Watsonia 27: 1–21 (2008) SPUR LENGTH IN 1

Is spur length of Platanthera in the British Isles adaptively optimized or an evolutionary red herring?

R. M. BATEMAN Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS*

and R. SEXTON 22 Alexander Drive, Bridge of Allan, Stirling FK9 4QB

ABSTRACT inexperienced researchers into morphometric invest- igations of the British and Irish flora. A consistent methodology was used by the authors, and by several volunteers lacking experience of KEYWORDS: adaptation, allometry, citizen scientist, morphometrics, to measure the length of the co-evolution, geographic cline, measuring error, nectiferous labellar spur in populations of two sister morphometrics, spur, orchid, Platanthera species within the orchid Platanthera, P. chlorantha, , pollinator attraction. bifolia and P. chlorantha. Although they differ morphologically, primarily in details of the column and labellar spur, these two species appear INTRODUCTION molecularly identical. Much previous research using Platanthera as an evolutionary model system THE GENUS PLATANTHERA AS AN EVOLUTIONARY assumed that spur dimensions play a critical role in MODEL SYSTEM attracting specific pollinators and thus in determining The Butterfly Orchids (genus Platanthera fitness. The present analysis of data from 1876 Rich.) have benefited from recent molecular individuals of 73 Platanthera populations from the phylogenetic studies (based on the nuclear British Isles, plus 10 populations from Continental Europe, suggests that spur length in both species ribosomal ITS region: Hapeman & Inoue 1997; follows a latitudinal cline, diminishing northward at Bateman et al. 2003), which have resolved the an average rate of 2·2% per 100 km. More southerly genus into six monophyletic sections (including populations are more likely to occur in shaded Piperia, which was previously widely recog- habitats, where their spurs (especially P. bifolia) are nized as a separate genus in North America: on average longer than in open habitats. The present Bateman et al. 2003; Lauri 2007). Several data are insufficient to determine whether this trend species of Platanthera have increasingly been is adaptive and pollinator-limited, as traditionally used as models of evolution in general and believed, or allometric and resource-limited, as putative adaptation in particular. In North suggested here. Most previous floras and mono- graphs (a) offer ranges for spur length that are America, hybridization, allopolyploid spec- seriously erroneous, and (b) are surprisingly reluctant iation and facultative autogamy have been to admit the occurrence of hybrids between the two documented in the P. dilatata–P. aquilonis species, preferring to prioritise speculative assertions aggregate of Section Limnorchis (Wallace of strong fidelity between pollinator and orchid 2004, 2006), while in Section Lacera above field observations recording morphologies inbreeding depression and mutational processes intermediate between the two parental species. were described in P. leucophaea (Wallace Statistical comparisons of various kinds of 2003; Holzinger & Wallace 2004) and polli- duplicated datasets show that, given an explicit nation frequency, outcrossing-to-inbreeding protocol, acceptably consistent measurements can be achieved among multiple, largely inexperienced, ratio and selection for spur length were analysts. The main source of potential measurement explored in P. lacera (Little et al. 2005). error, premature measurement of flowers prior to full The remaining evolutionary studies have anthesis, is easily avoided. The success of this focused on Section Platanthera – specifically, project should open the way for more ambitious on the classic Eurasian pairing of P. bifolia and initiatives that recruit substantial numbers of P. chlorantha. The former has been used to

*e-mail: [email protected] 2 R. M. BATEMAN AND R. SEXTON investigate variable selection for male and attracting those pollinators and persuading female function (Maad 2000; Maad & Alex- them to probe sufficiently deeply to acquire andersson 2004), while the latter has yielded pollinia. Only a minority of terrestrial orchid details of nectar secretion and resorption species in the Northern Hemisphere offer their (Stpiczynska 2003a, b). Perhaps the most pollinators a genuine food reward (usually interesting investigations have compared and nectar); the remainder deceive insects into contrasted aspects of the two species. A few mistakenly believing that they will be rewarded studies have explored the potential of these with food or, less frequently, sex (e.g. van der species for non-adaptive macroevolution Cingel 1995; Cozzolino & Widmer 2005). through saltation (a heritable, genetic or Although there have been repeated evolu- epigenetic modification that is expressed as a tionary transitions between the rewarding and profound phenotypic change across a single non-rewarding conditions (Bateman et al. generation and results in a potentially indepen- 2003; Cozzolino & Widmer 2005), the genus dent evolutionary lineage), notably via Platanthera appears to be uniformly rewarding modification of the labellum to resemble the (Hapeman & Inoue 1997), substantial quan- sepals (a transition termed pseudopeloria: tities of nectar being secreted by labellar spurs Bateman 1985; Bateman & DiMichele 2002; (Stpiczynska 2003a, b) that differ considerably Bateman & Rudall 2006b). However, most in mean length among the many species in the studies have focused on aspects of presumed genus. adaptation. Nilsson (1978, 1983) used morpho- metric measurements from both herbarium RATIONALE OF THE PRESENT STUDY collections and in situ populations in Not surprisingly, given their reputedly strong Scandinavia to quantify the features that influence over pollination efficiency, the spurs encouraged placement of the pollinia on the of orchids in general (Darwin 1877; Rudall & tongues of pollinating moths in P. bifolia and Bateman 2002; Box et al. in press), and of the eyes of other, closely related moths in P. Platanthera in particular (Nilsson 1983, 1985; chlorantha. As predicted from morphological Maad & Nilsson 2004; Little et al. 2005; observations by Summerhayes (1951), optim- Claessens & Kleynen 2006; J. Maad pers. ization of morphological characters across ITS- comm. 2007), have long received considerable based phylogenies suggested that the eye attention from evolutionary biologists. attachment of pollinia in P. chlorantha evolved The present authors independently estab- from the more conventional proboscis lished projects that involved extensively attachment that characterizes P. bifolia measuring the labellar spurs of the Greater (Hapeman & Inoue 1997; Bateman 2005). The Butterfly-orchid ( two species differ in fragrance chemotypes (Custer) Rchb.) and Lesser Butterfly-orchid (P. (Tollsten & Berstrom 1993). Admixed bifolia (L.) Rich.) in the UK. Working with populations of these two species containing Paula Rudall, Bateman began his project in phenotypic “intermediates” (i.e. presumed 2003 and focused on southern England hybrids) have been studied in the Low (Bateman 2005). This study, which sought to Countries (Claessens & Kleynen 2006) and better delimit the presumed boundary Scandinavia (Nilsson 1985), where the effic- separating the two species, involved measuring iency of pollinium import and export between 42 characters from each , which limited the two species was estimated (Maad & the number of individuals that could Nilsson 2004). Thus far, DNA sequencing realistically be studied. In addition, many of studies have revealed extraordinary similarity the subjected to morphometric analysis between the two taxa, despite the fact that they were also subjected to sequencing of the are morphologically distinct and so universally nuclear ribosomal ITS region (Bateman, James accepted as bona fide species (Bateman 2005; & Rudall in prep.). Bateman subsequently Bateman, James & Rudall in prep.). added some data on spur lengths of P. bifolia Most of the evolutionary scenarios emerging from northwest Scotland. from these studies paid particular attention to Sexton’s project began in 2005 and focused the key role of the morphology of the column on south-central Scotland (Sexton & McQueen (and, to a lesser degree, of the pollinia them- 2005). Having summarized prior evidence of selves) in placing the pollinia at appropriate considerable variation in the spur length of P. locations on visiting Lepidoptera, and on the chlorantha across Europe, this project sought role of the nectar-secreting spur in initially to uncover and explain local variation in spur SPUR LENGTH IN PLATANTHERA 3

FIGURE 1. Lateral view of flower of Platanthera chlorantha, illustrating the method used to measure spur length in the present study (note that in the actual study the flowers remained attached to their parent plants). Photo: Roy Sexton. lengths within the chosen study area. Other MATERIALS AND METHODS likely factors attracting pollinators to the flowers, including the volume of nectar and We made considerable efforts to formulate composition of scent, were considered. Success clear instructions that contained sufficient in attracting pollinators was judged indirectly, detail to ensure consistency of measurement through frequency of fruit set and frequency of among different analysts (Bateman & Sexton viable seed within individual fruit (Sexton & 2007). Spur length is a more obviously self- McQueen 2005). defining parameter than, for example, spur It soon became evident that, given their width (see below). It is most readily measured contrasting geographical foci, our respective by placing a 150 mm steel rule against the back datasets on spur length would increase consid- of one of the lateral sepals (Fig. 1), with the erably in value if pooled. In addition, it was flower still firmly attached to the inflorescence. clearly highly desirable to extend the geo- As flower size has recently been shown to graphic coverage of the project to encompass decrease considerably from the base to the apex the whole of the British Isles, and ideally to of the inflorescences of several European expand into mainland Europe. The most orchid species (Bateman & Rudall 2006a), we obvious way of achieving these more ambitious specified that the measured flower should be goals within a reasonable timescale was to chosen from the middle of the inflorescence involve substantial numbers of inexperienced and be fully open. We requested that analysts researchers (often described as ‘interested gather a sample of 20 individuals, though the amateurs’ or, more recently in science policy small size of most Platanthera populations in circles, “citizen scientists” – we recognize that the UK meant that some sample sizes were neither term is especially appropriate). appreciably smaller (in partial compensation, Therefore, in April 2007 we asked members of other samples were substantially larger). the UK’s Hardy Orchid Society to measure Datasets were submitted electronically by most Platanthera populations in their local area and analysts, and were collated by Bateman in an donate the resulting data to us for collation Excel spreadsheet immediately after the 2007 (Bateman & Sexton 2007). flowering season had ended. 4 R. M. BATEMAN AND R. SEXTON

TABLE 1. SIGNIFICANCE TESTS OF COMPARISONS OF DUPLICATED ESTIMATES OF MEAN SPUR LENGTH OBTAINED FROM THE SAME POPULATION OF PLATANTHERA

