Photosynthetic Responses of Plant Communities to Sand Burial on the Machair Dune Systems of the Outer Hebrides, Scotland

Abstract

• Background and Aims The effects of both short-term (2 weeks) and long-term (6 weeks) burial on the photosynthetic efficiency of four typical plant sub-communities of the machair sand dunes of the Outer Hebrides are described. Previous studies have examined the photosynthetic responses on individual species rather than the response at the community level.

• Methods Three replicate turves from four different sub-community types (foredune grassland, dune slack, three-year fallow and unploughed grassland) were subjected to short- and long-term burial treatments after acclimatisation in an unheated greenhouse for approximately 10 weeks. Three replicate control turves from each sub-community were left unburied. After treatment, photosynthetic rate was measured at 16–20 h and 40–44 h after re-exposure, using an infra-red gas analyser, with standardization by total leaf area for each turf. Effects of sub-community type, burial duration and time since re-exposure were analysed by 3-factor split-plot analysis of variance (ANOVA) with repeated measures for time since re-exposure in the subplots.

• Key Results Buried turves were characterized by a low dark respiration rate, which may represent a maintenance response to burial. After removal of sand, each machair sub-community showed some capacity for an elastic photosynthetic response. There were no differences between the effects of short- and long-term burial on the photosynthetic efficiency of machair vegetation, although turves buried for 6 weeks generally attained photosynthetic rates approaching those of control rates sooner than turves buried for 2 weeks. Photosynthetic responses to burial varied between sub-communities, with the slack turves exhibiting the poorest capacity for recovery within the investigated 44-h period.

• Conclusions In the machair environment, the ability to maintain photosynthetic equipment whilst buried, and the ability to bring about a relatively rapid reinstatement of photosynthetic mechanisms on emergence or exposure, is an important adaptation for survival. Survival is closely related to the ability of a plant to replenish carbohydrate reserves before the next burial event.

INTRODUCTION

Research into the effects of burial on plants has focused principally on impacts on individual species, rather than whole communities (Antos and Zobel, 1987; Kent et al., 2001; Owen et al., 2004; Shi et al., 2004). The main aim of the present study was to determine the effects of short- and long-term burial by sand on the photosynthetic efficiency of four typical sub-communities of the machair sand dune systems of the Outer Hebrides of Scotland. Coastal dune species are characterized by varying degrees of adaptation and response to burial (Antos and Zobel, 1987; Kent et al., 2001; Maun, 2004). Apart from the work of Farrow (1919) in the Brecklands of England, Martinez et al. (1997, 2001) on tropical dunes in the Gulf of Mexico and Owen et al. (2004) in the Outer Hebrides of Scotland, research into the effects of burial has concentrated on species-specific responses and tolerances (e.g. Sykes and Wilson, 1990a, b; Brown, 1997; Chen and Maun, 1999; Shi et al., 2004). Individual components of the flora of dune systems, such as bryophytes (Birse et al., 1957; Martinez and Maun, 1999), have also been examined, again using a single-species approach.

Burial and ‘machair stratification’

The Outer Hebrides of Scotland are subject to high wind speeds all year and particularly in winter (Angus, 1997). Sand movement due to anthropogenic activity is also widespread, since the machair dune plains are ploughed and cultivated for cereals and ‘lazy-bedded’ for potatoes (Boyd, 1979). Burial is thus frequent in the Hebridean environment. Gilbertson et al. (1995) and Kent et al. (2001) argued that perennial machair species have been evolutionarily selected because of their ability to tolerate repeated burial. Provided burial is not too deep, plants will recover within weeks or months of inundation. However, if burial is too deep, or there is frequent repeated shallow burial, regrowth fails and the existing vegetation and surface organic horizon are permanently buried, resulting in multiple ‘fossilized’ organic horizons under many areas of dunes, a process called ‘machair stratification’. Through analysis of species' abundances within controlled burial experiments for turves from four Hebridean machair sub-communities, Owen et al. (2004) showed that intermittent and repeated burial was more damaging than a single burial event of greater magnitude. The aim of this paper is to examine photosynthetic responses to the burial process in the same environment.

