Phosphorus response and efficiency of four adventive annual clovers grown in a New Zealand high country soil under glasshouse conditions

Abstract

Maintaining and increasing legume abundance is a critical component of pastoral intensification, increasing nitrogen inputs to nitrogen deficient New Zealand high and hill country pastures and improving feed quality. Establishment and persistence of traditionally sown legume species white clover (Trifolium repens) and subterranean clover (T. subterraneum) is often limited in summer-dry high country. In contrast, naturalized, adventive annual pasture legume species such as cluster clover (T. glomeratum), haresfoot clover (T. arvense), striated clover (T. striatum) and suckling clover (T. dubium) persist on low fertility, summer-dry high and hill country slopes, although little is known about the edaphic requirements of these species. A glasshouse study was conducted to determine the response and efficiency of these four pasture legume species to increasing levels of available phosphorus (P) in a typical, low fertility, acidic New Zealand high country soil, comparing against white clover and subterranean clover as ‘reference’ species. Trifolium subterraneum was the most productive species (4.4 g dry matter [DM]/pot) and T. striatum yielded the least (0.8 g DM/pot). The order of greatest yield DM response was T. subterraneum > T. arvense > T. repens > T. dubium ≥ T. glomeratum > T. striatum, while the P application rates at which maximum yield occurred varied between species. Mean shoot P uptake was highest for T. subterraneum and lowest for T. striatum (24.1 and 3.8 mg P/pot, respectively). P-response efficiency by species was in the order of T. subterraneum > T. arvense > T. glomeratum > T. repens > T. dubium > T. striatum. Implications for low input, extensive grazing systems in high and hill country are discussed. Trifolium arvense and T. glomeratum show potential for further P response investigation under field conditions.

Keywords:

• annual clovers
• high country
• P response
• P uptake
• P efficiency

Introduction

Phosphorus (P) is second only to nitrogen (N) as the key macro nutrient driving the productivity of legume-based grazed pasture systems in New Zealand high and hill country (Haynes & Williams 1993). Often the sole N inputs to these systems are sourced from sward legumes (clovers) through biological N fixation. This N input, as the key driver of overall pasture yield, is strongly influenced by soil P fertility (Moir et al. 2000). In general, higher soil P fertility often allows for a larger proportion of pasture legume to be present and to persist in the sward, resulting in greater annual biological soil N inputs (Bowatte et al. 2006; Gillingham et al. 2008) and increased soil plant available N for sward growth (Moir et al. 1995, 1997). As such, highly productive legume species such as white clover, Trifolium repens L., (and pastures) are generally adapted to high fertility soil conditions and do not perform well in infertile and/or acid soil conditions (Haynes & Williams 1993). ‘Improved’ pastures therefore require annual maintenance fertilizer inputs, such as single superphosphate (SSP) to elevate soil P and sulphur (S) fertility and sustain high production. In contrast, annual fertilizer inputs to these farming systems are, in reality, low, driven by the economics of fertilizer inputs. As a result, many high and hill country soils in New Zealand have soil fertility levels far below optimum for many common pasture legume species, such as white clover. As such, the long-term sustainability of extensive agroecosystems may therefore depend on greater P efficiency, involving sustained production with reduced fertilizer P inputs (Gilbert 2009; Vaccari 2009).

An important component of pastoral intensification is to increase legume abundance, so as to provide increased feed and nitrogen inputs to N deficient, high country grassland. However, the establishment and persistence of sown legume species such as the perennial white clover (T. repens L.) and the annual subterranean clover (T. subterraneum L.) is often limited in summer-dry, low soil fertility areas (Knowles et al. 2003; Power et al. 2006). This is in contrast to the common presence of other naturalized, adventive and unsown annual legumes, such as cluster clover (T. glomeratum L.), haresfoot trefoil (T. arvense L.), striated clover (T. striatum L.) and suckling clover (T. dubium) (Boswell et al. 2003; Power et al. 2006) that may be more suited to the low soil fertility and microclimates that exist on these hill slopes (Boswell et al. 2003; Maxwell et al. 2010). Much research has been conducted investigating white clover (T. repens) dry matter (DM) and growth responses to added P in New Zealand (Caradus & Snaydon 1986; Caradus et al. 1995). Considerable research has also been conducted in Australia investigating subterranean clover (T. subterraneum) responses to added P (Paynter 1993; Bolland & Paynter 1994; Bolland 1995; Cayley & Hannah 1995; Cayley et al. 1998). The literature addressing P responses of T. subterraneum and naturalized adventive annual legumes under New Zealand conditions is limited (Beale et al. 1993; Dodd & Orr 1995; Maxwell et al. 2010, 2012). In a pot trial using two P rates, Caradus (1980) reported a P response by striated, haresfoot and suckling clovers. Likewise, Hart & Jessop (1984) reported that suckling clover responded to P, although the response was low.

