to him in Petworth Church and a portrait in Petworth
House (reproduced in Povey 1928).

Sockett was also godfather to F. H. Arnold, who wrote the first Sussex Flora (Arnold 1887).

NOTES 291

Arnold’s father was choir master in Sockett’s church at Petworth. It is almost certain that Arnold
learnt some of his plants from Sockett, and in his History and antiquities of Petworth thanked him for
ten years of tuition (Arnold 1864). Oddly, there are no records attributed to Thomas Sockett in
Arnold’s Flora of Sussex, but he does include some records by one of his sons, Henry Sockett.
Henry Sockett also entered the church and had the living of nearby Sutton for many years.

RECORDS FROM PETWORTH

In the Petworth House archive held at the West Sussex Office, there is an unpublished, incomplete
manuscript from 1805 entitled ““Mr Sockett’s Journal’ (no. 1679). This diary indicates that he
regularly hunted, played Real Tennis and read many classical works in Latin, especially those of
Horace. He also gives a fascinating account of seeing Lord Nelson on the Isle of Wight before he
embarked on the HMS Victory prior to the Battle of Trafalgar. He was also a botanist and the
fragment of his journal contain three specific plant records:

Thursday 19 September 1805 ‘‘went into the paddock [at Petworth] to get Sedum telephium to dry
but it was all out of flower’.

Wednesday 25 September 1805 “went into the pleasure ground to look for fungi— found Asplenium
dilatatum [probably Dryopteris dilatata| which I have brought home and dried”. The pleasure
ground was a wild garden, walled on one side, to the north of Petworth House.

Saturday 5 October 1805 “‘went in a boat to Wiggonholt and brought back four plants of Butomus
[wmbellatus| which I planted in the pond”. Wolley-Dod (1937) records Butomus from Petworth as
“Gntroduced?)”, and attributes the record to F. H. Arnold. This record is not listed in Arnold’s
(1887) Flora of Sussex.

The records for Sedum and Butomus predate the first localised records for Sussex given by
Wolley-Dod (1937). Sockett is not mentioned in British and Irish herbaria (Kent & Allen 1984) and
no herbarium specimens are known to survive. Arnold’s herbarium at Christ’s Hospital, Horsham
(HSM) does not contain any of Sockett’s specimens.

ORCHIS SIMIA In SUSSEX

Wolley-Dod (1937) gives a record for Orchis simia ‘‘Petworth, Sussex, 1801, Mr Sokot, in F. Bauer’s
drawings of British orchids, pp. 69-70. This is the only record known and is no doubt a good one,
though the species has never been recorded since in Sussex’’. This ‘Mr Sokot’ must refer to Thomas
Sockett, the spelling being a corruption by Bauer.

Franz Bauer (1758-1840) was Kew’s first botanical artist and was outstanding (Stewart & Stern
1993). His original drawings are held in the Botany Library at the Natural History Museum,
London. The Orchis simia illustration is annotated at the bottom in black ink “‘Mr Sokot from
Petworth, Sussex, June 4 1801’’. Other drawings are annotated in pencil, and the annotation
appears to have been added at a later date.

There is no doubt about the identity of the drawing of the Orchis simia, but the origin of the plant
is another matter. The most obvious reading of the wording in light of what is now known about
Sockett is that it is he, and not the Orchis simia, that was from Petworth. Petworth is in the middle of
the Weald, and the nearest suitable calcareous soils occur on the South Downs. Arnold (1864, 1887)
makes no reference to the plant, which would surely have been of sufficient note for Sockett and he
to have discussed in relation to Petworth. Unfortunately it has not been possible to trace where
Sockett or the Egremont family were at the time the orchid was collected. The record of Orchis
simia for Sussex is therefore rejected.

ACKOWLEDGMENTS

We would like to thank Lord Egremont for access to the Petworth House archive, Judith Magee for
help with Bauer’s drawings at the Natural History Museum, London, and Mary Briggs.

292 NOTES
REFERENCES

ARNOLD, F. H. (1864). History and antiquities of Petworth. A. J. Bryant, Petworth.

ARNOLD, F. H. (1887). Flora of Sussex. Simpkin, Marshall, Hamilton, Kent & Co., London.

Kent, D. H. & ALLEN, D. E. (1984). British and Irish herbaria. B.S.B.1., London.

Povey, K. (1928). The rise of Thomas Sockett. The Sussex county magazine 2: 38-40.

STEWART, J. & STEARN, W. T. (1993). The orchid paintings of Franz Bauer. The Herbert Press and the Natural
History Museum, London.

WoLLeEy-Dop, A. H. (1937). Flora of Sussex. Chatford House Press, Bristol.

TCG RICH

Botany Department, National Museum of Wales, Cardiff, CF1 3NP
B. C. RICH

Chase House, June Lane, Midhurst, West Sussex, GU29 9EL

INE J STURT

Peacehaven, The Crescent, West Wittering, West Sussex, PO20 8EE

Watsonia 21: 299-302 (1997) 293

Book Reviews

Flowers of Britain and Ireland. J. R. Akeroyd. Pp. 255. Harper Collins Publishers, London. 1996.
Price £6.99. ISBN 000—220004-X.