Burial effects

Modification of the normal micro-environment for dune plants following sand accretion often increases moisture, nutrient availability, bulk density and anaerobic micro-organisms, and decreases soil temperature and aeration, oxygen, light intensity and competition (Maun, 1994, 1998, 2004). Burial also increases the available soil volume for roots to exploit in terms of moisture and nutrients (Maun, 1998). However, possibly the most immediate effects will be due to the changed light environment. If there is no mechanical damage and the plant is completely covered, burial by sand is synonymous with putting the plant in a dark environment (Harris and Davy, 1988; Sykes and Wilson, 1990a, b). Lack of light has clear implications for the immediate photosynthetic activity of a buried plant or plant community: photosynthetic capacity is inhibited and survival of the plant is dependent on the use of stored carbohydrate reserves (Harris and Davy, 1988). However, burial may also have ramifications for the long-term photosynthetic competence of a plant, which, in turn, will have important implications for the abilities of a plant to re-establish itself successfully following exposure or emergence from the sand.

The general physiology of plant responses to sand accretion has not been well studied, and specific considerations of the effects of burial by sand on the photosynthetic mechanisms of dune vegetation are limited to those of Harris and Davy (1988) and Yuan et al. (1993). Their investigations described the burial responses of individual plants of a single species when isolated from the plant community as a whole. However, the individual members of a plant community typically have complex inter- and intra-specific relationships, including competition for light, nutrients and space. The response of an individual removed from the effects of competition and other interactions is likely to differ from that of an individual of the same species examined as part of a community. Additionally, burial by sand in the dune environment is not limited to individuals of a single species, but is a phenomenon that affects whole communities and sub-communities. Although community-level measurement of plant physiological responses to environmental factors is a more ecologically sound approach, and recognizes that a plant community is greater than the sum of its constituent species, there is no literature pertaining to photosynthetic responses to sand inundation at the level of the community. Such a community-based perspective is the basis of this paper.

Photosynthetic responses to environmental stresses are typically ‘elastic’ (Levitt, 1972). Elastic responses to stress can be defined as those changes in a plant's functioning that return to the optimal level once the stress has been removed (Levitt, 1972; Salisbury and Ross, 1992). In a dynamic and unpredictable environment, such as the machair, where sand is highly mobile and the risk of burial is great, the vegetation may therefore be expected to maintain its photosynthetic capacities even when buried, permitting a rapid elastic response in the event of exposure by wind/rabbit action, or of emergence through the deposit.

MATERIALS AND METHODS

Site selection

Angus (1996, 1997), English Nature (1999) and Kent et al. (2003) state that, botanically, the South Uist machairs represent the best examples of Hebridean machair systems. Two typical but differing examples of machair vegetation that were subject to cultivation at Drimsdale and Kildonan (Fig. 1) were selected.

Fig. 1.

The location of the two sampling sites at Drimsdale and Kildonan in the Outer Hebrides of Scotland.

The Drimsdale research area lies within the Loch Druidibeg National Nature Reserve. Owen et al. (1996, 2000) have described the plant communities and sub-communities in detail. In the crofting township formed in 1924, cereal crops are cultivated on a rotational basis, while production of potatoes occurs in an area of lazy beds (linear earth mounds) towards the north end of the site. Hay fields are situated in the wetter areas. At Kildonan, domestic sheep and some cattle are grazed and the area has a large population of feral rabbits. Both plough and spade cultivation are extensive, with the production of both cereals and potatoes. The vegetation of the whole machair system at Kildonan is described in detail in Owen et al. (1998, 2000). Fourteen plant communities and sub-communities were defined using Two-way Indicator Species Analysis (TWINSPAN) (Hill, 1979), and their similarities to those detailed more widely for the machair systems of South Uist, Barra and Vatersay by Pankhurst (1991), Kent et al. (1996) and Dargie (1998) were noted. At both sites, the township gates are opened during the winter months to allow common grazing.