Locality Species Analyst(s) and Year(s) Difference between mean values (mm) (a) Same analyst, same year (4) Wolstenbury chlorantha Stott/Pearce, 2007 0·79 Latterbarrow chlorantha Fenton, 2007 1·83 Levens Wood chlorantha Fenton, 2007 0·37 Quoigs Meadow chlorantha Sexton, 2007 0·42 (b) Different analysts, same year (5) Newtimber Hill chlorantha Stott vs Pearce, 2007 1·12 Latterbarrow chlorantha Gendle vs Fenton, 2007 0·48 A590/A591 junction chlorantha Gendle vs Fenton, 2007 2·33* Waitby Greenriggs chlorantha Gendle vs Fenton, 2007 0·11 Kernsary bifolia Bateman vs D.+C. Hughes, 2007 0·17 (c) Different analysts, different years (1) Pewsey bifolia Bateman, 2004 vs A. Hughes, 1·35 2007 (d) Same analyst, successive years (10) Bomains Meadow chlorantha Sexton, 2005 vs 2006 0·68 Bomains Meadow chlorantha Sexton, 2006 vs 2007 0·91 Plean Park chlorantha Sexton, 2005 vs 2006 1·56 Plean Park chlorantha Sexton, 2006 vs 2007 1·06* Kippen (A) chlorantha Sexton, 2006 vs 2007 2·26** Kippen (B) chlorantha Sexton, 2006 vs 2007 1·65* Wester Balgair chlorantha Sexton, 2006 vs 2007 1·34* Braeleny Farm chlorantha Sexton, 2006 vs 2007 0·99 Quoigs Meadow chlorantha Sexton, 2006 vs 2007 0·49 Quoigs Meadow bifolia Sexton, 2006 vs 2007 2·21* (e) Same analyst, same year, different stages of anthesis (1) Carnan bifolia Bateman, 2007 3·38**

Significance: * = p <0·05, ** = p <0·01

RESULTS from Scotland by David and Christine Hughes). The net result was clustering of data- The results of the study are summarized in points in the Alps, southern England, Cumbria, Appendix 1 and Tables 1–2. By the close of the southern Scotland and northeastern Scotland 2007 field season, our combined database of (interestingly, the middle three areas constitute spur lengths contained 120 datasets (49 for P. the geographical “hot-spots” for P. chlorantha bifolia) totalling 1876 individual plants (624 explicitly identified in the UK by Foley & for P. bifolia); datasets ranged in sample size Clarke [2005]). from a single plant to 118 plants (Appendix 1). Fifteen datasets consisted of only one or two Of these 120 datasets, 33 were generated by plants, and so yielded data of very limited Bateman, 26 by Sexton and the remaining 61 value. Each of the remaining 105 datasets by other recorders (notably two from southern yielded values for the mean and sample England and five from Austria by Tony standard deviation, omitting a few rare plants Hughes, four from southern England by wherein spur development was clearly Katherine Stott and David Pearce, 11 from seriously retarded. A further 21 datasets (four Cumbria by Alan Gendle, a further four from for P. bifolia) duplicated other datasets based Cumbria by James Fenton, seven from west- on the same locality (Table 1). Most central Scotland by Sarah Longrigg, and 15 duplications of measurement were intentional, SPUR LENGTH IN PLATANTHERA 5

TABLE 2. COMPARISON OF SPUR LENGTHS FROM THE LOWERMOST, MIDDLE AND UPPERMOST FLOWERS IN TEN PLANTS OF TWO POPULATIONS OF PLATANTHERA FROM SOUTHWEST ENGLAND

Measurement Locality Bull’s Cross Pewsey Mean SD CV Mean SD CV Species bifolia + hybrids bifolia Flowers per inflorescence 14·20 3·08 21·7 11·30 3·56 31·5 Spur lengths (mm) lowermost 23·50 1·73 7·4 18·80 1·75 9·3 middle 23·80 1·58 6·6 19·75 1·32 6·7 uppermost 22·67 2·19 9·7 18·75 3·60 19·2

None of the pairwise comparisons of positions on the inflorescence is statistically significant. (Data collected by Tony Hughes). though three duplicated datasets from Cumbria, DISCUSSION one from Wiltshire and one from northeast Scotland resulted from wholly accidental (but nonetheless welcome) convergences between INTRAPOPULATION VARIATION AND POTENTIAL two independent analysts (Gendle vs Fenton, ANALYTICAL ERROR Bateman vs A. Hughes, Bateman vs D. and C. Values for mean and sample standard dev- Hughes, respectively). Five duplications were iations were used to calculate coefficients of conducted by different analysts in the same variation for each usable dataset. Coefficients year, one by different analysts in different of variation ranged from 6 to 15% and were years (Bateman vs A. Hughes), four by the typically 8–12%; this moderate level of same analyst in the same year, and ten by the variation relative to other features of the plant same analyst (Sexton) in different years. suggests a considerable degree of functional Another kind of duplication study (Bateman) latitude rather than the constraints inherent in compared plants within the same population strong selection pressure (cf. Bateman, James that had at least 50% open flowers with those & Rudall in prep.). There were, however, two where the lowermost flower was just beginning exceptions. The unusually low coefficient of to open. The last category of duplication study variation (2%) for Murroch is probably an (by A. Hughes) compared the spur lengths of artefact of the unusually small sample size flowers at the specified position in the middle (Appendix 1). However, the exceptionally high of the inflorescence with the lowermost and coefficient (23%) for the Bix Bottom uppermost flowers of the same inflorescence, population of P. bifolia may be biologically focusing on two populations of P. bifolia from meaningful, indicating that hybridization at the southwest England (Table 2). site is more extensive than was previously These duplicate datasets, summarized in supposed, and that some of the plants Table 1, allowed us to infer how much of the designated “pure” by us on the basis of overall observed variation in spur length could be morphology may actually have experienced attributed to measuring error or non-genetic introgression (see below). Indeed, since this (i.e. environmental and/or developmental) statement was first written, multivariate factors rather than genetic factors. analysis required re-identification of one of the Subsequently excluding duplicated datasets Bix plants originally assigned to P. bifolia as (retaining the dataset from the site that was the hybrid between P. bifolia and P. based on the largest sample size) left 83 chlorantha. datasets that could be statistically analyzed and Comparison of mean values and sample so used to seek geographically related patterns standard deviations for different phases of in spur length variation for each species. measurement of the same population (Tables 1, 6 R. M. BATEMAN AND R. SEXTON

2) provided much useful information. Four overall length of the spur in question). We populations were measured by the same analyst initially suspected that, in these cases of in the same year, the two phases of measure- divergence between analysts, different flowers ment typically occurring about a week apart. had been measured on the same inflorescence. The mean values differed on average by 3·2% Our original emphasis on measuring flowers at of the total spur length (range 1·3–6·3%), much consistent position in the middle of the inflor- of which can presumably be ascribed to escence was prompted by the study of Bateman sampling error. None of the disparities between & Rudall (2006a), who demonstrated that spur these pairs of means was statistically and especially labellum dimensions decreased significant (Table 1). significantly from the base to the apex of the A further five populations were analyzed by inflorescences of several species of European different analysts in the same year; four pairs orchids. In contrast, these trends were barely of measurements were taken within a week of evident in the single inflorescence of each other but the fifth pair, at Kernsary, were Platanthera chlorantha studied by them; separated by a remarkable three-week period. indeed, spur length appeared to peak slightly in The mean values differed on average by only the middle of the inflorescence (Bateman & 0·4% more than when the same analyst Rudall 2006a, fig. 9). collected both tranches of data (3·6%, range This pattern seemed most likely an artefact. 0·6–7·8%), suggesting that our efforts to ensure However, data collected for the present study consistency among analysts in the way that by A. Hughes show that this is in fact a they measured spurs had been successful (but genuine trend in P. bifolia that can be detected see below). Admittedly, one such population (albeit without statistical significance) at the (A590/A591 junction near Kendal, Cumbria) population level. Table 2 shows maximum spur showed a difference between the two pop- length in the centre of the inflorescence in both ulation means that was marginally statistically study populations, with lower coefficients of significant, but contrasting subpopulations may variation than are evident in either the have been sampled in this extensive, linear lowermost or the uppermost flowers. We roadside population, which encompassed a suspect that the unusually large, widely spaced range of habitats. flowers of these Platanthera species permit the More interesting are ten comparisons made development of a cylindrical rather than a by Sexton in south-central Scotland, where in conical inflorescence. Whatever the reason, each case the same population was sampled (in there is evidently less risk of positional most cases extensively) in successive years. differences causing measuring error in These comparisons showed typical mean Platanthera than in most other European length differences of 5·1% (range 1·8–10·3%), orchids. and five of the ten comparisons were statist- Our data suggest that the largest potential ically significant, one strongly so (Table 1). In source of error is the stage of anthesis reached four of these five cases, spurs were on average by the inflorescence. At the Carnan shorter in 2007 than in 2006, suggesting that (Benbecula) population of P. bifolia, Bateman there is a significant environmental influence sampled 20 inflorescences where at least the on mean spur length from year to year; lower half of the inflorescence bore fully open presumably, this discrepancy primarily reflects flowers, and a further 10 inflorescences where local weather conditions. The single comp- even the lowermost flowers were only just arison of means determined by different beginning to open or were about to open. We analysts in different years (Pewsey) also fell estimate that anthesis in the two cohorts of naturally into this category; although strictly inflorescences only differed by less than a not statistically significant, the two means week, yet there was a statistically highly differed by 6·8%. significant difference of 24% in their respective Another interesting experiment was mean spur lengths (10·9 mm vs 14·2 mm: conducted by Katherine Stott and David Table 1). This observation supports recent Pearce, who independently measured the same scanning electron microscope studies which small number of plants at Newtimber Hill, suggested that spur elongation occurs surp- Sussex. For most plants examined their risingly late in the development of the flowers respective estimates of spur length differed by of European orchids (Box et al. in press); it at most only one unit of measurement (i.e. 0·5 clearly demonstrates that the measured flower mm), but in a few cases measurements differed must be fully open before relevant and reliable by up to a worrying 3 mm (i.e. c. 10% of the spur dimensions can be obtained. SPUR LENGTH IN PLATANTHERA 7

TABLE 3. RANGES OF SPUR LENGTHS GIVEN FOR PLATANTHERA SPECIES IN FLORAS OF THE BRITISH ISLES AND CONTINENTAL EUROPE

Flora/monograph Spur length (mm) P. bifolia P. chlorantha (a) British Isles Godfery (1933) 15–25 25–30+ Clapham et al. (1962) 15–20 19–28 Sell & Murrell (1996) 15–20 19–28 Stace (1997) 15–20 19–28 Harrap & Harrap (2005) 13–23(–27) 19–35 (b) Europe Webb (1980) 25–30 18–27 Davies et al. (1983) 25–30 18–27 Baumann & Künkele (1988) 13–40 20–401 Buttler (1991) 12–23 18–40 23–412 Bournérias & Prat (2005) 20–30 20–45 Delforge (2006) 13–23 18–41 20–412

1Given erroneously in the original text as 2–4 mm. 2Both authors distinguished a short-spurred calcifugic ecotype in northwest Europe from a longer-spurred calcicolous ecotype occurring further south and east.