This softback book is a member of the latest series from Collins, the ‘‘Collins Wild Guide’’. It is a
photoguide that aims to introduce the interested reader to botany. With such a large taxonomic
group as flowering plants the difficult task is to cover a sufficient number of commoner species, and
yet produce a guide which is adequate in the field. This book achieves this in two ways. The rare and
showier species (for example, the rarer members of the orchid family) are excluded from the book
(unlike many other photoguides and wild flower books) and the common and less photogenic plants
(for example, the plantains and thistles) are included.

The combination of good photographs and an “ID fact file’’ enables the reader to identify many
common species encountered in a variety of habitats. In the cases of similar species, such as the St
John’s-worts (Hypericum spp.), identification of some species simply is not possible with this guide.

The inclusion of distribution maps is a good idea, but in practice they are of limited use since they
are quite small and since the species in the book are common and therefore occur generally
throughout Britain. The provision of information about the medicinal uses of plants is very
interesting and gives the book a wider range of appeal.

For the beginner, the good photographs and the choice of plants makes this book an ideal
purchase at £6.99.

M. THOMASON

A natural history of Guernsey, Alderney, Sark & Herm. B. & J. Bonnard. Pp. xii + 160. Guernsey
Press Co., Guernsey. 1995. Price £9.95. ISBN 0902550-60-8.

This is an excellent book covering the very different islands as well as space allows, embellished with
some of Brian Bonnard’s photographs. It starts with the formation of the islands, their pre-history
and settlement, and then goes through the anomalously distributed forms of wildlife to animals,
birds, fish, marine life, ecology, weather, conservation and all (including the Alderney Cow). Most
of the special plants of all sorts are duly brought in. Brian Bonnard has been studying them closely in
Alderney for several years and is their recorder. His knowledge of them is well brought out in his
Out and about in Alderney which came out at much the same time from the same publishers for
£4.95. Both are well worth getting, and using.

D. McCLINTOCK

The Flora of Hampshire. A. Brewis, P. Bowman & F. Rose. Pp. 408, 16 colour plates, 25 black and
white plates, 13 figures and 579 maps. Harley Books, Colchester. 1996. Hardback £45.00, ISBN 0-
946589-34-8. Paperback £25.00, ISBN 0—946589-53—4.

Hampshire has long been the only county on the south coast without a modern Flora, and in
southern England only Oxford and Buckinghamshire now lack any post-war record. The Flora of
Hampshire has been so many years in gestation that I must confess I was deeply sceptical as to the
outcome, but, with only a few reservations, we now have a relatively up-to-date, superbly produced
and well-balanced Flora.

The introductory chapters are excellent. Those on geology, climate and history are just the right
length. The chapter on conservation is welcome and informative, and that on previous botanists is
exemplary. But it is the writing on habitats that is almost worth the price of the book — a masterly

294 BOOK REVIEWS

overview of the vegetation, with sensible grid references to allow one to find the places described. I
am not sure how Hampshire came to annex the Allen Valley (p. 61) from deepest Dorset, but that is
a small matter.

The chapter on Hampshire’s flora vis-a-vis other counties is fun, is reminiscent of Victorian Floras
and is mischievous! On balance I would agree that Hampshire’s vascular flora is slightly richer than
that of Dorset, but it is a desperately close-run thing. There are 28 species in widest political Dorset
not in widest political Hampshire against 30 in Hants not in Dorset. And of course Dorset is much
richer in lichens and Hampshire consists of two vice-counties. Readers will appreciate that this can
run and run, but I think it is a useful chapter. Already v.c. 69 (Westmorland) claims more
Cyperaceae than Dorset or Hampshire!

It is surely a mistake that one has to wait until page 80 before one discovers the area covered by
the book. This turns out to be all of v.cc. 11 and 12, plus all the bits of surrounding vice-counties now
part of political Hampshire, and seems the correct approach (and of course helps the richness!). The
text and map of this chapter are first-rate and offer a convincing defence of the permanence of the
vice-county system. In fact all the maps are excellent and clear, although I would have liked a bigger
and more detailed general map of the county — Fig. 1 is just not good enough.

On to the heart of the book, the systematic accounts of the Flora. I found this admirable, with just
the right balance of caution, where appropriate, over the status of species such as Pulsatilla vulgaris,
Erica ciliaris, Utricularia vulgaris, Cochlearia officinalis and Poa infirma. Briza minor is properly
treated as a cornfield weed rather than as an introduction as in Stace’s New Flora of the British Isles.
The tetrad maps are well-chosen rather than omnipresent, but they are post-1930, which is a very
long time ago, and must be my major caveat over this book. There is just no way of telling whether
the plants are still present, nor is there any indication of how thorough was the coverage of each
tetrad. I can find only a few errors in this section, in that Eleocharis parvula is not in Dorset other
than in v.c. 11, that Rhynchospora fusca is probably three times as common in Dorset rather than
the other way about, and I thought everyone knew that the Isle of Wight recent record for
Centaurium tenuiflorum was an error. Is Persicaria mitis an accepted new name? The treatment of
aliens is excellent, with not too many ephemerals covered, and with a very useful date of first record,
something that later workers will appreciate.

The lichen and bryophyte chapters are what they should be — concise coverage by 10-km squares,
leaving the field free for a proper Flora as for Sussex.

Finally, the indexes are exemplary, although I might have put last that for vacua plants, on the
grounds that it will be used most, or even combined all three. They have a clear integration of Latin
and English names and are a pleasure to consult. I would have liked a gazetteer, but four-figure grid
references are consistently given in the text.