Collection and maintenance of plant material

Turf transplant experiments were used, for reasons of practicality and to attempt to standardize conditions. During mid-April 1997, turves of machair vegetation, each approximately 15 × 15 × 10 cm, were collected from areas of foredune (first main dune ridge, not embryo dunes) grassland (FD) and dune slack (SL) at Kildonan, and 3-year fallow grassland (3F) and unploughed grassland (UP) at Drimsdale (Fig. 1). The 3F and UP grasslands were typical of the main machair plain. In this respect, the last three sub-communities and their constituent species were quite different from those typically examined elsewhere, where studies have concentrated predominantly on foredune and main dune ridge species (e.g. the various researches of Maun and colleagues on dune systems around the Great Lakes of Canada).

Nine turves were collected from each of the four sub-community types, giving a total of 36 turves. Every effort was made to ensure that replicates within each sub-community were as similar as possible in species composition, although clearly some variability existed in relation to local mosaics in the vegetation. The initial species composition of the turves within each of the four sub-communities immediately prior to experimental burial is shown in Table 1. Ideally, a larger number of replicates would have been taken but the logistics of transporting the turves and maintaining them precluded this.

Table 1.

Initial species composition of the turves from the four machair sub-communities, foredune grassland (FD), dune slack (SL), three-year fallow grassland (3F) and unploughed grassland (UP)

Sub-community	FD	SL	3F	UP
Achillea millefolium	4·1		3·2	4·1
Agrostis stolonifera	12·6	43·2		6·5
Bellis perennis	1·5		24·5	16·7
Calliergon cuspidatum		45·4
Campylium chrysophyllum			0·7
Cardamine pratensis		9·5
Carex arenaria		0·9		4·3
Carex flacca		2·7
Carex nigra		8·1
Cerastium fontanum	3·5	2·1		3·25
Epilobium palustre		0·5
Festuca rubra	63·7	6·8	65·1	82·0
Galium palustre		16·7
Galium verum	5·2		0·6	3·0
Geranium molle			0·9	4·6
Holcus lanatus	0·8	5·6	0·6	17·1
Homalothecium lutescens			6·8
Hydrocotyle vulgaris		3·4
Juncus articulatus		1·8
Juncus bufonius		7·3
Lophocolea bidentata	6·1		0·8
Lotus corniculatus			15·0	4·7
Lychnis flos-cuculi		2·3
Plagiomnium rostratum	20·4	3·0	3·5	19·3
Plantago lanceolata	8·8			10·8
Poa pratensis	7·5	10·9	0·5
Poa subcaerulea			11·2	0·9
Potentilla anserina		1·7	0·7
Prunella vulgaris		4·7	2·7	6·2
Ranunculus acris			5·5	3·5
Ranunculus repens	6·7	2·8	4·8	4·4
Rhytidiadelphus squarrosus		0·7		9·2
Sagina procumbens		4·9
Samolus valerandi		0·6
Senecio jacobaea			1·5	2·3
Taraxacum brachyglossum	1·9			1·1
Thalictrum minus	1·5
Tortula ruraliformis			9·5	1·0
Trifolium pratense	10·8
Trifolium repens	10·4	12·9	20·5	20·6

Data recorded as local rooted frequency (Greig-Smith, 1983) within a 10 × 10 grid of 1 cm squares in the centre of each 15 × 15 cm turf. Scores are the means of the nine replicate turves from each sub-community type. Nomenclature: higher plants, Pankhurst (1991), Stace (1997); bryophytes, Watson (1981).

The turves were transported to the University of Plymouth in large plastic crates and were immediately transferred to an unheated polythene tunnel under natural light. The turves were watered daily by an aerial sprinkler system and were acclimatised for approximately 10 weeks prior to experimentation. Since machair vegetation is subject to frequent heavy grazing by rabbits, cattle and sheep, each turf was artificially ‘grazed’, using shears, to a height of approx. 3 cm at 2-week intervals during this period.