How these naturalized adventive legumes respond to increases in the plant-available soil P pool is therefore a crucial question. Further, establishing a ‘critical’ level of available soil P and plant P efficiency for these legume species is vital for extensive farming systems (Syers et al. 2008, 2010).

The objective of this study was to determine the influence of P supply on the DM yield and P response efficiency of six legume species grown in a typical low P fertility New Zealand high country soil under glasshouse conditions. This study is part of a larger suite of field and climate-controlled experiments examining the role of adventive annual clovers in extensively grazed high and hill country environments (Maxwell et al. 2010, 2012).

Materials and methods

Soil and growth conditions

The soil used was a ‘Cass Series’ High Country Southern Brown soil (New Zealand classification: Upland Allophanic Brown Soil—Hewitt [1998]. USDA: Dystrochrept—Soil Survey Staff [1998]). Soil (0–7.5 cm horizon) was collected from a high country site (43°19′02.88″S, 171°8′31.41″E) on Glenfalloch Station, a central Canterbury sheep and beef farm in the South Island, New Zealand in June 2008. The hill site (25° slope) has an altitude of 740 masl and a north-facing aspect. At most, a total of one or two fertilizer applications of SSP have been applied at this site since development (100 years; 30 kg P, 40 kg S ha⁻¹). The field moist soil was prepared by passing through a 4 mm sieve, removing all plant material, then mixing thoroughly. Soil analyses were conducted before commencement of the experiment, with results confirming an initial low soil fertility status (Table 1).

Table 1 . Initial fertility status of a high country Allophanic Brown soil used in the glasshouse experiment. Topsoil sample (0–75 mm) was collected from a mid-altitude (740 m) central South Island, New Zealand high country site in June 2008.

Soil properties
pH	5.2
Olsen P	11 mg kg^−1
Sulphate S	4 mg kg^−1
CEC	20 cmolc kg^−1
Base Saturation	39.9%
Exchangeable Ca	5.7 cmolc kg^−1
Exchangeable Mg	1.59 cmolc kg^−1
Exchangeable K	0.53 cmolc kg^−1
Exchangeable Na	0.11 cmolc kg^−1
Total N	0.63% w/w
Total C	9.62% w/w
C/N Ratio	15
Mineralisable N	288 kg ha^−1
P Retention (ASC)	44%

The glasshouse experiment was conducted at the Lincoln University glasshouse facilities, Canterbury, New Zealand. Phosphorus, in the form of Ca(H₂PO₄).4H₂0, was applied at eight different rates; 0, 30, 60, 100, 250, 500, 1000 and 2500 mg P kg⁻¹ soil. Treatments were replicated four times for each of the six pasture legume species. Basal nutrient addition prior to experiment commencement included sulphur (S) as calcium sulphate (CaSO₄) to address the soil plant-available S deficiency, and potassium (K) as potassium chloride (KCl), applied at rates of 500 mg element kg⁻¹ soil. All pots also received lime at a rate of 4 t ha⁻¹ equivalent of laboratory grade CaCO₃, which corrected soil pH to 6.1. The various P treatments and basal nutrients were added to 400 g of the air-dried soil and then mixed thoroughly. The soil was then lightly packed into 0.8 L (10 cm height×12 cm diameter) pots with saucers, at a bulk density (ρb) of 700 kg m⁻³ . Deionized (DI) water was slowly added to each pot to wet up the soil to a gravimetric water content of 40%. The pots were then moved to the glasshouse and arranged on tables in a randomized block design. The glasshouse was maintained within the range of 10–30 °C, with a mean temperature of 19.9 °C for the duration of the experiment.