To sum up then, we have this beautifully produced new Flora, of a sensible size rather than the
clinical A4, with perfectly acceptable colour plates of relevant plants — special plants, beautifully
photographed. I do not think I have ever spent more than a few seconds looking in any Flora at these
now obligatory colour plates, but these are almost an exception. The long gestation which meant the
relatively unsatisfactory position of the tetrad maps is probably outweighed by the excellence of
both the introductory chapters and the systematic accounts. The price is a new high in county Flora
terms, but, for one’s money, one has what is probably the best produced county Flora.

D. A. PEARMAN

Fléra Chorca Dhuibhne. Aspects of the flora of Corca Dhuibhne. M. ui Chonchubhair, photographs
by A. O Conchtir. Pp. xviii + 270 with 370 colour photographs and 75 line drawings. Oidhreacht
Chorca Dhuibhne, Ballyferriter. 1995. Price IR£20.00. ISBN 0—906096-07-3.

Oidhreacht Chorca Dhuibhne (The Heritage of Corkaguiny) was founded in 1980 with the aim to
make available to people of the locality all aspects of the information relating to the physical
environment. An archaeological survey was published in 1983 and, following this, a study of the
natural history was begun. The wealth of information generated forms the basis of the present Flora.

BOOK REVIEWS 295

The field work was conducted throughout in Irish but for publication the section on the flora is
bilingual. The Irish names are those in use in the Dingle peninsula. A glossary is provided.

The publication treats of the herbaceous flowering plants and ferns of Corca Dhuibhne, one of
five baronies which make up South Kerry, and one third of the vice-county, H1. This east-west
peninsula — the Dingle peninsula — is some 50 km long. The peninsula projects well into the Atlantic
Ocean. Its extremity, Dunmore Head, is the most westerly land in Europe (Iceland excepted), and
the Blasket Islands form an offshore extension of the ridge.

In the foreword the background to the study is set out. It was initiated to encourage local interest
in natural history. Introductory chapters narrate in brief the history of the area. The soils are in the
main acidic and blanket bog is widespread. We are informed that the ‘Neolithic people. . . begana
process of altering the landscape that has continued to the present... the. . . total lack of natural
tree cover in the area is one result of this’. Elements in some place names provide evidence of
woodland in times past.

The main chapter deals with some 300 common species of plants including naturalised aliens.
Each page begins with a colour photograph accompanied by three columns of text, one in English.
A page is devoted to a species, but in the case of grasses and sedges, there are four species per page.
The short descriptions are followed by folk, medicinal and herbal notes. We learn that Heracleum
sphondylium is known about Tralee as “‘the singer . . . from the whistling of the wind through the
hollow stems in winter’’. Fuchsia magellanica is ‘‘widely planted as a durable hedge” — in places it
seems to be the only hedging plant. Under Euphorbia helioscopia it is noted ‘“‘Spurge has been used
by poachers to. . . kill fish’. E. hyberna (Irish spurge) is the plant used in the past in the south-west
for this purpose, but it is not mentioned at all though it occurs in the extreme west. A hawthorn
standing alone is believed “‘to belong to the Otherworld’’, to interfere with it would “‘invite
disaster’. There are few printing errors. The publisher has drawn attention to an error on p. 193,
where Cymbalaria muralis carries an incorrect illustration.

Colour photography by Aodan O Conchuir is a feature of the Fléra. Most of the photography is of
high quality, as in Pinguicula grandiflora (showing the overlapping lobes of the corolla lip), and a
close-up of Menyanthes trifoliata (flower). Seeking to show the image and the habitat in the one
frame does not always work as the pictures of Lemna minor, Juncus effusus, Hyacinthoides non-
scripta, Teucrium scordium demonstrate.

Fléra Chorca Dhuibhne is not a Flora in the conventional sense. The lay-out is not in the standard
format, species are not localised, records are not dated. The author choose “‘Flora”’ deliberately as
“it provides a balanced flavour of the accessible flora of the region’’, but “‘is not exhaustive”. Flora
Chorca Dhuibhne meets admirably the aims of the Oidhreacht. Great thought went into the
production and all concerned are deserving of praise. The book is well printed and is firmly bound,
but it is not a Flora for the pocket. Dr Caroline Mhic Daeid carried out a skilful job as botanical
editor.

To the reviewer the publication is a manual of botany, providing plant instruction, ecology anda
guide to the flora in one cover. The book is recommended to botanists and environmentalists.

M. J. P. SCANNELL

James Bolton of Halifax. J. R. Edmondson. Pp. 72. National Museums & Galleries on Merseyside,
Liverpool. 1995. Price £12.95. ISBN 0—906367—80-8.

Between, on the cover, full-colour reproductions of his immaculate paintings of Vaccinium
oxycoccos and Amanita muscaria, James Bolton is brought to life in a copiously illustrated account
of his work as an 18th century naturalist and artist. He is shown to have acquired an intimate
knowledge of his local flora of West Yorkshire, and to have added considerable impetus to the
development of taxonomic thought in Britain and abroad. Through a wide circle of friends and
correspondents, he was able to extend his interests to many parts of Britain. He illustrated, for
instance, Relhan’s (1785) Flora Cantabrigiensis, and explored as far afield as Snowdonia. Two
centuries later, the accuracy of Bolton’s work, as demonstrated by this book, still provides reliable
base-line data of the kind which gives meaning to the aims of A‘las 2000.