Lastly, conducting our own measurements covering (a) the British Isles and (b) part or all led to a better appreciation of qualitative of Continental Europe. differences in spur morphology between the Within the British Isles, Sell & Murrell two species. In particular, Ettlinger (1997) (1996) and Stace (1997) followed the dim- noted that the spur of P. bifolia appears to be a ensions given for spurs of both species by simple cylinder, whereas that of P. chlorantha Clapham et al. (1962), but their ranges are too is expanded towards the apex. We found spurs narrow for P. bifolia and are skewed toward of both species to be bilaterally compressed the lower end of the scale for P. chlorantha, (i.e. oval rather than circular in transverse failing to accommodate the long-spurred (mean section), but in P. chlorantha there was also a >32 mm) populations that predominate in significant dorsiventral increase in wall southern England (Fig. 3). Only the ranges thickness in the distal half to two-thirds of the given by Harrap & Harrap (2005) match our spur. This often conferred on the spurs of P. own observations for both Platanthera species, chlorantha a gently sigmoid shape when though we have not yet located any of the viewed laterally (Fig. 1); these then require short-spurred (mean <23 mm) populations of straightening prior to measurement. P. chlorantha reported by all authors except Godfery (1933). PREVIOUS REPORTS OF SPUR LENGTH IN Moving on to mainland Europe, Webb P. CHLORANTHA AND P. BIFOLIA (1980, echoed by Davies et al. 1983) gave Appendix 1 and Figure 3 summarize our lower and upper limits to P. chlorantha that extensive data on Platanthera spur lengths for were less than those for P. bifolia, apparently the British Isles and our more sporadic data having given undue emphasis to the extra- obtained from populations occurring in ordinary reversal in this character reported Continental Europe. For comparison, in Table along the Baltic coast by Nilsson (1978, 1985). 3 we have abstracted ranges of spur lengths The ranges fail utterly to accommodate British given for the two species in several widely- populations. The remaining authors (Baumann read general floras and orchid monographs & Künkele 1988; Buttler 1991; Bournérias & 8 R. M. BATEMAN AND R. SEXTON

Prat 2005; Delforge 2006) succeeded in MUTATION encompassing variation in both species, but Visual inspection of the data matrix suggested only by offering ranges of spur length that are that frequencies of both presumed hybrid- so broad that they lose any potential for being ization and presumed mutation radically diagnostic of either species (and even then, reducing spur length were surprisingly low. Bournérias & Prat failed to accommodate the With regard to putative mutation, of 1876 lower end of the spectrum of spur length plants measured, only four plants (0·21%), recorded by us in both British and Continental apportioned equally between P. chlorantha and P. bifolia). P. bifolia, yielded abnormally short spurs Perhaps the most interesting decision was (defined as those less than 60% of the mean that taken by Buttler (1991) and Delforge length for the population). This result contrasts (2006) to separate two putative ecotypes of P. with, for example, data for Gymnadenia bifolia, partly on the basis of ecological conopsea and its close relatives, where the preference and partly on the basis of vegetative frequency of abbreviated spurs is substantially vigour and spur length. Delforge (2006) greater and has been inferred as a likely driver offered slightly more liberal spur-length ranges of speciation within the genus (Bateman 2005). than Buttler (1991), allowing the northwestern, Pseudopeloric individuals of P. chlorantha, later flowering, calcifugic ecotype (subsp. possessing only simplified sepal-like labella graciliflora Bisse, ? = var. robusta Seemen) to bearing greatly reduced (essentially vestigial) extend from 13–23 mm and the southern and spurs, have previously been reported from eastern calcicolous ecotype to cover the range Keltneyburn, near Perth (averaging approx- 20–41 mm. However, calcicolous populations imately 3% of the population: cf. McKean from the UK approximate both the upper 1982; Bateman 1985) and Kennishead, Glasgow boundary of the smaller ecotype and the lower (averaging approximately 6% of the boundary of the larger ecotype (cf. Appendix 1 population: Dickson 1990). During the present and Table 3). In Continental Europe, the four study in 2007, Sexton identified 21 widely calcicolous populations of P. bifolia measured distributed pseudopeloric individuals out of by Bateman & Rudall in southeastern France 802 P. chlorantha flowering at Bomains occupy the lower part of the range of the Meadow, Falkirk (i.e. 2·6%). calcicolous ecotype, whereas the calcicolous Recent assertions that these pseudopeloric population measured by Sexton in the plants could constitute incipient species Dolomitic Alps of northeastern Italy shows a (Rudall & Bateman 2002; Bateman & Rudall typical “calcifugic” mean spur length of 17 2006b) have arguably been weakened by mm. At best, the delimitation of these Sexton’s observation that the Bomains mutants infraspecific taxa clearly requires further verge on being male-sterile. We suspect that refinement. they are epimutants, reflecting hypermethyl- The above discussion may appear to have lation of the DNA bases rather than non- become enmeshed in unnecessarily fine details. synonymous mutation of the bases themselves; However, our aim is to address a broader a project is underway to test this hypothesis underlying theme. Specifically, spur length is (Bateman, Rudall, Kidner & James, in fact diagnostic of the two northern European unpublished). species of Platanthera, but only if the latitude of populations is taken into account (Fig. 3). HYBRIDIZATION When latitude is not taken into account, the Only five (0·27%) of the measured platantheras majority of the published ranges of spur length were accused by us of being hybrids between either fail to accommodate a substantial the two study species, identified primarily on proportion of Platanthera populations, if they the basis of their intermediate pollinaria are too narrow, or fail to suggest any diagnostic positions. All five plants were reported by potential, if they are too broad. We suspect that Bateman & Rudall: four from Bix Bottom, similar patterns are commonly reflected in Oxfordshire (Fig. 2), plus one less certain quantitative characters presented in diagnoses identification from St Anne’s Chapel, Cornwall of herbarium-based floras. There is no subs- (regrettably, no data were obtainable from its titute for access to geographically and gynostemium). Disappointingly, our published ecologically extensive field data. request for data from additional mixed SPUR LENGTH IN PLATANTHERA 9

FIGURE 2. Flowers of Platanthera bifolia (A left), P. chlorantha (C right) and a putative hybrid (B centre) from the Warburg Reserve, Bix Bottom, Oxfordshire. Scale bar = 5 mm. Photos: Richard Bateman. populations (Bateman & Sexton 2007) yielded [Nilsson 1983]. However, recent field obser- only one new dataset (gathered near Perth, by vations question this assumption, as moths Sexton), and no putative hybrids were reported have been photographed removing pollinia from this locality. On the other hand, the Bull’s from hybrids: Claessens et al. in press.) Cross locality studied by Tony Hughes In adopting this sceptical viewpoint, these contained convincing hybrids, though they authors were echoing Charles Darwin, who were not reliably separated from their co- took considerable exception to the decision of occurring parents. Bentham & Hooker (e.g. 1886) to treat the two The hybrids recorded from southern England British platantheras as mere varieties of a by Bateman & Rudall (Fig. 2), Tony Hughes single species – a decision taken on the (Hughes 2007) and Sarah Whild (Whild & grounds that “intermediate” forms occurred Lockton 2007) during the present survey have between them (cf. Summerhayes 1951). some significance, as most authors of floras Darwin was driven to uncharacteristically have been uncharacteristically reluctant to waspish exaggeration in his opposition, stating accept the occurrence of hybridization between that “the two forms differ in a large number of the two species of Platanthera, especially characters, not to mention general aspect and within the British Isles. For example, Godfery the stations inhabited”, and that “these two (1933) reported an occurrence of a putative forms certainly differ from one another more hybrid from Sligachan, Skye, but then qualified than do most species belonging to the same this observation by noting that the “hybrid” had genus”. Having noted that loose waxy scales on co-occurred with a pseudopeloric mutant the heads of visiting moths most likely confine (sensu Bateman 1985). This sceptical view- placement of pollinia to the proboscis (P. point was adopted by most subsequent authors, bifolia) or eyes (P. chlorantha), Darwin (1877, who argued that evidence of hybridization in pp. 73–4), apparently prompted in part by British platantheras was inconclusive and that Muller (1865), concluded that there is no most if not all putative hybrids were actually “doubt that the Larger and Lesser Butterfly aberrant individuals of one of the parents (cf. Orchids are distinct species, masked by close Hunt 1975; Sell & Murrell 1996; Stace 1997, external similarity” (thereby somewhat in press; Foley & Clarke 2005; Harrap & undermining his own earlier statement Harrap 2005). Further questionable arguments regarding their appreciable morphological frequently deployed against the existence of distinctness). Given the continuing hybrids included supposed differences in uncertainties over the nature of morphological habitat preference, peak flowering period and, intermediates between these species, the most bizarrely, the presumed reduced fitness of surprising failure thus far of molecular the hybrids caused by suboptimal presentation systematists to identify genetic markers that of the viscidia to pollinators. (This is actually reliably separate the two species is particularly an argument against introgression, rather than unfortunate – indeed, it once again raises against hybridization per se, since the key questions regarding the now generally accepted assumption is that primary hybrids could only status of these taxa as separate species act as pollen recipients, not as pollen donors (Bateman 2005; Bateman et al. in prep.). 10 R. M. BATEMAN AND R. SEXTON

FIGURE 3. Mean spur length (mm) regressed against latitude for populations of Platanthera bifolia (left, n = 38, r2 = 0·56) and P. chlorantha (right, n = 43, r2 = 0·42). Shade and open habitats are also distinguished. The dashed line connects mean values for both parents and the putative hybrids at the Bix Bottom locality; the two juxtaposed arrowed populations of P. bifolia show evidence of introgression. The Alpine population of three flowering plants labelled “indeterminate” was reported as P. bifolia but has a mean spur length more characteristic of P. chlorantha; this, together with the hybrid group, was omitted from the regression analysis.