Plant material and rhizobia inoculant

Volunteer seedlings were removed 10–14 days post wetting up of the soil. Seeds of the target pasture legume species were pre-treated by soaking and scarification and then sown directly, using a ‘grid’ sowing pattern using five sowing positions and two seeds at each position, on to the soil surface in the pots as monocultures, in late September (late spring). A small quantity (2 mm depth) of soil was sprinkled over the seed and gently packed. Soon after germination, all legume species were thinned to a final plant density of five plants per pot, to a plant density of 450 plants m⁻², based on field observations (Maxwell et al. 2010). The pasture legume species evaluated were: white clover (T. repens cv. Nomad), subterranean clover (T. subterraneum cv. Mt Barker), striated clover (T. striatum), suckling clover (T. dubium), cluster clover (T. glomeratum) and haresfoot trefoil (T. arvense). Seeds of the traditionally sown pasture legume species investigated were sourced commercially. Seed of the naturalized legume species were collected from several hillslope farmland sites at Glenfalloch and Mt Grand Stations.

Commercial rhizobia inoculant (Nodulaid, Group B and C, Becker Underwood), Group B for all white clover pots and Group C for all annual clover pots, was added 5 weeks after seed germination to ensure that an active soil rhizobia population was present. At 5 weeks post germination of seeds, a small quantity of N (30 kg N ha⁻¹) nutrient solution, in the form of ammonium nitrate (NH₄NO₃), was applied to all pots in order to overcome any plant N deficiencies during the seedling establishment phase. Beyond this point, plants were dependent on N sourced from N fixation or soil N for growth (plants of all species were later confirmed to have nodulation on visual inspection). In addition, a nutrient solution (Booking 1976; Caradus & Snaydon 1986) containing trace elements only was applied on a regular basis (twice monthly over winter, then weekly thereafter during rapid growth in spring and early summer), to ensure adequate trace element nutrition. Total quantities of trace elements applied were 500, 200, 40 and 1000 g element ha⁻¹ for boron (B), cobalt (Co), molybdenum (Mo) and zinc (Zn), respectively. Throughout the experiment, all pots were watered with DI water daily to maintain a gravimetric soil moisture content of 40%, and watered to weight twice weekly.

Yield and herbage analyses

Seven herbage harvests were conducted from late November 2008 to early May 2009, representing 9, 11, 13, 15, 20, 27 and 30 weeks post germination. The total duration of the experiment was 229 days. Clover plants were harvested by cutting 2 cm from the crown of each plant. Plant material from each pot was then dried at 70 °C for 48 h and weighed for DM yield. Samples were then ground, acid digested (Kjeldahl digest procedure; Blakemore et al. 1987) and analysed for total P concentration by molybdenum blue using an FIA (Flow Injection Analyser, Tecator Inc, Sweden). This information was then combined to give total P uptake for the duration of the experiment. Individual Olsen P tests were run on soil from all pots to determine the level of plant-available P at the conclusion of the experiment (Table 2). Further soil analyses were conducted on a bulked soil sample from all P treatment pots.

Table 2 . Values of total accumulated shoot yield, P concentration and P uptake by six pasture legume species, grown under glasshouse conditions in a New Zealand high country soil supplied with increasing rates of P (eight levels of P; increasing from 0 to 2500 mg P kg⁻¹ soil) and Olsen P values of soil in which plants were grown.

Species		Mean shoot yield g DM/pot	Mean shoot P concentration % P	Mean shoot P uptake mg P/pot	Olsen P^1 mg kg^−1
T. glomeratum		1.31	0.41	5.3	82
T. arvense		3.37	0.40	14.1	77
T. subterraneum		4.40	0.59	24.1	81
T. dubium		1.44	0.55	8.1	64
T. striatum		0.81	0.46	3.8	79
T. repens		1.67	0.39	7.5	78
Grand mean		2.168	0.470	10.47	77
Species	LSD (5%)	0.173	0.059	1.43	4.3
P rate	0	1.26	0.33	4.3	11
	30	1.87	0.35	6.4	13
	60	1.79	0.36	6.1	16
	100	2.12	0.38	8.1	21
	250	2.57	0.46	10.9	46
	500	2.82	0.50	14.1	84
	1000	2.52	0.61	16.0	151
	2500	2.39	0.74	17.9	271
	LSD (5%)	0.199	0.068	1.65	4.9
P	Species	***	***	***	***
	P rate	***	***	***	***
	Sp×P Rate	***	***	***	***
^1Olsen P values at the completion of the experiment.
***Significant at P<0.001

Statistical analyses

The effects of applied P on DM yield, plant shoot P concentration, plant shoot P uptake and final soil Olsen P was analysed by an analysis of variance (ANOVA) of a randomized block design, using GenStat 12.2 (Lawes Agricultural Trust, Rothamsted, UK). The model included P rate, legume species and the P rate×species interaction as fixed effects. Highly significant (P <0.001) interaction effects were observed and therefore regression analysis and curve fitting was undertaken for all species individually in order to better interpret the response of species to P rate.