Publication of this book was timed to coincide with an exhibition at Liverpool Museum, but it is
by no means a mere catalogue. It stands alone as a gripping tale of Bolton’s achievements and place

296 BOOK REVIEWS

in the advancement of biological science, for Dr Edmondson has unearthed long-lost biographical
detail and has had access to much of Bolton’s original art-work, based on vascular plants, fungi,
bryophytes, lichens and birds. Some of it is now published for the first time.

The book is well-presented and provides compelling reading. It also presents historians with
tabulated information on James Bolton’s contemporaries and associates, as well as the places he
visited. Typographical errors are generally minor, although Bolton’s patroness is said to have died
in 1785 and 1786. I was particularly disappointed to be promised, on p. 36, a list of the moss-
drawings in the unpublished manuscript, Genera Muscorum, only to be denied it. However, I am
glad to dispel the author’s doubt about the locality known to Bolton as Clattering Syke, a
wonderfully evocative name that is still with us at SD/86.64.

Dr Edmondson is to be congratulated on producing a valuable and attractive addition to the
literature. I enjoyed reading it, and have no doubt that others will, too.

M. E. NEwTron

Historical ecology of the British flora. M. Ingrouille. Pp. xi + 347. Chapman & Hall, London. 1995.
Price £22.56. ISBN 0—412-56150-6.

In this book Ingrouille has attempted to provide an account of the historical ecology of the British
flora that deals not just with the historical period normally considered by historical ecologists, nor
even only with the Quaternary, but that begins with the earliest land plants of the genus Cooksonia
found in Silurian rocks dating from over 420 million years ago in Tipperary. The book has only three
chapters, each with numerous sections, and, as might be expected, the temporal resolution
increases towards the present day. Chapter 1 (93 pp.) takes us from the Silurian up to the end of the
last (Devensian) glaciation. Chapter 2 (109 pp.) then deals with the period from the late Devensian
up to the mid-Holocene “‘elm decline’’, whereas the period since then, that is marked by intensifying
human impact, is dealt with in Chapter 3 (107 pp.). The book is well illustrated, with numerous
monochrome photographs and four sides of colour plates, and is rounded off by a useful
bibliography and an adequately comprehensive index (15 pp.).

Unfortunately I found the prose did not stimulate the natural excitement that I have for
palaeobotany; indeed much of the account of the earlier geological record was ‘catalogue-like’ and
made great demands upon one’s ability to maintain concentration. When I reached those sections
dealing with the Quaternary and recent periods, about which I have greater knowledge and
expertise, then I found the treatment wanting also in that it was somewhat dated in approach. That
is not to say that recent literature is not cited, rather that it is interpreted and presented in a way that
defines a view more in keeping with the prevalent views amongst Quaternary palaeoecologists
perhaps a decade or more ago. In this respect the book singularly fails to reflect the growing
importance of palaeoecology with respect to the need to understand the response of organisms to a
changing environment. Section 3.6.5 on climate change is less than one page in length, and presents
a minority view that suggests that there is uncertainty over ‘‘how carbon dioxide pollution might
cause global warming’’, failing either to discuss or even hint at the records of natural change in
‘““CO*” (sic) levels in the atmosphere during the late Quaternary and their relationships to past
climate changes, or to refer to the work and conclusions of the Intergovernmental Panel on Climate
Change (IPCC).

Notwithstanding my reservations, the book is a useful compilation to which I shall want to refer
first- and perhaps second-year undergraduates. It is unfortunate, however, that it fails to satisfy the
demands that I would make of a senior undergraduate textbook, in terms of a rigorous and up-to-
date treatment, whilst at the same time being written in prose that I for one was unable to read for
any extended period, ruling it out as a successful ‘popular’ text. The author attempted a difficult task
and in doing so has provided a useful addition to our library shelves. Perhaps it was inevitable, given
the scope of his task, that academic reviewers would be likely to find the product wanting in some
respects; I feel, however, that it could have been done more successfully and with a more up-to-date
treatment, although this probably would have required more than one author in order to provide
adequate depth across the entire time span to be considered.

B. HUNTLEY

Watsonia 21: 303-305 (1997) 297

Obituary

CHRISTINA MAYNE DONY
(1910—1995)

Chris Dony (née Goodman) passed away peacefully on 23 May 1995. Although not in the best of
health during the previous months, her death came as a great shock to her family and friends. I first
met Chris in 1975, soon after I moved to Bedfordshire and became interested in the local flora. From
that time onwards I joined some of the botanical exploits of Chris and her late husband John around
the county. These were wonderful occasions and from both of them I learned much of my field
botany. Chris herself was an excellent field botanist, with a keen eye backed up by a wealth of
knowledge. She was particularly good at the grasses and her enthusiasm for wool aliens, encouraged
by her friend Charles Bannister (1918-1984), was unbounded. She seemed to have a second sight for
locations of unusual species, often leaving the beaten track for no apparent reason, and then one
would soon hear the call to see a new species for the day’s records. One of her most interesting finds,
on a visit to an old railway line in north-east Bedfordshire, was the discovery of an extensive colony
of a Petrorhagia sp. This, after much deliberation, was identified as P. prolifera which has recently
been designated as a British native species. In the intervening years, with no protection, the site has
been nearly destroyed by sand extraction with the plant apparently being eliminated from the site.
However, earlier this year (1995) a few plants were discovered at the edge of the sand quarry but,
sadly, Chris did not live to hear that the plant had been rediscovered. I hope that the colony can be
conserved, perhaps in her memory.