Returning to the present study, the putative chlorantha – an extraordinary character hybrids from Bix had spur lengths intermediate reversal relative to the more typical condition between those of the parental species co- (Bateman 2005). Indeed, intermediacy of spur existing at the two relevant localities dimensions has characterized all of the many (Appendix 1, Fig. 3). Indeed, we suspect that quantitative case-studies of hybridization hybridization in the Bix and Bull’s Cross among spurred European orchids known to us populations may be more extensive than we (reviewed by Bateman & Haggar in prep.). originally believed, since spur-length measure- This reliable intermediacy renders extra- ments for supposedly pure plants of P. bifolia ordinary the report of extensive hybrid-ization in these two populations are the longest of any in two mixed populations of Platanthera in the in southern England (Fig. 3), suggesting the Netherlands, where the putative hybrids show genetic influence of co-occurring P. longer average spur lengths (32 mm) than co- chlorantha. Similar observations of spur existing populations of both P. chlorantha (26 intermediacy in hybrids were made on larger mm) and P. bifolia (23 mm) (Claessens & numbers of putative hybrids of P. chlorantha × Kleynen 2006; Claessens et al. in press). It bifolia (= P. × hybrida Bruegger) along the seems likely that the evolutionary processes Baltic coast of Sweden by Nilsson (1985), operating in these ambiguous Dutch though here P. bifolia typically has spurs that populations are more complex than simple are substantially longer than those of P. introgression. SPUR LENGTH IN PLATANTHERA 11

GEOGRAPHICAL AND ECOLOGICAL TRENDS differences could, for example, ultimately Prior to the present survey, we wrote that “as reflect the greater relative degree of insolation expected, spurs of P. chlorantha were typically impacting on the (usually paired) leaves of considerably longer than those of P. bifolia plants at lower latitudes (though note that, (averaging 34 mm vs 19 mm) … Data from during the summer months when Platanthera Scotland (very limited in the case of P. bifolia) leaves are fully expanded, day length at least is suggest that P. bifolia retains its English longer in higher latitudes). Spur length would dimensions there, whereas the spurs of P. then be inferred to be under at most relatively chlorantha are significantly shorter” (Bateman weak and/or sporadic selection pressure (a & Sexton 2007, p. 61). The addition of conclusion reached, on the basis of more substantial datasets from southeast France, sophisticated data, for the North American P. the Alps, Cumbria and Scotland requires lacera by Little et al. 2005). substantial revision of that preliminary view- Furthermore, re-examination of the present point. Regression of mean spur length against data (Fig. 3) suggests that application of more latitude (Fig. 3) revealed strong correlation complex algorithms for fitting curves to the coefficients, especially for P. bifolia (r2 = 0·56, spur-length data might ultimately achieve a vs 0·42 for P. chlorantha), and suggested more exact fit. For example, the spur-length similar trends in both species; each shows a data for P. bifolia could legitimately be inter- reduction in mean spur length of approximately preted as showing a constant size of approx- 2·2% per 100 km of increasing latitude. Figure imately 18 mm for most of the latitudinal 3 also demonstrates that spur length dist- range, deviating only at the two extreme ends inguishes the two species only if latitude is of the latitudinal gradient (i.e. upwards in taken into account, as populations of P. bifolia southeast France and downwards in northwest from the Vercors in southeast France show Scotland). Similarly, data for P. chlorantha similar mean spur lengths to populations of P. could be considered to be constant for most of chlorantha occurring approximately 1300 km its latitudinal range, plateauing at approximate- to the north in southern Scotland (c. 26 mm). ly 28 mm (a similar figure also characterizes Those occurring a little further north of the this species in the Low Countries: Claessens & Vercors in the Alps also fit into the general Kleynen 2006), but to have deviated upward to trend, though the spread of populations appears values of 32–37 mm over much of southern relatively wide (even discounting the excep- England (a mean spur length of 32 mm was tional mean length of 35 mm obtained in also reported for a co-latitudinal population of Pettneu, western Austria, which is rendered P. chlorantha from Poland by Stpiczynska suspect by both the unusually small sample 2003a). Positive deviations in the spur length size of three plants and uncertain species of P. bifolia (mean values 27–34 mm) charac- assignment). terize much of Scandinavia, falling to 18–21 The obvious temptation is to interpret these mm in the birch of the north (Nilsson latitudinal trends as reflecting adaptation of the 1983, 1985; Tollsten & Berstrom 1993; Maad spur to proboscis length in the respective 2000). Moreover, on the Baltic island of Oland pollinators of these two orchid species. there has been divergence between short- However, data gathered by Bateman and co- spurred populations of P. bifolia in grassland workers from all major features of the plants of (means 19–23 mm) and those occupying decid- both species suggest that, at least in southern uous woodland (means 28–40 mm) (Nilsson England, most features of P. bifolia are on 1983; J. Maad pers. comm. 2007). average only two-thirds of the dimensions of This observation suggests that, beyond equivalent features in P. chlorantha (Bateman latitude, the other extrinsic factors that could 2005; Bateman et al. in prep.). A similar ratio potentially have influenced spur length are was evident among floral dimensions recorded those relating to habitat preference, particularly in south-central Scotland by Sexton soils (pH and moisture content) and degree of (unpublished). If this ratio is repeated across shade experienced by the orchids. Our present most of the (largely coincident) ranges of the data for soils are inadequate, though there two species, the latitudinal gradient in spur exists much anecdotal evidence that P. bifolia length observed in this study (Fig. 3) could is more tolerant than P. chlorantha of wet simply represent an allometric ratio across acidic soils. In contrast, we were able to most organs of the plant that is merely reflected amalgamate the habitat descriptions listed in in spur length. In this case, the overall size Appendix 1 into two broad categories of 12 R. M. BATEMAN AND R. SEXTON exposure to light: shaded (woodland and scrub) of reproductive isolation separating the two and open (grassland, heathland and moorland). species and different geographic/ habitat races Among our sampled populations, shaded within P. bifolia; indeed, the observed mechan- populations do not extend northward beyond ism has become a model system of plant– southern England for P. bifolia and beyond pollinator co-evolution (e.g. Hapeman & Inoue northern England for P. chlorantha. In the case 1997). of P. bifolia, within specific latitudinal zones, The classical model states that P. bifolia all populations occurring in shade have longer places its pollinaria anywhere along the length spurs than those occurring in the open (all of the moth’s proboscis, whereas P. chlorantha occur on the positive side of the regression line requires the moth to ram its head into the in Fig. 3). The pattern is less clear for P. column in order to attach the viscidia to the chlorantha, for which in addition we presently surface of its eyes. This contrast in the mechan- lack data from Continental Europe. In southern ism of pollinium placement would suggest that England, shaded populations appear to span a spur length should be more critical to the re- wider range of spur lengths than exposed productive success of P. chlorantha than to that populations (only shaded populations span the of P. bifolia. Setting aside sites that yielded upper end of the range of means, at 34–37 very small sample sizes and those showing mm). However, no significant difference evidence of introgression, coefficients of var- between shaded and open populations was iation for spur length within our study observed in northern England, and no shaded populations range from 5–18% for P. bifolia populations were recorded in Scotland. None- (typically approximately 11%) and 6–14% for theless, potential correlations between average P. chlorantha (typically approximately 9%: spur length and environmental parameters Appendix 1). Thus, if there is a contrast in sel- evidently merit more detailed scrutiny. ective pressure on spur length between the two Lastly, we note that although some of our species it does not appear to be a strong one. study populations of both Platanthera species Darwin (1877) and subsequent commentators encompassed several hundred flowering plants, (e.g. Summerhayes 1951; Nilsson 1983) argued the majority were small, many typically adamantly that only long-tongued moths effect yielding less than ten flowering plants in any pollination in European Platanthera. However, one year. Where such small effective pop- Darwin also persistently rejected observations ulation sizes pertain, deviation from previous of other natural historians, notably Müller distributions of spur length can be readily (1865), that many European orchids did not achieved via genetic drift, whereas the potential actually reward their pollinators, whereas for long-term directional or disruptive selection subsequent studies have proven Müller right; is greatly reduced (e.g. Tremblay et al. 2005). many orchid species succeed by deceiving Genetic drift merits consideration in any study rather than rewarding pollinators (e.g. van der of spur-length variation in Platanthera, though Cingel 1995; Neiland & Wilcock 1998; its random effects might have been expected to Cozzolino & Widmer 2005). Many authors obscure the apparent correlation between have noted that species offering genuine habitat preference and mean spur length. rewards to pollinators generally achieve sub- stantially higher frequencies of pollination than POLLINATOR SPECIFICITY REVISITED do co-existing food-deceptive species (Neiland As noted earlier, past assertions of strong & Wilcock 1998; Cozzolino & Widmer 2005). selection pressure on spur length have focused However, fewer authors have noted that this on morphological and ecological data for the comparatively successful pollination rate has two Platanthera species in Scandinavia not translated into greater numbers of (Nilsson 1978, 1983, 1985; Maad 2000; Maad rewarding than non-rewarding species or to & Nilsson 2004), where mean spur length of P. greater average size or number of populations bifolia in particular appears to show greater per rewarding species. And even fewer authors regional variation than is evident in our have noted that there is no obligation on measurements taken elsewhere in western pollinators to accept any reward offered. Europe. The Swedish data suggest that strong Surely, an insect that is demonstrably willing to pollinator specificity among different species be duped into pollinating an unrewarding of sphingid moth provides a meaningful level orchid will not balk at the prospect of SPUR LENGTH IN PLATANTHERA 13 pollinating a species that offers a reward that CONCLUSIONS the insect cannot satisfactorily exploit? Having thus hypothesized that at least some We have confidently identified latitudinal and occasional pollinators of Platanthera would habitat influences on mean spur length in show behaviour patterns that are unlikely to be Platanthera bifolia and P. chlorantha. Once influenced by spur length, we were briefly these factors have been compensated for, P. excited when our 2007 spur-length survey chlorantha maintained spurs approximately prompted Tony Hughes (2007) to capture 50% longer than those of P. bifolia throughout daytime images in Gloucestershire of a our latitudinal transect, though previous studies distinctive Oedermera beetle transporting (Nilsson 1983, 1985; Maad & Nilsson 2004; pollinia across a putatively hybrid Platanthera Boberg et al. 2007; J. Maad pers. comm. 2007) inflorescence at Bull’s Cross. Unfortunately, suggest that this apparently linear relationship more careful scrutiny suggests that the pollinia breaks down in parts of Scandinavia. We in question were obtained from adjacent plants recognize that detection of further patterns of of Neottia (Listera) ovata. Feasibility studies regional highs and lows would once again using moth traps at P. chlorantha sites in encourage consideration of potential local south-central Scotland showed that moths with adaptation of spur length to optimally fit the pollinia attached to their eyes were caught morphology of co-evolving pollinating moths, fairly readily (Sexton & McQueen 2005). and that our observations do not necessarily Moreover, John Knowler (pers. comm. 2006) reject previous selectionist interpretations. reported that catches included the Beautiful However, we do suggest that, in future, Golden Y (Autographa pulchrina), Silver Y (A. potential non-adaptive causes of variation in gamma) and Gold Spangle (A. bractea); all spur length, relating to allometric relationships three moths are prominent pollinators of P. with other structures of the plant and/or to chlorantha in Sweden (Nilsson 1978). vigour of growth influenced by habitat and Nonetheless, recent combined population latitude, are considered alongside more genetic and morphological studies of gene flow conventional co-evolutionary explanations. in other groups of European orchids also The new data have allowed us to formulate counsel caution when assuming strong polli- more precise questions regarding the evolution nator specificity. Even in the notionally most of spur length in the study species, even if they pollinator-specific genus, Ophrys, natural have not provided unequivocal answers. hybrids occur among all major species groups, Adding further spur-length data, preferably demonstrating both their interfertility and the targeted to fill current geographical lacunae in fact that pollinators frequently transfer pollinia our trans-European sampling, should help to between species (cf. Paulus & Gack 1990; clarify patterns of variation. However, the Devey et al. in press; S. Cozzolino pers. comm. above hypothesis of a non-adaptive allometric 2007). In addition, the ability to identify relationship between spur length and other pollinia through DNA sequencing (e.g. Soliva parts of the plant can only be tested effectively & Widmer 2003) means that we can assess by expanding the study to encompass measure- pollinator error rates, which even in Ophrys are ments of other characters from the same plants proving to be relatively high. Also, artificial that are presently yielding spur-length data. crosses among Ophrys species have demon- Top priority for such supplementary characters strated fertility in F1s, F2s and backcrossed would be metric measurements of the column/ progeny (S. Malmgren pers. comm. 2006; S. pollinia, labellum and (typically paired) Cozzolino pers. comm. 2007). Current (albeit expanded leaves. Data on the ontogeny of the limited) evidence suggests that P. bifolia and spurs of Platanthera, and on the underlying P. chlorantha may similarly introgress when- developmental genetic mechanisms, would also ever they co-exist in substantial numbers in the be welcome. Similarly, data on pollinator same or immediately adjacent habitats. specificity are needed across the full geograph- For the present, the jury is out regarding the ical ranges of both species before meaningful degree of reproductive isolation enjoyed by the interpretations of the possible relationship two European Platanthera species. Clearly, it between adaptation and speciation can be is desirable to gather data on the range and made. frequency of success of pollinators from other More generally, we hope that we have regions of Europe to test the long-standing successfully demonstrated that field-oriented hypotheses of strong specificity. projects recruiting inexperienced researchers 14 R. M. BATEMAN AND R. SEXTON can be successfully extended beyond the ACKNOWLEDGMENTS conventional limit of simply recording the presence of a particular species, into the realm We warmly thank the following members of of quantitatively measuring features of the Hardy Orchid Society for donating data to potentially high biological significance. the project (number of datasets is given in Returning to butterfly-orchids, we note that it is parentheses): Peter Daltry (1), Ian Denholm B.S.B.I.-inspired field surveys that identified (2), Jonathan Fenton (5), Alan Gendle (11), Platanthera bifolia as one of the most rapidly Lois and Nigel Harbron (2), Tony Hughes (7), declining species of flowering plants in the Christine and David Hughes (15), David British Isles (Preston et al. 2002; Braithwaite et Johnston (2), Harold and Jane Lambert (1), al. 2005), and led to a subsequent resurvey of Sarah Longrigg (7), Kathy Stott and Dave its localities in Scotland that relied largely on Pearce (5), Jonathan Tyler (1) and Iain Wright “amateur” botanists (Farrell 2006; Lavery 2007). (1). Several of these contributors showed It would therefore be appropriate if inexper- considerable inventiveness in designing their ienced researchers also contributed further own supplementary analytical approaches. scientific data with the potential to explain Salvatore Cozzolino, Jean Claessens and those declines. Such an explanation should be Barbara Gravendeel kindly shared their facilitated by the co-occurrence of a very observations on orchid pollination. Richard closely related species, P. chlorantha, that does Bateman thanks the Botanical Research Fund not appear to be experiencing a rapid decline and Botanical Society of the British Isles for and therefore constitutes an ideal comparative partially supporting his fieldwork through yardstick. Comparison of rare and/or rapidly small grants, and is grateful to Johanne Maad, declining species with closely related species Paula Rudall and an anonymous referee for that are common and/or rapidly expanding is critically appraising the manuscript. presently an under-exploited approach to exploring the British and Irish flora.