Experimental burial of machair turves

During the final week of June 1997, at the end of the 10-week acclimation period, the nine turves from each of the four sub-community types were assigned to one of three different treatments. Sand used for burial during the course of the investigation was collected from exposed areas at Daliburgh (Landranger map reference NF745210) and Kildonan machairs and sterilized in a Camplex electric soil sterilizer (Simplex Instruments, Cambridge, UK) to destroy any viable propagules. Wooley and Stoller (1978) found that burial by 2–3 mm of coarse sand attenuated light to 1 % of the surface flux. A deposit of 2 cm of sand was, therefore, sufficient to inhibit all photosynthetic activity within the experimental turves. For the long-burial (LB) treatment, three turves from each sub-community type were buried under a depth of 2 cm of sand during the last week of June 1997, approximately 6 weeks prior to their photosynthetic measurement. A further three turves of each type were subjected to the short-burial (SB) treatment and were buried under 2 cm of sand in the final week of July 1997, only 2 weeks prior to photosynthetic measurement. The final three turves from each sub-community served as controls and remained unburied throughout the course of the investigation. A re-survey of the species composition of the controls showed no significant change in species composition at the end of the experiment.

Measurement of exchange rate

All photosynthetic investigations were undertaken at the Institute of Biological Sciences, University of Wales Aberystwyth, during the period 4 to 15 August 1997, using an ADC 225 Plant Physiology Infra-Red Gas Analyser (IRGA) Mk 3.

Rates of net carbon dioxide exchange were taken for all twelve control turves to provide a baseline photosynthetic rate for each investigated machair sub-community and to enable comparisons between unburied vegetation and vegetation recovering following burial. For completeness, and in order to confirm that no photosynthesis occurred during burial by sand, the net carbon dioxide exchange of all twelve LB treatment turves, complete with 2 cm layer of sand, was also measured.

The 2 cm layer of sand was removed from the twelve LB treatment turves on 5 August 1997, approximately 6 weeks after burial. The net carbon dioxide exchange of each LB turf was measured at 16–20 h following re-exposure and subsequently at 40–44 h after re-exposure. To prevent excess disturbance during the 20–40 h period, the exposed turves were stored within the IRGA laboratory. Vegetation turves subjected to the SB treatment were uncovered on 10 August 1997, approximately 2 weeks following sand inundation. Net carbon dioxide exchange was measured at 16–20 h, and again at 40–44 h, following re-exposure.

The configuration of the IRGA and the method of gas exchange measurement were identical for each experimental run. Due to the dimensions of the light chamber, it was necessary, prior to the use of the IRGA, to trim each turf to 13 × 13 × 10 cm to ensure a good fit. The turf was carefully lowered into the plant chamber to stand on a grid above a circulating fan. The temperature of the plant chamber was controlled by means of a water jacket supplied by a Grant Instruments (Cambridge, UK) SB3 water cooler equipped with a refrigerated coil and a thermostatically controlled heater/stirrer unit. The addition of ice to the water cooler brought the ambient temperature of the chamber to a level similar to that in the field (approx. 12 °C). The light source was a 1000 W sodium lamp, operating at a photon flux density of approx. 700 µmol m⁻² s⁻¹. The rates of flow through the analysis and reference cells were kept approximately equal at all times.

Leaf area determination

Photosynthetic rate is most usefully expressed as µg CO₂ m⁻² s⁻¹. Although each experimental turf conformed to a uniform surface area of 13 × 13 cm and it would, therefore, have been valid to express photosynthetic rate per unit area of turf; turves from different sub-communities were characterized by different species composition and different vegetation densities. To determine the photosynthetic rate of each experimental turf, it was, therefore, necessary to obtain a measure of total leaf area for each turf. Following the experiment, the vegetation was trimmed to soil level using scissors and the total leaf area of each turf was determined using a DT area measurement system (Delta-T Devices Ltd, Cambridge, UK) connected to a JVC TK-5310 video camera. Dead, senescing and diseased plant material was disregarded, as the contribution to total turf photosynthesis was considered to be negligible.

Where no net photosynthesis occurred, the amount of carbon dioxide discharged was also expressed as µg CO₂ m⁻² s⁻¹. Although this is not a true measure of the respiratory rate of a turf, as such rates are generally expressed as a function of plant dry mass, the time allocated for IRGA investigations at Aberystwyth was not sufficient to record the total plant biomass on each experimental turf. However, expressing the amount of carbon dioxide discharged as a function of leaf area permits a comparison of these values with those calculated for photosynthesis and provides a measure of the flux of carbon dioxide across the leaf area of each turf.