Results

Yield response

Plant growth response to applied soil P differed among pasture legume species, and between P application rates (Fig. 1, Table 2, P rate×Sp interaction, P<0.001). Averaged across P rates, mean total accumulated dry matter (TDM) yield ranged from 0.8 to 4.4 g DM/pot. Trifolium subterraneum was the most productive species (4.4 g DM/pot) and T. striatum yielded the least (0.8 g DM/pot) (Table 2).

Figure 1.  Total accumulated shoot dry matter (DM) yield response of pasture legume species—A, Trifolium glomeratum; B, T. arvense; C, T. subterraneum; D, T. dubium; E, T. striatum; and F, T. repens—to increasing levels of soil phosphorus (eight levels of P; ranging from 0 to 2500 mg P kg⁻¹ soil), grown in a New Zealand high country soil. Data are mean values ± SEM (n =4), with R ² and P values for fitted curve showing data trend.

The soil P level at which maximum DM yield occurred varied among species (Fig. 1, Table 3). Three species (T. subterraneum, T. repens and T. glomeratum) showed a clear rise to maximum DM yield at 500 mg P kg⁻¹, followed by a decline (excluding T. repens) in yield at higher P rates (Fig. 1). Two species (T. arvense and T. dubium) showed a rapid rise in yield up to 250 mg P kg⁻¹ soil, then plateaued, with no decline in yield. In the case of T. dubium, yield increased again slightly at the highest soil P level (Fig. 1). Trifolium striatum rose quickly to a maximum yield response at 250 mg P kg⁻¹ soil, then showed a steady decline in yield as soil P level increased.

Table 3 . Rate of P application and shoot P concentration at which 95% maximum yield was observed for the pasture legume species.

Species	Maximum yield g DM/pot	P Rate mg P kg^−1	Shoot P concentration %
T. glomeratum	1.9	500	0.44
T. arvense	4.3	250	0.38
T. subterraneum	6.3	500	0.61
T. dubium	1.7	2500	0.96
T. striatum	1.3	250	0.46
T. repens	2.6	500	0.47

Shoot P concentration and uptake

Mean shoot P concentration was highest for T. subterraneum and T. dubium, intermediate for T. striatum and lowest for T. glomeratum, T. arvense and T. repens (Table 2). All species, except T. striatum, showed the general pattern of increasing shoot P concentration as soil P availability increased (Fig. 2). Shoot P concentration increased at a similar rate from 0 to 2500 mg P kg⁻¹ soil for T. arvense, T. subterraneum, T. dubium and T. repens. Shoot P concentration of T. glomeratum increased more rapidly between 0 and 250 mg P kg⁻¹ soil before flattening off and steadily increasing in a similar pattern to the other species. Trifolium striatum shoot P concentration was high with no added P, but showed a general decline and then minimal change as soil P level increased. The latter trend was weakest of all the species (R ²=0.64; Fig. 2).

Figure 2.  Comparison of shoot P concentration of pasture legume species—A, T. glomeratum; B, T. arvense; C, T. subterraneum; D, T. dubium; E, T. striatum; and F, T. repens—grown in a New Zealand high country soil supplied with increasing levels of soil phosphorus (eight levels of P; ranging from 0 to 2500 mg P kg⁻¹ soil). Data are mean values ± SEM (n =4), with R ² and P values for fitted curve showing data trend.

The relationship between shoot P concentration and 95% max yield varied among the species (Table 3). P concentration at which 95% maximum yield was achieved was highest for T. dubium (0.96%) and lowest for T. arvense (0.38%) (Table 3). P uptake also varied among the six pasture legume species (P rate×Sp interaction, P<0.001). The mean level of P uptake ranged from 3.8 mg P/pot in T. striatum to 24.0 mg P/pot in T. subterraneum (Table 2).

P-response efficiency

The P rate at which the highest P-response efficiency (defined as P applied g⁻¹ DM grown; Syers et al. 2008) was observed varied among species (P rate×Sp interaction, P<0.001). Trifolium glomeratum, T. arvense, T. subterraneum and T. repens all showed the highest P-response efficiency at 30 mg P kg⁻¹ soil (Fig. 3). In contrast, the highest P-response efficiency for T. dubium and T. striatum occurred at 100 mg P kg⁻¹ soil. In general, the order of P-response efficiency by species was as follows: T. subterraneum>T. arvense>T. repens>T. glomeratum>T. dubium >T. striatum (Fig. 3).