Christina Mayne Goodman was born in Selly Oak in Birmingham, into a family of three older
sisters and later they were joined by a brother. She was educated at Edgbaston High School for Girls
in Birmingham and, after leaving school, went to work in the family business, a builders and coal
merchant, in Selly Oak. She worked there from 1932 for 30 years, eventually becoming a director.
During the last war she did voluntary work in the ARP and at the Birmingham Service Club, and, in
addition, she worked nights once a week in an aircraft parts factory.

The Goodmans were a sporting family and Chris followed this course for many years with hockey
being her main interest. She played centre forward for Edgbaston Ladies Club from 1929-1950, for
the County of Warwickshire from 1939-1949 and for Midland Counties from 1932-1948. In 1935 she
was a reserve for the England team and, finally, played for her country in 1937, 1946 and 1947, in the
latter year being England vice-captain. Recently she reminisced to a reporter on the way
international sport was conducted some 60 years ago “*. . . we just turned up at the ground about an
hour before the game and had dinner together afterwards” (Daily Telegraph, 12 May 1995).
Although Chris had an early interest in natural history, it was to sport that much of her life was
devoted at this time.

After the war, although continuing to play hockey, she became involved in natural history,
joining the Birmingham Natural History Society in 1947. She served on its Council and was, for
seven years, secretary of the Botanical Section. She was involved in recording for the Flora of
Warwickshire up to 1965, and, in the Flora, when published in 1971, it was noted that she was one of
the main contributors of the field records.

One of her early botanical interests was the study of the alien flora, in particular the wool aliens of
Worcestershire and it was through this that she met her future husband, John Dony. They were
married in 1962 and she was soon involved in the botany of her adopted county, Bedfordshire, an
area where there was also a good diversity of wool aliens. However, Chris always maintained an
interest in the botany of her home area. At the time of their marriage John was the Honorary
Secretary of the Botanical Society of the British Isles (BSBI). Chris had joined the Society in 1948,
but now became involved in the administration. She was Membership Secretary for ten years from
1964 and it was during her tenure that the Society’s membership increased to over 2000. She sat on
Council for four years and Meetings Committee for ten years. In 1975 she was elected an Honorary
Member in recognition of her sterling work for the Society. This was the first and, I believe, only

298 OBITUARY

time that both husband and wife have been accorded this honour. In later years, up to just before
John’s death in 1991, they both thoroughly enjoyed the annual Exhibition Meetings at the Natural
History Museum, then known as the BM (British Museum (Natural History)) where they were able
to meet up with many of their botanical friends and colleagues.

Chris joined the Wild Flower Society (WFS) in 1964 and immediately started her Diary. By 1967,
and in record time, she had attained Valhalla, and, by 1972, Parnassus. In a notebook recording
yearly totals since that year she noted that by 1988 the total had reached 2530 but that this was
‘““, . Our last year of active recording”’. By 1994 two more species had been added giving her final
total as 2532.

Following John’s retirement from teaching in 1964 they devoted their time to the study of the flora
of Bedfordshire and Hertfordshire. During the recording season it was not unusual for six days a
week to be spent in the field, and it was only at Chris’ firm insistence that at least half a day per week
was spent on housework and shopping! Certainly it was only by this single-minded approach that
they were able to complete field work for the local Floras of these counties within a relatively short
time period, an amazing six years for the Bedfordshire plant atlas. The work was definitely a team
effort and John often told me that Chris really should have been a co-author of this and the
Hertfordshire Flora. The wild flowers of Luton was published in 1991 and, indeed, Chris was
accorded joint authorship.

Chris was involved in many important publications associated with the flora of Bedfordshire and
the full references are detailed below. She carried out a detailed study of the spread of Puccinellia
distans along the major routes through the county. In 1981 she started studying the dandelions of
Bedfordshire and, in the following five years, with her co-workers, had recorded about 65
Taraxacum species for the county. This list was published jointly with Adrian Rundle and many
additions were added in subsequent years. This study was one of the most important of recent years
on the Bedfordshire flora, and it was due to her marvellous eye for detail that she became quite an
expert in this difficult group. Following the publication of the Bedfordshire plant atlas in 1976 the
study of the local flora did not stop and subsequent discoveries were duly published in a joint paper
with John in Watsonia. Many important sites were resurveyed resulting in a comprehensive report
detailing the important species present in virtually all the worthwhile sites of the county. This
document is an invaluable local resource and, whilst it was written in John’s inimitable handwriting,
the content owes a great debt to Chris.

For many years the Donys, either separately or jointly, presented exhibits at the Society’s Annual
Exhibition Meetings on the local flora. In addition Chris presented her detailed study of the
Society’s membership from 1942-1972. Reports of all these were duly published in Watsonia.

When, recently, I tentatively started work on recording for a new Flora for Bedfordshire, Chris
was very enthusiastic and, although somewhat frail, she took on the responsibility for the tetrad
where she lived in Dunstable and was actively recording up to a few weeks before her death. She was
still excited to be taken out to a favourite site to see that plants recorded many years ago were still
flourishing. She was a mine of information on the sites for many rare local species.

In her last year she maintained an active interest in the botanical exploits in Bedfordshire and was
always keen to hear about her friends still involved in the BSBI and WES. Chris made an
outstanding contribution to the Bedfordshire and Warwickshire floras and will be remembered with
much affection by all who knew her.