Statistical analysis

Firstly, two-way Analysis of Variance (ANOVA) was used to compare net photosynthesis in relation to community type and burial (2 cm: all 12 LB turves)/non-burial (control: 12 turves). Secondly, the effects of sub-community type, duration of burial and time since re-exposure on net carbon dioxide exchange rates were examined by a three-way (community ×4; burial time ×2; time since re-exposure ×2) split-plot ANOVA with the last factor in the subplots (Underwood, 1997). The analyses were performed using the Statistical Package for the Social Sciences (SPSS Version 11.5).

RESULTS

Net carbon dioxide exchange rates for all sub-community and treatment combinations are presented in Table 2. Negative values indicate a net uptake of carbon dioxide and hence represent net photosynthetic activity. Turves exhibiting positive values represent net respiration.

Table 2.

Net carbon dioxide exchange rates of four machair sub-communities in response to different burial regimes

		Carbon dioxide exchange rate (µg CO2 m−2 s−1)
Sub-community	Replicate	Control	B, buried	LB, 16–20 h	LB, 40–44 h	SB, 16–20 h	SB, 40–44 h
FD	1	−1·58	0·14	0·65	0·72	−0·32	0·03
	2	−2·37	0·14	−0·22	0·20	−0·36	−0·34
	3	−2·4	0·14	−0·03	−0·58	0·09	0·02
SL	1	−1·75	0·29	0·00	−0·78	0·10	0·28
	2	−1·99	0·38	0·09	0·20	0·02	0·02
	3	−2·01	0·35	0·00	0·71	0·08	0·19
3F	1	−0·79	0·17	−0·51	0·09	−0·18	−0·66
	2	−2·16	0·18	−0·10	0·53	0·21	−0·11
	3	−2·31	0·24	−0·70	0·30	−0·39	−0·17
UP	1	−1·65	0·07	−0·15	−0·15	−0·61	−0·36
	2	−1·94	0·04	−0·65	−0·81	−0·12	−0·09
	3	−1·61	0·07	0·53	−0·78	−0·57	–0·53

FD = main foredune grassland; SL = dune slack; 3F = 3-year fallow grassland; UP = unploughed grassland; B = all 12 LB turves once buried; LB = buried for 6 weeks before re-exposure; SB = buried for 2 weeks before re-exposure; 16–20 h = 16–20 h after re-exposure; 40–44 h = 40–44 h after re-exposure.

Within each sub-community, the greatest net photosynthetic activity (as indicated by a net uptake of carbon dioxide and a negative exchange rate value) was in the control turves (Table 2). All unburied control turves (C) had significantly greater net photosynthetic rates than the buried LB turves (B) (Table 2) and the two-way ANOVA indicated that all control turves showed a significantly higher net exchange of carbon dioxide than all other LB turves buried by 2 cm of sand (Table 3). However, rates of carbon dioxide exchange of control and buried turves were not significantly different between sub-communities (Table 3).

Table 3.

Two-way ANOVA for the effects of sub-community type (C) and burial (2 cm sand: all 12 LB turves)/control (non-burial: all 12 control turves) (B) on the net photosynthetic rate of machair vegetation

	d.f.	F	P
Sub-community (C)	3	0·463	0·712
Burial/control (B)	1	209·747	1·29 × 10−10
C×B	3	0·658	0·589
Residual	16	0·122

There was considerable variation in the data, which can be partly attributed to the fact that IRGA measurements were carried out on only three replicate turves for each treatment. Additionally, turves within the same sub-community subjected to the same treatment did not always show uniform responses in terms of net carbon dioxide exchange (Table 2). In some cases replicate turves showed both positive and negative carbon dioxide exchange.

Recovery rates after re-exposure at both 16–20 h and 40–44 h for all four sub-community types and both burial treatments were low in relation to their respective controls (Table 2). In the foredune (FD) grassland, there was evidence of poor net photosynthetic recovery on re-exposure after the long-burial treatment at both 16–20 h and 40–44 h, but net photosynthesis did recommence under the short burial regime at both exposure times. The slack (SL) turves showed the least response, with only one of the twelve buried and re-exposed samples (SB-40–44 h) registering actual net photosynthesis (Table 2). With the exception of LB at 40–44 h, carbon dioxide exchange rates of slack turves of all combinations of burial duration and time since re-exposure were significantly lower than control rates.