REFERENCES

BATEMAN, R. M. (1985). Peloria and pseudopeloria in British orchids. Watsonia 15: 357–359. BATEMAN, R. M. (2005). Circumscribing and interpreting closely related orchid species: Platanthera, Dactylorhiza and the crucial role of mutation. Journal of the Hardy Orchid Society 2(4): 104–111. BATEMAN, R. M. & DIMICHELE, W. A. (2002). Generating and filtering major phenotypic novelties: neoGoldschmidtian saltation revisited. In Q. C. B. CRONK, R. M. BATEMAN & J. A. HAWKINS (eds.), Developmental genetics and plant evolution, pp. 109–159. Taylor & Francis, London. BATEMAN, R. M. & HAGGAR, J. (in prep.). Infrageneric and intergeneric hybridization of Dactylorhiza Necker ex Nevski in the British Isles. Botanical Journal of the Linnean Society. BATEMAN, R. M., HOLLINGSWORTH, P. M., PRESTON, J., LUO, Y.-B., PRIDGEON., A. M. & CHASE, M. W. (2003). Molecular phylogenetics and evolution of Orchidinae and selected Habenariinae (). The Botanical Journal of Linnean Society 142: 1–40. BATEMAN, R. M., JAMES, K. E. & RUDALL, P. J. (in prep.). Morphological versus molecular divergence between two closely related species of Platanthera (Orchidaceae) in southern England: recent evolution with a strong allometric overprint. Annals of Botany. BATEMAN, R. M. & RUDALL, P. J. (2006a). Evolutionary and morphometric implications of morphological variation among flowers within an inflorescence: a case-study using European orchids. Annals of Botany 98: 975–993. BATEMAN, R. M. & RUDALL, P. J. (2006b). The Good, the Bad, and the Ugly: using naturally occurring terata to distinguish the possible from the impossible in orchid floral evolution. Aliso 22: 481–496. BATEMAN, R. M. & SEXTON, R. (2007). Survey of the spurs of European butterfly-orchids. Journal of the Hardy Orchid Society 4(2): 60–63. BAUMANN, H. & KÜNKELE, S. (1988). Die wildwachsenden Orchideen Europas. Kosmos, Stuttgart. BENTHAM, G. & HOOKER, J. D. (1886). Handbook of the British flora (5th edn). Reeve, London. BOBERG, E., ALEXANDERSSON, R., JONSSON, M., MAAD, J., AGREN, J. & NILSSON, L. A. (2007). Pollinator shifts and evolution of spur length in the moth-pollinated orchid Platanthera bifolia. Abstracts of the Sixth Biennial Conference of the Systematics Association (Edinburgh): 69. BOX, M. S., BATEMAN, R. M., GLOVER, B. J. & RUDALL, P. J. (in press). Floral ontogenetic evidence of repeated speciation via paedomorphosis in subtribe Orchidinae (Orchidaceae). Botanical Journal of the Linnean Society. SPUR LENGTH IN PLATANTHERA 15