BIBLIOGRAPHY

BOOK
Dony, J. G. & Dony, C. M. (with Boon, C. R. (ed.)). (1991). The wild flowers of Luton. 64 pp. Luton.

PAPERS

Dony, C. M. (1979). Puccinellia distans (reflexed saltmarsh-grass) in Bedfordshire. Bedfordshire naturalist 33:
68-69.

Dony, J. G. & Dony, C. M. (1986). Further notes on the flora of Bedfordshire. Watsonia 16: 163-172.

Runpb_e, A. J. & Dony, C. M. (1986). A provisional survey of Bedfordshire dandelions (genus Taraxacum,
family Compositae). Bedfordshire naturalist 40: 65-72.

OBITUARY 299

UNPUBLISHED REPORTS
Dony, J. G. & Dony, C. M. (1989). The course of the River Ivel from Beadlow to Shefford. 19 pp.

Dony, J. G. & Dony, C. M. (1989). Bedfordshire site reports. 316 pp.

MEETING REPORTS

Dony, Mrs J. G. (1970). BSBI membership. Watsonia 8: 86.

Dony, C. M. (1973). Rise and fall in BSBI membership, 1942-1972. Watsonia 9: 427.

Dony, C. M. (1982). Deschampsia danthonioides (Trin.) Munroe ex Benath. in Bedfordshire and Buckingham-
shire. Watsonia 12: 393.

Dony, C. M. & Dony, J. G. (1979). Maritime species in Bedfordshire. Watsonia 12: 393.

Dony, J. G. & Dony, C. M. (1973). Revision of the flora of Bedfordshire. Watsonia 9: 427.

Dony, J. G. & Dony, C. M. (1977). Some additional wool aliens in Bedfordshire. Watsonia 11: 428.

Dony, J. G. & Dony, C. M. (1982). Childing pink in Bedfordshire. Watsonia 14: 228.

Dony, J. G. & Dony, C. M. (1982). Therfield Heath, and Scales Park. Watsonia 14: 105 (field meeting report).

C. R. Boon

ig ree
~~ a
a ves
‘
4: ; ;
h y
awh, ; i:
ges ‘
j
ne rh Z
.. % An .
nT | P
bs oye :
oir ipa se
’
j
4
i
as
~
ae
°
a
‘
~

i he ase we Yih,
ie screen ots rd
oth ae ae ji r,
a scans Ree 4 tei od

7
*

Mi ee ry Gitex ae :
rie a iogepehiire .

: Sag mbes Selipooct
sce Sewn: Sheena
201} ‘aaa ere est

=<) tee NA — al; 5
q s mie witigelee Ec Ros —
a ee
| im hesed Feith

Watsonia 21: 307-309 (1997) 301

Report

ANNUAL GENERAL MEETING, 11 MAY 1996

The Annual General Meeting of the Society was held at the Jodrell Laboratory, Royal Botanic
Gardens, Kew, by kind invitation of the Director Sir Ghillean Prance, at 11.30 a.m. The President,
Mr D. A. Pearman took the Chair in the presence of 103 members.

Apologies for absence were read and Minutes of the 1995 Annual General Meeting, published in
Watsonia 21: 151-154 (1996), were accepted as correct, approved and signed by the President.

REPORT OF COUNCIL

The President took members through the Report of the Council, which had been previously
circulated to members, commenting on the main achievements of the Society during the year,
especially the launch of Atlas 2000 and the employment of Mr C. S. Crook as BSBI Co-ordinator
and Dr T. Dines as Atlas 2000 organiser. Mrs M. Briggs was congratulated on producing another
excellent report. The adoption of the Report was accepted unanimously.

HON. TREASURER’S REPORT AND ACCOUNTS

Mr M. Walpole, Hon. Treasurer, presenting the Accounts, invited queries. Dr T. G. F. Curtis asked
about the possibility of setting up a credit card system to ease payment of subscriptions, by Irish and
overseas members. The Treasurer reported that he was investigating the possibility. The adoption
of the report was proposed by Mr R. M. Burton, seconded by Mr P. A. Harmes and carried
unanimously.

ELECTION OF PRESIDENT-ELECT

The President took great pleasure in proposing Council’s nomination of Mrs M. Briggs. She had
been the Society’s Hon. General Secretary for 24 years and a Council member for over 30 years as
she had previously been Secretary of Meetings Committee. He could think of no other person, who
had done so much for the Society over so many years, who was better fitted to be President. The
adoption of Mrs M. Briggs as President-elect was proposed by Mrs M. Lindop, seconded by R. G.
Ellis and carried with acclamation. On behalf of Council, the President then presented Mary with a
gold brooch depicting the Society’s bluebell emblem. Mary responded that she was “‘lost for words —
almost.”

ELECTION OF VICE-PRESIDENT

The President commented on the importance to the Society of having experienced members as
Vice-Presidents who could assist the Chair when necesssary. There was one nomination, Dr R. J.
Gornall. His election was proposed by the President, seconded by Mr R. G. Ellis and carried
unanimously.

302 REPORT
ELECTION OF HON. GENERAL SECRETARY

As Mrs M. Briggs had now been elected President-elect, the President stated that there was a
vacancy for the post of Hon. General Secretary. Council had nominated Mr R. G. Ellis. His election
was proposed by Mrs M. Briggs, seconded by Miss A. Burns and carried unanimously.

RE-ELECTION OF HON. TREASURER

The President paid tribute to the enormous amount of work carried out by Mr M. Walpole on behalf
of the Society and his re-election was proposed by Mr D. E. Green, seconded by Dr N. K. B.
Robson and carried unanimously with applause.