The 3F turves exhibited the most variable response and the lack of photosynthetic response in the LB 40–44 h turves was surprising in view of the apparent recovery of the LB 16–20 h turves (Table 2). However, rates of carbon dioxide exchange of 3F turves for all combinations of burial duration and time since re-exposure were not significantly different from control exchange rates (Table 2). Unploughed grassland (UP) gave the best response in three of the four burial/re-exposure time treatments (SB-16–20 h and 40–44 h, and LB-40–44 h) and was second only to 3-year fallow (3F) in the LB-16–20 h treatment (Table 2). Carbon dioxide exchange rates of FD and UP grassland turves previously buried for 2 weeks (SB treatment) were significantly different from control rates at both 16–20 h and 40–44 h following re-exposure (Table 2).

In the three-way ANOVA, the period of time since re-exposure had no significant effect on the net carbon dioxide exchange of turves previously buried for either 2 (SB treatment) or 6 weeks (LB treatment) (Tables 2, 4a). Rates of net carbon dioxide exchange among sub-communities were also not significantly different (Table 4a), although the dune slack community (SL) showed virtually no photosynthetic recovery, since all measurements except one were positive (Table 2).

Table 4.

(a) Three-way ANOVA for the effects of sub-community type (C), burial time (B) and post-burial net photosynthesis (P), with repeated measures for time after re-exposure. (b) Table of means for the three-way interaction. 16–20 h and 40–44 h indicate the post-burial time of photosynthetic measurement

(a)
	d.f.	Mean square	F	P
Main plot
Plant sub-community (C)	3	0·4256	2·28	0·118
Burial duration (B)	1	0·1131	0·61	0·448
C × B	3	0·0664	0·36	0·786
Residual	16	0·1866	–	–
Sub-plot
Post-burial gas exchange (P)	1	0·0239	0·24	0·629
Interactions
C × P	3	0·1090	1·11	0·374
B × P	1	0·0035	0·04	0·853
C × B × P	3	0·3125	3·18	0·052
Residual	16	0·0982	–	–
Total	47	–	–	–

(b)
	Community
	FD			SL			3F			UP
	16–20 h	40–44 h	16–20 h	40–44 h	16–20 h	40–44 h	16–20 h	40–44 h	Burial time Means
Short burial (2 weeks)
Replicates	−0·197	−0·097	0·067	0·163	−0·120	−0·313	−0·433	−0·327	−0·157
Mean (s.e.)	−0·147 (−0·100)			0·115 (−0·096)			−0·217 (0·193)			−0·380 (−0·106)
Long burial (6 weeks)
Replicates	0·133	0·113	0·030	0·043	−0·436	0·307	−0·090	−0·580	−0·060
Mean (s.e.)	0·123 (0·020)			0·037 (−0·013)			−0·065 (−0·743)			−0·335 (0·490)
Community means	−0·012			0·076			−0·141			−0·358

Duration of burial by sand (LB–SB) prior to re-exposure had no significant effect on the net carbon dioxide exchange rate of turves uncovered for 16–20 h or for turves uncovered for 40–44 h (Tables 2, 4a). Net exchange rates were not significantly different between sub-communities at either 16–20 h following re-exposure or at 40–44 h following re-exposure (Table 4a).

The most interesting result of the three-way ANOVA was the existence of marginally significant interaction effects between sub-community type, length of burial and time since re-exposure (Table 4a). Perhaps the most surprising feature is that the later photosynthetic measurement often shows a lower value than the earlier one (five of the differences are negative and three are positive; Table 4b). However, there is no consistency to these differences, which makes the interpretation of the triple interaction difficult. Undoubtedly, the fact that in long burial there are two very different changes in the time of photosynthetic measurement factor (−0·743 for the 3F community, but +0·490 for the UP community), whereas the differences in short burial are relatively more uniform, is the cause of this marginally significant triple interaction.