BOURNÉRIAS, M. & PRAT, D., eds. (2005). Les Orchidées de France, Belgique et Luxembourg (2nd edn). Mezé. BRAITHWAITE, M. E., ELLIS, R. W. & PRESTON, C. D. (2006). Change in the British flora 1987–2004. Botanical Society of the British Isles, London. BRZSOKO, E. (2003). The dynamics of the island population of Platanthera bifolia in the Bierbrza National Park (NE Poland). Annali Botanici Fennici 40: 243–253. BUTTLER, K.-P. (1991). Field guide to orchids of Britain and Europe. Crowood, Swindon, Wilts. CLAESSENS, J. & KLEYNEN, J. (2006). Anmehungen zur Hybridbildung bei Platanthera bifolia und P. chlorantha. Journal Europäischer Orchideen 38: 3–28. CLAESSENS, J. GRAVENDEEL, B. & KLEYNEN, J. (in press). Cucullia umbratica L. als Bestäuber von Platanthera × hybrida Bruegg. in Süd-Limburg (Niederlands). Journal Europäischer Orchideen. CLAPHAM, A. R., TUTIN, T. G. & WARBURG, E. F. (1962). Flora of the British Isles (2nd edn). Oxford University Press, Oxford. COZZOLINO, S. & WIDMER, A. (2005). Orchid diversity: an evolutionary consequence of deception? Trends in Ecology and Evolution 20: 487–494. DARWIN, C. (1877). The various contrivances by which orchids are fertilised by insects (2nd edn). Murray, London. DAVIES, P., DAVIES, J. & HUXLEY, A. (1983). Wild orchids of Britain and Europe. Chatto & Windus, London. DELFORGE, P. (2006). Orchids of Europe, North Africa and the Middle East. A. & C. Black, London. DEVEY, D. S., BATEMAN, R. M., FAY, M. F. & HAWKINS, J. A. (in press). Friends or relatives? Phylogenetics and species delimitation in the controversial European orchid genus Ophrys. Annals of Botany. DICKSON, J. H. (1990). Aberrant Greater Butterfly Orchids (Platanthera chlorantha) in the Glasgow area. Glasgow Naturalist 21: 599–602. ETTLINGER, M. T. (1997). Notes on British and Irish orchids. Published by the author, Dorking, Surrey. FARRELL, L. (2006). Lesser Butterfly Orchid – Platanthera bifolia. SNH-BSBI-Plantlife survey leaflet, pp. 1–4. FOLEY, M. J. Y. & CLARKE, S. (2005). Orchids of the British Isles. Griffin Press, Cheltenham, Gloucs. GODFERY, M. J. (1933). Monograph and iconograph of native British Orchidaceae. Cambridge. HAPEMAN, J. R. & INOUE, K. (1997). Plant–pollinator interactions and floral radiation in Platanthera (Orchidaceae). In T. J. GIVNISH & K. J. SYTSMA (eds), Molecular evolution and adaptive radiation, pp. 433–454. Cambridge University Press, Cambridge. HARRAP, A. & HARRAP, S. (2005). Orchids of Britain and Ireland. A & C Black, London. HOLZINGER, K. E. & WALLACE, L. E. (2004). Bayesian approaches for the analysis of population genetic structure: an example from (Orchidaceae). Molecular Ecology 13: 887–894. HUGHES, A. (2007). Beetlemania. Journal of the Hardy Orchid Society 4(3): 88–90. HUNT, P. F. (1975). Platanthera Rich. In C. A. STACE (ed.), Hybridization and the flora of the British Isles, p. 488. Academic Press, London. LAURI, R. (2007). The systematic and phylogenetic study of the subgenus Piperia (Orchidaceae) and closely related Platanthera. Botanical Society of America abstracts (Chicago), p. 241. LAVERY, L. (2007). Species Action Framework report on Lesser Butterfly Orchid (Platanthera bifolia) [in Scotland]. Unpublished report to Scottish Natural Heritage. 30 pp. LITTLE, K. J., DIERINGER, G. & ROMANO, M. (2005). Pollination ecology, genetic diversity and selection on nectar spur length in (Orchidaceae). Plant Species Biology 20: 183–190. MAAD, J. (2000). Phenotypic selection in hawkmoth-pollinated Platanthera bifolia: targets and fitness surfaces. Evolution 54: 112–123. MAAD, J. & ALEXANDERSSON, R. (2004). Variable selection in Platanthera bifolia (Orchidaceae): phenotypic selection differed between sex functions in a drought year. Journal of Evolutionary Biology 17: 642–650. MAAD, J. & NILSSON, L. A. (2004). On the mechanism of floral shifts in speciation: gained pollination efficiency from tongue- to eye-attachment of pollinia in Platanthera (Orchidaceae). Biological Journal of the Linnean Society 83: 481–495. MCKEAN, D. R. (1982). × Pseudanthera breadalbanensis McKean: a new integeneric hybrid from Scotland. Watsonia 14: 129–131. MÜLLER, H. (1865). Beobachtungen en westfalischen orchideen. Verhandlungen des Naturalische Vereins fur Province Rheinlandeen Westfalischen Jahrg. 25(3): 36-38. NEILAND, M. R. M. & WILCOCK, C. C. (1998). Fruit set, nectar reward, and rarity in the Orchidaceae. American Journal of Botany 85: 1567–1581. NILSSON, L. A. (1978). Pollination ecology and adaptation in Platanthera chlorantha (Orchidaceae). Botanisk Notiser 131: 35–51. NILSSON, L. A. (1983). Processes of isolation and introgressive interplay between Platanthera bifolia (L.) Rich. and P. chlorantha (Custer) Reichb. (Orchidaceae). Botanical Journal of the Linnean Society 87: 325–350. 16 R. M. BATEMAN AND R. SEXTON

NILSSON, L. A. (1985). Characteristics and distribution of intermediates between Platanthera bifolia and P. chlorantha (Orchidaceae) in the Nordic countries. Nordic Journal of Botany 5: 407–419. PAULUS, H. F. & GACK, C. (1990). Pollinators as prepollinating isolating factors: evolution and speciation in Ophrys (Orchidaceae). Israel Journal of Botany 39: 43–97. PRESTON, C. D., PEARMAN, D. A. & DINES, T. D. (2002). New atlas of the British and Irish flora. Oxford University Press, Oxford. RUDALL, P. J. & BATEMAN, R. M. (2002). Roles of synorganisation, zygomorphy and heterotopy in floral evolution: the gynostemium and labellum of orchids and other lilioid monocots. Biological Reviews 77: 403–441. SELL, P. D. & MURRELL, G. (1996). Flora of Great Britain and Ireland. 5. Butomaceae–Orchidaceae. Cambridge University Press, Cambridge. SEXTON, R. & MCQUEEN, A. E. D. (2005). Greater Butterfly-orchids: Platanthera chlorantha (Custer) Reichenbach. Forth Naturalist and Historian 27: 77–90. SOLIVA, M. & WIDMER, A. (2003). Gene flow across species boundaries in sympatric, sexually deceptive Ophrys (Orchidaceae) species. Evolution 57: 2252–2261. STACE, C. A. (1997). New Flora of the British Isles (2nd edn). Cambridge University Press, Cambridge. STACE, C. A. (in press). Platanthera Rich. In C. A. Stace (ed.), Hybridization and the flora of the British Isles (2nd edn). Botanical Society of the British Isles, London. STPICZYNSKA, M. (2003a). Floral longevity and nectar secretion of Platanthera chlorantha (Custer) Rchb. (Orchidaceae). Annals of Botany 92: 191–197. STPICZYNSKA, M. (2003b). Nectar resorbtion in the spur of Platanthera chlorantha Custer (Rchb.) Orchidaceae: structural and microautoradiographic study. Plant Systematics and Evolution 238: 119–126. SUMMERHAYES, V. S. (1951). Wild orchids of Britain (1st edn). Collins, London. TOLLSTEN, L. & BERSTROM, L. (1993). Fragrance chemotypes of Platanthera (Orchidaceae): the result of adaptation to pollinating moths. Nordic Journal of Botany 13: 607–613. TREMBLAY, R., ACKERMAN, J. D., ZIMMERMAN, J. K. & CALVO, R. N. (2005). Variation in sexual reproduction in orchids and its evolutionary consequences: a spasmodic journey to diversification. Biological Journal of the Linnean Society 84: 1–54. VAN DER CINGEL, N. A. (1995). An atlas of orchid pollination: European orchids. Balkema, Rotterdam. WALLACE, L. E. (2003). Molecular evidence for allopolyploid speciation and recurrent origins in (Orchidaceae). International Journal of Plant Sciences 164: 907–916. WALLACE, L. E. (2004). A comparison of genetic variation and structure in the allopolyploid Platanthera huronensis and its diploid progenitors, Platanthera aquilonis and Platanthera dilata (Orchidaceae). Canadian Journal of Botany 82: 244–252. WALLACE, L. E. (2006). Spatial genetic structure and frequency of interspecific hybridization in Platanthera aquilonis and P. dilata (Orchidaceae) occurring in sympatry. American Journal of Botany 93: 1001– 1009. WEBB, D. A. (1980). Platanthera L.C.M. Rich. In T. G. TUTIN, V. H. HEYWOOD, N. A. BURGES, D. M. MOORE, D. H. VALENTINE, S. M. WALTERS & D. A. WEBB (eds.), Flora Europaea 5, pp. 331–332. Cambridge University Press, Cambridge. WHILD, S. & LOCKTON, A. (2007). New records. Shropshire Botanical Society Newsletter (Autumn 2007): 4– 5 + cover. (Accepted November 2007) SPUR LENGTH IN PLATANTHERA 17