The President then warmly thanked the Editors of Watsonia and BSBI news, the compiler of BSBI
Abstracts, and all Representatives on Council and other Committees for the hard work they carried
out, voluntarily, on behalf of the Society. This was greeted by applause.

ELECTION OF COUNCIL MEMBERS

In accordance with Rule 11 nominations had been received for Mrs P. P. Abbott, Dr F. J. Rumsey
and Dr M. C. Sheahan. Profiles had been published; election of these members was proposed by
Comd. J. M. W. Topp, seconded by Mr P. A. Harmes and carried unanimously.

ELECTION OF HONORARY MEMBER

In nominating Mr R. G. Ellis, the President paid tribute to his work on behalf of the Society in
Wales and especially as Editor of BSBI news as he had now produced 30 issues over ten years. He
also mentioned that the nomination had been made before Mr Ellis’ acceptance of nomination as
Hon. General Secretary. The election was carried unanimously with warm applause.

PRESIDENT’S AWARD
David Bellamy (President, Wild Flower Society) and David Pearman (President, BSBI) had great
pleasure in giving the award to Mr C. D. Preston for his Pondweeds of Great Britain and Ireland
(BSBI Handbook No 8).
RE-ELECTION OF HONORARY AUDITORS
The President warmly thanked Grant Thornton, West Walk, Leicester, for their exemplary auditing
of our accounts. Their re-election as Hon. Auditors was proposed by Mr P. H. Oswald, seconded by

Mrs M. E. Perring and carried unanimously with appreciation and applause.

There being no other business, the meeting closed at 12.05.

GwynNn ELLIS

FIELD EXCURSION HELD IN CONJUNCTION WITH THE A.G.M.

WAKEHURST PLACE, SUSSEX (Vv.c. 14) 12 May 1996

Forty seven members attended this meeting at ‘““Kew in the Country’. After making friendly
greetings in the car park we repaired to a lecture hall to hear Roger Smith, the Head of Seed

REPORT 303

Conservation at Wakehurst, give us a potted history of Kew’s arrival at Wakehurst followed by an
interesting insight into the collection of seeds and their storage. This was followed by a visit to the
Seed Bank where drying, sorting and packaging were demonstrated. An invitation to enter the
—20° C storage chamber was declined, we preferring to stay in the comfort of the drying chamber.

Leaving the Seed Bank we were conducted around the newly laid out trial beds and greenhouses
where seeds of rare and threatened species of our native flora are tested for viability. Some members
were pleased to see Veronica peregrina (American Speedwell) as a weed in one of the plots.

After lunch we reconvened outside the mansion ready to tour the gardens. The party was divided
up into three groups, each group covering the same areas but in different directions so as not to
clash. Andrew Jackson, the Conservation and Woodland Manager at Wakehurst led one group
whilst Chris Clennett the Nursery Manager took his group away in the opposite direction. Arthur
Hoare a local BSBI member and frequent visitor to Wakehurst led the third party.

We stopped first at the Slips to admire the many patches of Lathraea clandestina (Purple
Toothwort) which has been in the gardens since before Kew took on the lease 31 years ago. It is now
appearing quite naturally in many places both in and outside the gardens. Also here was Orchis
morio (Green-winged Orchid), introduced after a rescue operation from a redevelopment site in a
nearby town. Close by Orchis laxiflora (Loose-flowered Orchid) was just coming into flower; this
was an experimental introduction which has proved very successful and the plants are now
increasing by natural regeneration.

Our next stop was to see Hymenophyllum tunbrigense (Tunbridge Filmy-fern) which for my
group meant a fairly steep climb up a bank to pay homage to this last remaining patch within the
confines of the garden although it is to be found in many other areas throughout the estate. Our
president was rather pleased to be shown this as it was a new plant for him.

A number of sectors within the garden are being managed for the conservation of our native flora.
It was at one of these sectors that we made our third stop, a sloping site with a few wet flushes
mostly acidic but one which was slightly alkaline. This gave a diverse flora and kept the members
happy for quite a while. With the season being a little late only a few plants were in flower; one,
Valeriana dioica (Marsh Valerian) was much appreciated. It was the Carex in this sector that created
the most interest with C. laevigata (Smooth-stalked Sedge), C. pallescens (Pale Sedge), C. pendula
(Pendulous Sedge), C. sylvatica (Wood-sedge) and C. viridula subsp. oedocarpa (Yellow-sedge) all
found and identified.

Continuing our perambulation past sandstone outcrops to a bank where Luzula forsteri (Southern
Wood-rush), L. pilosa (Hairy Wood-rush) and their hybrid, L. x borreri, were seen all within a
couple of metres of each other. This aroused the interest of some of our northern members who are
unfamiliar with L. forsteri.

The meeting ended in the tea room for well earned refreshments and genial conversation.

Our thanks are due to Norman Robson for making the arrangements and to the various leaders
for making this a most enjoyable day.