DISCUSSION

The net photosynthetic capacity of all machair sub-communities was completely inhibited by a 2 cm layer of sand, confirming that no photosynthesis is possible during burial (Table 2: LB buried). Harris and Davy (1988) found that the net photosynthesis of Elytrigia juncea after 2 d of complete burial was 30 % lower than unburied controls. Net photosynthesis was reduced to nearly zero by 5 d of complete burial (Harris and Davy, 1988). On this basis, it is probably correct to assume that the 6-week buried machair turves had shown no active photosynthesis for a period of 4–5 weeks and the positive carbon dioxide exchange rates recorded for each sub-community (Table 2) can, therefore, be taken as a true measure of dark respiration. Reduced dark respiration rates are an adaptation of severely shaded plants, particularly as a survival mechanism during periods of temporary stress (Fitter and Hay, 1987; Salisbury and Ross, 1992). The low dark respiration rates recorded for the machair sub-communities may, therefore, represent a maintenance response to burial. However the possibility of oxygen stress due to an alteration of the micro-environment (Maun, 1994, 1998) should not be ignored.

Within sub-communities, different burial durations and periods since re-exposure had similar effects on photosynthetic rates; the only significant differences recorded being those between the photosynthetic rates of control and all buried (LB) turves (Table 3). Control turves from all four sub-communities exhibited comparable photosynthetic rates. However, the effects of burial (in terms of the ability of a sub-community to recover following inundation) were variable between sub-communities. However, there was no evidence that, collectively, species in any of the buried machair sub-communities were photosynthesising at a higher rate than unburied ones, as described for the single species Ulmus pumila in controlled burial experiments by Shi et al. (2004).

Relative to other sub-communities, slack (SL) turves showed poor recovery from both burial regimes. Slack turves previously subjected to burial for 6 weeks (LB) showed no net photosynthetic activity at 16–20 h following re-exposure. However, by 40–44 h after re-exposure, photosynthetic response was highly variable with two turves showing no response but the third approaching the photosynthetic rate of the controls. However, for those turves subjected to burial for 2 weeks (SB), carbon dioxide uptake was still significantly lower than that of control turves even at 40–44 h after re-exposure and no net photosynthesis was recorded at either stage of re-exposure. The poor photosynthetic response of the slack turves relative to turves from other sub-communities must have been due to the inherent species composition of the slack sub-community. Owen et al. (1998, 2001, 2004) found that slacks contain some species with characteristically poor burial tolerances, for example Juncus articulatus, Lychnis flos-cuculi and Holcus lanatus (see also Table 1). It is likely, therefore, that burial resulted in a greater loss of biomass from slack turves than from turves of other sub-communities. This loss of biomass, specifically in the form of dead and senescing leaf matter, would clearly reduce the capacity of the slack turves for efficient photosynthesis, ultimately leading to a poor photosynthetic response.

Machair dune slacks are routinely flooded during the winter months (Gimingham, 1964): thus, whereas waterlogging is a temporally predictable stress in the slack environment, burial is less so (Maun, 1998, 2004). The relatively poor photosynthetic recovery of the slack turves following burial may, therefore, be additionally due to the fact that the mechanisms of photosynthesis in slack species are primarily adapted to the waterlogging that inevitably occurs during winter. It is possible that, in some slack species, physiological and biochemical adaptation of the photosynthetic apparatus to waterlogging has precluded the development of specialized adaptations to the effects of burial by sand (Maun, 1998, 2004).

FD and SL turves, subjected to either burial treatment, were unable to achieve photosynthetic rates comparable to the corresponding controls, even at 40–44 h after re-exposure (Table 2). The 3-year fallow grassland (3F) sub-community made some photosynthetic recovery following burial, although both SB and LB turves were well below those of the control turves. The apparent recovery of the 3F sub-community following burial should be viewed with caution, as it is possible that the variable nature of the data for the control turves has partly masked the true photosynthetic responses of the experimental turves. Also the apparent recovery of the 3F turves under the LB treatment at 16–20 h is contradicted by the absence of recovery at 40–44 h.