·3 ·8 ·5 ·4 ·2 ·3 ·7 ·2 ·5 ·2 ·2 ·6 ·6 ·6 ·5 ·8 CV (%) (%) CV 16 12 11 7 14 9 18 12 5 7 7 NA NA 8 10 10 9 6 11

·79 ·75 ·97 ·07 ·44 ·64 ·39 ·62 ·3 ·64 ·41 ·51 ·57 ·07 ·83 ·78 ·24 SSD 3 2 2 2 2 1 3 4 1 1 1 NA NA 1 3 2 1 1 3

·2 ·98 ·25 ·55 ·8 ·5 ·33 ·2 ·8 ·63 ·38 ·56 ·5 ·2 ·46 Mean Mean 23 22 25 27 17 17 18 36 25 21 19 26 23 18 33 19 19 27 27 3 3 3 N 10 20 20 11 48 5 6 3 5 5 8 1 1 8 9 10 4 4 37 Shaded Shaded N N Y Y N N N Y N N N N N N N N N Y Y mean excluded one plant with exceptionally an 3 Habitat Habitat grassland/ coarse scrub grassland coarse open scrub mixed woodland calcareous flush open rocky slope grassland dense woodland grassland (hay meadow) grassland (hay meadow) grassland coarse grassland coarse grassland coarse wet heath/bog grassland coarse heath heath woodland woodland , LISTED ACCORDING TO INCREASING LATITUDE TO ACCORDING , LISTED not plotted, not as these datasets duplicate other largerdatasets; 2 PLATANTHERA Locality Locality S BARNAVE, S Die, SE of Vercors, SE France Robinia site, S CHAMALAC, SE of Vercors, SE France SE Vercors, of SE Clelles, NW Road, TRESANNE France Croix des Toutes AURES, NE Royans, NW Vercors, France CORVARA, SE Brixen, Dolomites, Italy Main road, OBERGURGL, NW Bozen, W Austria Road W from VENT toward Rofen, NW Bozen, W Austria PETTNEU, 10km E St Anton, E Bludenz, W Austria MOOS, NW outskirts of St Anton, E Bludenz, W Austria OBERLECH, 13 NW Stkm Anton, E Bludenz, W Austria Sylvia’s Meadow, ST ANNE’S Chapel, Tavistock, Cornwall Sylvia’s Meadow, ST ANNE’S Chapel, Tavistock, Cornwall Sylvia’s Meadow, ST ANNE’S Chapel, Tavistock, Cornwall Elkham’s Grave, MEAD END, Rhinefield, S Hants BADBURY Rings, Shapwick,Wimborne, NW Dorset PIG BUSH, Lodge, Denny S Hants STEPHILL Bottom, Denny Lodge, S Hants Sussex Hill, NEWTIMBER Sussex Hill, WOLSTENBURY AND IRISH AND IRISH Recorder(s), Date(s) Date(s) Recorder(s), R Bateman/P Rudall,2007 R Bateman/P Rudall,2007 R Bateman/P Rudall,2007 R Bateman/P Rudall,2007 2006 Sexton, R A Hughes, 2007 A Hughes, 2007 A Hughes, 2007 A Hughes, 2007 A Hughes, 2007 R Bateman/P Rudall,2003 R Bateman/P Rudall,2003 R Bateman/P Rudall,2003 R Bateman/P Rudall,2003 R Bateman/P Rudall,2004 R Bateman/P Rudall,2003 R Bateman/P Rudall,2003 2007 Pearce, Stott/D K 2007 Pearce, Stott/D K 1 1 APPENDIX 1.APPENDIX DETAILS OF LOCALITIES AND SPUR MEASUREMENTS ENCOMPASSED STUDY PRESENT BY THE BRITISH OF not plotted.sampleas size one only orplants; two Species ALPS bifolia bifolia bifolia bifolia bifolia bifolia bifolia indeterminate bifolia bifolia UK bifolia chlorantha hybrids bifolia chlorantha bifolia bifolia chlorantha chlorantha Appendix 1 continued on next page page next on continued 1 Appendix 1 short spur. 18 R. M. BATEMAN AND R. SEXTON

·5 ·6 ·4 ·9 ·1 ·5 ·6 ·7 ·7 ·7 ·3 ·1 ·1 ·6 CV (%) (%) CV 13 11 NA NA 8 6 NA NA NA 14 7 10 6 6 9 8 19 10 7 NA 6 6 NA 7

·6 ·15 ·03 ·55 ·47 ·53 ·95 ·32 ·3 ·58 ·2 ·24 ·42 ·08 ·52 ·46 ·23 SSD 3 3 NA NA 3 2 NA NA NA 4 1 1 1 2 1 3 4 3 2 NA 1 1 NA 2

·58 ·1 ·2 ·7 ·33 ·4 ·75 ·3 ·5 ·75 ·91 ·33 ·25 ·1 ·5 ·9 Mean Mean 26 27 28 28 36 37 36 17 23 31 20 18 19 34 17 36 22 33 29 29 23 24 34 31 N 12 5 1 1 5 20 1 1 1 10 3 5 10 20 10 8 4 11 3 2 5 5 1 10 Shaded Shaded Y Y N N Y Y N N Y N Y N N Y N Y Y Y Y ?Y Y Y Y Y Habitat Habitat woodland woodland grassland coarse grassland open woodland scrub chalk grassland/ coarse scrub grassland coarse woodland grassland/ coarse scrub open woodland grassland grassland woodland grassland woodland/scrub woodland/scrub/ grassland woodland/scrub/ grassland woodland/scrub/ grassland ?woodland woodland scrub scrub/open woodland scrub/coarse grassland ood, Elham, W Kent Locality Locality Sussex Hill, WOLSTENBURY Sussex Hill, CLAYTON Broad Down, WYE Downs, SE Wye, E Kent PARK GATE Down, Elham, E Kent YOCKLETTS Bank, Waltham, E Kent BONSAI BANK, SE Canterbury, E Kent SHEEPLEAS Wood, West Horsley, Surrey Wilts Great Lavington, CHEVERELL Hill, SW Coppice NW SHELDWICH, S Faversham, E Kent Traffic island, LONGREACH Wood, Stockbury, W Kent W Hill STOCKBURY Walkers Hill, PEWSEY Downs,Alton N Barnes, Wilts Walkers Hill, PEWSEY Downs,Alton N Barnes, Wilts CULVERSTONE Woods, Meopham, NW Kent MORGAN’S HILL, Calstove Wellington,Wilts Calne, SE HOMEFIELD Wood, Medmenham, Bucks Warburg Reserve LNR, BIX Bottom, Oxon Warburg Reserve LNR, BIX Bottom, Oxon Warburg Reserve LNR, BIX Bottom, Oxon Near Great MISSENDEN, Bucks BULL’S CROSS, B4070, NE Stroud, Gloucs BULL’S CROSS, B4070, NE Stroud, Gloucs DANCERS END LNR, Aston Clinton, Bucks ASTON CLINTON RagpitsBucks LNR, Recorder(s), Date(s) Date(s) Recorder(s), 2007 Pearce, Stott/D K 2007 Pearce, Stott/D K R Bateman/P Rudall,2004 R Bateman/P Rudall,2004 2003/4 Rudall, Bateman/P R D Johnson, 2007 R Bateman/P Rudall,2003 R Bateman/P Rudall,2004 R Bateman/P Rudall,2004 R Bateman/P Rudall,2004 2003/4 Rudall, Bateman/P R R Bateman/P Rudall,2004 A Hughes, 2007 D Johnson, 2007 R Bateman/P Rudall,2004 2003/4 Rudall, Bateman/P R 2003/4 Rudall, Bateman/P R 2003/4 Rudall, Bateman/P R 2003/4 Rudall, Bateman/P R P Daltry, 2007 A Hughes, 2007 A Hughes, 2007 2004 Bateman, R R Bateman/P Rudall,2004 2 1 1 1 1 1 2 1 1 2 APPENDIX 1 CONT…. APPENDIX Species UK chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha bifolia bifolia chlorantha bifolia bifolia bifolia chlorantha bifolia chlorantha bifolia chlorantha hybrids chlorantha bifolia hybrids? chlorantha chlorantha SPUR LENGTH IN PLATANTHERA 19

·6 ·3 ·3 ·9 ·6 ·7 ·1 ·3 ·2 ·7 ·2 ·7 ·9 ·8 ·8 ·5 ·8 ·8 ·1 ·3 ·5 CV (%) (%) CV 11 13 9 8 NA NA 11 9 9 6 7 13 9 7 8 5 9 12 7 8 9 1 8 8 9 8

·05 ·45 ·69 ·47 ·27 ·73 ·61 ·55 ·07 ·45 ·47 ·15 ·61 ·59 ·85 ·23 ·22 ·3 ·61 ·5 ·23 ·23 ·49 ·16 SSD 3 4 2 2 NA NA 2 2 2 1 2 3 2 2 2 1 2 2 1 2 2 0 2 2 2 2