A. C. HOARE

~*~

eee

stron

! 4 : eh ee a oy : ee : =, !
, ; Ly es . 7, ‘ 7 on F
7 Pu whe ‘ 14 U5 ‘ nak Pe ; ; : ‘

r ig he GAREY GR STS Ab a ad peers veo ti sc as
EOE REA TO SI TTY 6 SUIS MAES Bae wos sy Jil eaigee elaaeee to abe Te a ak ~

gs tity S 4 ate at r ; ; P : ee:
io i J rongeeD 4 ry : 4 "et. Sir ra oh , : ae
- eld rae ; i eines yee Fav ih hae, Wath eet Nae weg ov oir ge is dee
: : oe wv ' Sl re ee
’ ; i "4 (‘ii Pe Thue
f i ¢
- 4 J oO
We ; At
4 j oe : ms
hy ty De Wee phe el de fe! an bod
aD | A ’ nye Pre ale rie
} 4 v J soe ] (i
é % hd
? PAP
: - é <p v =!
7 A - by 8 tiryrgnd Peis
' -
' ; Core ;
+ : ; .
64 eke if 4 é v3 reps bei
. 7 =
;
Py qo . i,
H
od 1
hy
1
=
¥
i
7 a

INSTRUCTIONS TO CONTRIBUTORS

Scope. Authors are invited to submit Papers and Notes concerning British and Irish vascular plants,
their taxonomy, biosystematics, ecology, distribution and conservation, as well as topics of a more
general or historical nature. Authors should consult the Hon. Receiving Editor for advice on
suitability or any other matter relating to submission of manuscripts.

Papers and Notes must be submitted in duplicate, typewritten on one side of the paper, with wide
margins and double-spaced throughout. Submission of copy on computer discs is acceptable but
must also be accompanied by two hard copies of the text, double spaced throughout.

Format should follow that used in recent issues of Watsonia. Underline where italics are required.
Names of periodicals should be given in full, and herbaria abbreviated as in British and Irish
herbaria (Kent & Allen 1984). The Latin names and English names of plants should follow the New
Flora of the British Isles (Stace 1991). Further details on format can be obtained from the Hon.
Receiving Editor.

Tables, figure legends & appendices should be typed on separate sheets and attached at the end of
the typescript.

Figures should be drawn in black ink or be laser-printed and identified in pencil on the back with
their number and the author’s name. They should be no more than three times final size, bearing in
mind they will normally be reduced to occupy the full width of a page. Scale-bars are essential on
plant illustrations and maps. Lettering should be of high-quality and may be done in pencil and left
to the printer. Black and white photographs can be accepted if they assist in the understanding of the
article.

Contributors must sign a copyright declaration prior to publication which assigns the copyright of
their material to the Botanical Society of the British Isles. Twenty-five offprints are given free to
authors of Papers, Notes and Obituaries; further copies may be purchased in multiples of 25 at the
current price. The Society takes no responsibility for the views expressed by authors of Papers,
Notes, Book Reviews or Obituaries.

Submission of manuscripts

Papers and Notes: Dr B. S. Rushton, School of Applied Biological and Chemical Sciences,
University of Ulster, Coleraine, Co. Londonderry, N. Ireland, BT52 1SA.

Books for Review: C. D. Preston, Biological Records Centre, Monks Wood, Abbots Ripton,
Huntingdon, PE17 2LS.

Plant Records: the appropriate vice-county recorder, who should then send them to C. D. Preston,
Biological Records Centre, Monks Wood, Abbots Ripton, Huntingdon, PE17 2LS.

Obituaries: Dr B. S. Rushton, School of Applied Biological and Chemical Sciences, University of
Ulster, Coleraine, Co. Londonderry, N. Ireland, BTS2 1SA.

Back issues of Watsonia are handled by Dawson UK Limited, Cannon House, Folkestone, Kent,
CT19 5EE to whom orders for all issues prior to Volume 20 part 1 should be sent.

Recent issues (Vol. 20 part 1 onwards) are available from the Hon. Treasurer of the

B.S.B.I., 68 Outwoods Road, Loughborough, Leicestershire, LE11 3LY.

Parser

r I i il

BRARIES

[ HNN

Contents

PEARMAN, D. Presidential Address, 1996: Towards
and scarce plants ... A Ce ae .

| STEVENS, D. P. & BLACKSTOCK, T. H. Gynodioeey inB
Eriophorum vaginatum L. (Cyperaceae)

ARMSTRONG, J. V. A morphometric study of the ust
campestris (L.) DC., L. multiflora subsp. multiflora ( i
__L. congesta (Thuill). Lej. (Juncaceae) in the Bri
Kirkpatrick, A. H. & MacDonatp, A. J. Calluna
dieback in Cie Scotland Pe
NoTEs ae

Blackstock, T. H. & Jones, R. A. Juncus ps Ww
rediscovered near its eee cea in ira (v

(Ronee in Britian Sugeest ee nee

FitzGerald, R., Field, G. D. & Chatters, C.The status of ic
Gaertner in Britainin 1995... ve ee oe

Primavesi, A. L. A Wolley-Dod letter of historical i in aes

Reynolds, S. C. P. Alien pines at Foynes Port, pe
1988-1994 ooo a ee =

Reynolds, S. C. P. The e putative eae between two. tea:

botanical records by Thomas Sockett (177-1858)
Book REVIEWS ...__... Oe es ee
OBITUARY Jee S Se

REPORT | Saas
Annual General Meeting, 11 May 1996... |

Published by the Botanical Society of the British Isles
ISSN 0043-1532

Filmset by =
WILMASET LTD, BIRKENHEAD. MERSEYSIDE
Printed in Great Britain by ae Dis, Perea
EATON PRESS LIMITED, WESTFIELD ROAD, WALLASEY, :, MERSEYSIDE. Das,