Overall, where differences existed within a sub-community between the control and LB photosynthetic rates and between the control and SB photosynthetic rates (i.e. FD, UP and SL sub-communities), turves subjected to the SB treatment took longer than LB turves to achieve photosynthetic rates approaching that of controls, although differences in recovery time between SB and LB treatments were not significant. This observation may appear surprising in light of the knowledge that the time taken for a plant to reactivate photosynthesis by repair mechanisms, following removal of a stress factor, is generally in proportion to the duration of stress (Larcher, 1995). However, the more rapid recovery of LB turves, in comparison with SB turves, may be because the former were exposed to a longer period of burial stress, which would have resulted in a greater utilization of stored carbohydrate reserves (Maun, 1998). The relationship between photosynthetic activity at sites of carbohydrate production (‘sources’, notably photosynthesising leaves) and the demand for photosynthate at sites of carbohydrate consumption (‘sinks’ such as growing apices, meristems and leaves) is well documented (e.g. Canny, 1984; Salisbury and Ross, 1992). Plant sources and sinks are inter-related, so that an increase in ‘sink strength’ (i.e. demand for photosynthate) leads to an associated increase in ‘source strength’ (i.e. production of photosynthate). It is possible, therefore, that the depletion of stored carbohydrate in the LB turves during burial resulted in an increase in sink strength which, on re-exposure of the photosynthetic tissue to light on re-exposure, led to a rapid re-instatement of source activity. However, further research is clearly needed to validate these ideas.

Although, as described above, photosynthetic rates of uncovered SB and LB turves were never comparable to controls, all sub-communities, with the major exception of the slack, exhibited some degree of positive photosynthesis following re-exposure. Clearly, the photosynthetic functioning of machair vegetation is capable of an elastic response to burial. Individuals of E. juncea exhibited a similar recovery of photosynthetic competence after complete burial with sand (Harris and Davy, 1988). Previously buried individuals showed a net photosynthetic rate similar to that of unburied controls at 24 h following re-exposure (Harris and Davy, 1988). The dune grasses Ammophila breviligulata and Calamovilfa longifolia exhibited an even more efficient photosynthetic response to burial by sand. Seedling and adult individuals showed a higher net carbon dioxide uptake after emergence from sand compared with unburied controls (Yuan et al., 1993).

The cessation of photosynthesis during burial, followed by a resumption of activity on exposure, indicates that although there may have been partial physical deterioration of the photosynthetic apparatus of machair vegetation as a result of burial, most species maintained their photosynthetic capability. It is more likely that burial resulted in the reduction of levels of inducible photosynthetic enzymes (e.g. ATPases) within the chloroplasts. Re-exposure to light on removal of the sand layer would, therefore, have stimulated the production of new enzymes and the subsequent re-activation of photosynthetic mechanisms. Morphological changes in response to burial are well documented and can have pronounced effects on a plant's physiological functioning (Maun, 1998). Changes in leaf morphology in response to burial, for example, have obvious implications for photosynthetic efficiency. Disraeli (1984) and Yuan et al. (1993) reported an increase in leaf width, and hence a larger leaf area for the interception of radiation, in buried dune grasses. An increase in leaf thickness, resulting in an increased number of photosynthetic bundle sheath cells and more efficient absorption of light energy, has also been recorded in buried grasses (Yuan et al., 1993). Additionally, Disraeli (1984) found that the concentration of chlorophyll in leaves of A. breviligulata increased as a result of burial with sand. Although these factors were not considered during this investigation, the possible role of morphological changes in aiding the plants' photosynthetic recovery following re-exposure, should not be discounted.

Various aspects of the overall experimental design may be commented on. Ideally, the number of sample turves from each sub-community would have been larger, and it would have been interesting to compare other treatments, such as the effects of repeated intermittent burial, as well as single burial at varying depths (Owen et al., 2004). However, given their size and weight, the handling of the larger numbers of turves would rapidly have become very difficult. The results clearly demonstrate, however, that such turf transplant experiments can be used to assess photosynthetic efficiency at the community level.

CONCLUSIONS

Determination of photosynthetic efficiency following experimental burial of turves from four machair sub-communities was possible and demonstrated that suspension of photosynthetic activity is a physiological adaptation of many machair species. The measurement of photosynthesis for species assemblages on turves, rather than for individual species in controlled environments, is a novel aspect of the research. Nevertheless, more detailed investigation on the post-burial characteristics of certain dominant machair species, such as Festuca rubra, Ranunculus repens and Prunella vulgaris, would also allow estimation of the relative contributions of key machair species to the total community photosynthesis.