·2 ·44 ·02 ·75 ·75 ·65 ·23 ·71 ·63 ·85 ·13 ·5 ·83 ·87 ·08 ·04 ·45 ·34 ·73 ·25 ·52 ·83 ·15 ·06 Mean Mean 26 33 29 27 28 29 19 28 28 24 28 25 27 29 29 27 29 17 17 27 26 27 27 26 26 27 3 N 10 9 26 4 1 2 20 20 12 16 17 12 21 20 19 6 67 20 19 6 11 4 29 55 19 24 Shaded Shaded N Y Y Y N N N N N N Y? N Y Y N N Y N N N N N N N N N mean excluded one plant withan exceptionally 3 Habitat Habitat grassland coarse woodland open woodland woodland grassland coarse grassland coarse valley mire grassland grassland grassland ?woodland meadowhay woodland woodland grassland coarse grassland coarse woodland grassland (grassland) grassland (meadow) grassland coarse grassland (meadow) glassland coarse (waste) grassland coarse grassland coarse grassland coarse not plotted, not as these datasets duplicate other largerdatasets; 2 Locality Locality Powys Crickhowell, LLANBEDR, Ty-Commins, Lineage Wood, E Bridge Street,LAVENHAM, W W Suffolk LLYNCHLYS Common, Oswestry, Shrops ARNSIDE Cemetery MEATHROP Quarry SANDSIDE Quarry OUTLY Moss LATTERBARROW LNR, SW Kendal, SE Cumbria LATTERBARROW LNR, SW Kendal, SE Cumbria Bridge NEWBY Hill, Canny LEVENS Wood, SW Kendal, SE Cumbria Piers Gill, DENTDALE A590/A591 junction, SW Kendal, Cumbria SE A590/A591 junction, SW Kendal, Cumbria SE A6/A591 junction, S Kendal, SE Cumbria FIRBANK churchyard SMARDALE NNR, W Kirkby Stephen, SE Cumbria WAITBY Greenriggs, W Kirkby Stephen, SE Cumbria WAITBY Greenriggs, W Kirkby Stephen, SE Cumbria MILGAVIE, East Dumbartonshire CARDROSS, W. Dumbartonshire MURROCH, Dumbarton, E. Dumbartonshire RENTON, East Dumbartonshire BOMAINS Meadow LNR, Bo’ness, Falkirk,?Fife BOMAINS Meadow LNR, Bo’ness, Falkirk,?Fife BOMAINS Meadow LNR, Bo’ness, Falkirk,?Fife Recorder(s), Date(s) Date(s) Recorder(s), R Bateman/H+J Lambert, 2004 R Bateman/J Tyler, 2004 I Wright, 2004 2007 Gendle, A 2007 Gendle, A 2007 Gendle, A 2007 Gendle, A J Fenton, 2007 2007 Gendle, A 2007 Gendle, A J Fenton, 2007 2007 Gendle, A J Fenton, 2007 2007 Gendle, A 2007 Gendle, A 2007 Gendle, A L+N Harbron, 2007 2007 Gendle, A J Fenton, 2007 2007 Longrigg, S 2007 Longrigg, S 2007 Longrigg, S 2007 Longrigg, S 2005 Sexton, R 2006 Sexton, R 2007 Sexton, R 1 1 2 2 2 2 2 not plotted. as sample size one only or two plants; Species UK chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha bifolia chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha bifolia bifolia chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha 1 short spur. Appendix 1 continued on next page page next on continued 1 Appendix 20 R. M. BATEMAN AND R. SEXTON

·7 ·5 ·4 ·5 ·6 ·9 ·4 ·2 ·8 ·8 ·2 ·5 ·6 ·8 ·8 ·8 ·8 ·7 ·5 ·5 ·7 ·1 ·4 ·6 CV (%) (%) CV 8 6 6 8 9 5 8 9 7 4 10 8 7 8 8 7 7 7 12 13 9 10 11 8 12

·51 ·94 ·66 ·3 ·73 ·57 ·47 ·73 ·12 ·38 ·64 ·53 ·07 ·27 ·63 ·17 ·18 ·11 ·36 ·49 ·86 ·89 ·1 ·21 ·45 SSD 2 1 1 2 2 1 1 1 2 1 1 2 2 2 1 2 2 2 2 2 1 2 2 2 3

·9 ·98 ·88 ·94 ·5 ·53 ·41 ·75 ·06 ·71 ·05 ·65 ·36 ·69 ·5 ·78 ·8 ·31 ·86 ·48 ·6 ·09 ·47 ·46 Mean Mean 28 29 25 26 28 26 17 18 27 28 16 29 27 25 18 27 27 27 18 18 20 27 19 26 27 N 20 20 75 28 6 49 22 15 25 6 20 118 24 47 5 8 24 10 36 24 12 27 8 99 6 Shaded Shaded N N N N N N N N N N N N N N N N N N N N N N N N N Habitat Habitat grassland coarse grassland coarse grassland (meadow) grassland (meadow) grassland (meadow) grassland (meadow) wet meadow wet meadow grassland coarse grassland coarse wet meadow grassland coarse grassalnd coarse grassland (meadow) wet meadow grassland (meadow) grassland coarse grassland coarse grassland coarse grassland coarse grassland coarse grassland (meadow) grassland coarse grassland coarse grassland coarse Locality Locality BALLOCH, East Dumbartonshire Stirlingshire Croftamie, PIRNIEHALL, PLEAN Country PARK PLEAN Country PARK PLEAN Country PARK PLEAN Primary SCHOOL Wester BALGAIR, Stirling Wester BALGAIR, Stirling Glinns Road, KIPPEN (B), Stirling Glinns Road, KIPPEN (B), Stirling Bute & Argyll LUSS, Glen KIPPEN (A) Horse Field, Stirling KIPPEN (A) Horse Field, Stirling KINROSS Hills, Keith BALLANGREW Meadow, Stirling ASHFIELD, Dunblane, Perth QUOIGS Meadow, Perth QUOIGS Meadow, Perth QUOIGS Meadow, Perth QUOIGS Meadow, Perth QUOIGS Meadow, Perth FLEECEFAULDS Meadow, Ceres, Fife Stirling Course, Golf CALLANDER W Stirling Callander, Farm, BRAELENY W Stirling Callander, Farm, BRAELENY Recorder(s), Date(s) Date(s) Recorder(s), 2007 Longrigg, S 2007 Longrigg, S 2007 Sexton, R 2005 Sexton, R 2006 Sexton, R 2007 Sexton, R 2007 Sexton, R 2006 Sexton, R 2007 Sexton, R 2006 Sexton, R 2007 Longrigg, S 2006 Sexton, R 2007 Sexton, R 2005 Sexton, R 2006 Sexton, R 2006 Sexton, R 2007 Sexton, R 2006 Sexton, R 2007 Sexton, R 2007 Sexton, R 2006 Sexton, R 2005 Sexton, R 2006 Sexton, R 2007 Sexton, R 2006 Sexton, R 2 2 2 2 2 2 2 2 2 Species UK chlorantha chlorantha chlorantha chlorantha chlorantha chlorantha bifolia bifolia chlorantha chlorantha bifolia chlorantha chlorantha chlorantha bifolia chlorantha chlorantha chlorantha bifolia bifolia bifolia chlorantha bifolia chlorantha chlorantha APPENDIX 1 CONT…. APPENDIX SPUR LENGTH IN PLATANTHERA 21

·5 ·3 ·8 ·6 ·2 ·4 ·6 ·9 ·5 ·9 ·6 ·7 ·29 ·9 ·8 CV (%) (%) CV 10 NA 14 8 NA 13 NA 9 9 14 15 14 16 14 5 12 11 12 15 NA 5

·45 ·51 ·29 ·53 ·88 ·35 ·05 ·69 ·09 ·71 ·42 ·9 ·91 ·8 ·91 ·92 SSD 2 NA 3 2 NA 2 NA 1 1 2 2 2 2 2 0 1 1 1 2 NA 1

·44 ·5 ·32 ·5 ·63 ·6 ·23 ·29 ·42 ·2 ·14 ·31 ·93 ·1 Mean Mean 23 23 24 26 15 18 24 19 14 14 17 14 16 16 16 15 16 15 16 22 17 N 18 1 8 20 2 22 2 20 20 20 7 12 12 20 7 7 13 28 10 1 3 Shaded Shaded N N N N N N N N N N N N N N N N N N N N N mean excluded one plant with exceptionally an 3 Habitat Habitat grassland moorland grassland verge grassland coarse moorland marshy meadow damp meadow grassland coarse calcareous flush/ heath wet heath/bog grassland grassland moorland moorland moorland calcareous flush/ heath grassy bog calcareous flush/ heath calcareous flush/ heath grassland meadow not plotted, not as these datasets duplicate other largerdatasets; 2 Locality Locality Westerness Morvern, LOCHALINE, Roadside, Morvern, Lochaline, ALNE, LOCH Lochside, Westerness Westerness Morvern, ARDTORNISH, Roadside, KELTNEYBURN LNR, Fortingall, Perth Mountainside, Beinn na h-UAMHA, Morvern, Westerness Woodside, FORFAR, Angus Westerness Morvern, DRIMNIN, Inverness Abernethy, Garten, of Boat S GLENCAIRN, Loch a Mhuilinn, S APPLECROSS, W Ross CARNAN, N Benbecula, Outer Hebrides Ross Wester KINLOCHEWE, Roadside, Roadside E TALLADALE, S Loch Maree, Wester Ross Loch na h-ORDNACHE, S Flowerdale, Wester Ross E DUBH LOCH, S Flowerdale, Wester Ross E Kerrysdale, FLOWERDALE, Wester Ross NE Lochdrum Farm, LOCHDROMA, Braemore, W Ross Ross Wester MAREE, Loch NW Kernsary, S N bank Loch KERNSARY, E Poolewe, NE Gairloch, Ross W N bank Loch KERNSARY, E Poolewe, NE Gairloch, Ross W Ross Wester INVEREWE, N Roadside MIDTOWN, NW Poolewe, Wester Ross Recorder(s), Date(s) Date(s) Recorder(s), 2007 Hughes, D+C 2007 Hughes, D+C 2007 Hughes, D+C 2006 Denholm, I 2007 Hughes, D+C 2007 Hughes, D+C 2007 Hughes, D+C 2006 Denholm, I R Bateman/I Denholm, 2007 R Bateman/I Denholm, 2007 2007 Hughes, D+C 2007 Hughes, D+C 2007 Hughes, D+C 2007 Hughes, D+C 2007 Hughes, D+C R Bateman/I Denholm, 2007 2007 Hughes, D+C R Bateman/I Denholm, 2007 2007 Hughes, D+C 2007 Hughes, D+C 2007 Hughes, D+C 1 1 1 1 2 not plotted.sampleas size one only orplants; two Species UK chlorantha chlorantha chlorantha chlorantha bifolia bifolia chlorantha bifolia bifolia bifolia bifolia bifolia bifolia bifolia bifolia bifolia bifolia bifolia bifolia chlorantha bifolia 1 short spur.