Figure 3.  Comparison of P-response efficiency of pasture legume species—A, T. glomeratum; B, T. arvense; C, T. subterraneum; D, T. dubium; E, T. striatum; and F, T. repens—grown in a New Zealand high country soil supplied with increasing levels of soil phosphorus (eight levels of P; ranging from 0 to 2500 mg P kg⁻¹ soil). Data are mean values ± SEM (n =4), with R ² and P values for fitted curve showing data trend.

Discussion

P yield response

Total DM yields differed between species, in the order of T. subterraneum>T. arvense>T. repens>T. dubium≥T. glomeratum>T. striatum. Also, the annual T. arvense out-yielded the perennial T. repens cv. Nomad at every soil P level. This suggests that T. arvense may have a stronger agronomic potential than T. repens cv. Nomad, a cultivar selected for increased persistence in dry environments prone to low summer moisture levels (Agricom 2010). Trifolium glomeratum and T. dubium yields did not differ significantly from each other, although both out-yielded T. striatum as soil P levels increased.

The superior yield performance of T. subterraneum cv. Mt Barker can be explained by its large seed size, with greater endosperm reserves relative to all the other species. This enabled rapid early seedling development, and subsequent exhibition of its good agronomic potential, under the optimum soil and climatic glasshouse conditions. Caradus (1980) reported on the responsiveness of T. repens cv. Huia, T. subterraneum cv. Woodenellup, T. striatum, T. arvense and T. dubium to added P, in terms of shoot dry weight and total shoot P content, grown at relatively high P rates of 300 and 2000 mg P kg⁻¹, in a P-deficient soil (volcanic Stratford coarse sandy loam) over 24 weeks. Similar to our study, T. striatum was the lowest yielding least P responsive annual legume species. The yield results for T. arvense and T. dubium, however, contrast our results, as T. arvense yielded less and T. dubium yielded more in that study. This was possibly due to ecotypic variation in germplasm used between the two experiments. Trifolium subterraneum was the most productive species across all P rates. Differences in cultivar of T. subterraneum and T. repens grown by Caradus (1980) to those in our study may explain the contrasting trends in yield responses between the two comparable P rates; both species had greater yields at 2000 mg P kg⁻¹ soil than at 300 mg P kg⁻¹ soil which contrasts the yield response trends for these species in our study. However, the two relatively high P rates used by Caradus (1980) do not allow for the examination of P response of these species at low soil P fertility, which is a critical focus of our experiment for low-P high country environments. Additionally, the volcanic soil examined by Caradus (1980) contrasts with the sedimentary high country Allophanic Brown soil of our study, in terms of soil parent material (high allophane content and very high P retention capacity) and environmental soil-forming conditions.

Blair & Cordero (1978) also found that T. subterraneum out-yielded T. glomeratum in a 10-week pot trial. Caradus et al. (1995) reported only minor yield responses of T. repens beyond 400 mg P kg⁻¹, which is comparable to our study. Hart & Jessop (1984) examined the growth responses to P of T. repens growing in an Egmont sandy loam, a very N deficient soil of high phosphate fixing capacity. Phosphorus was added to the soil at levels equivalent to those used in our study. In accord with our results, they observed T. repens responding more strongly to added P than T. dubium, with the latter having lower shoot dry weights. Steepest increases in shoot dry weight of both species occurred at lower P rates (0–500 mg P kg⁻¹ soil). They concluded T. dubium to be a species that has a relatively small response to improvements in P availability, which is in agreement with the results of this experiment.

Maximum yield

The soil P level at which maximum yield was observed varied between species suggesting that the optimum P requirement of these species is different. Maximum yield response for T. subterraneum, T. glomeratum and T. repens occurred at an equivalent rate of 221 kg P ha⁻¹. Beyond this point however, these three species were unable to utilize further increases in available P for shoot growth, suggesting that factors other than plant-available soil P were limiting yield. For T. dubium, maximum yield was observed at the highest soil P level treatment suggesting this species has the ability to take up P available in the soil at very high levels. This species appears to be able to continue to take up P over a wide range of availability, although remaining very unresponsive to P in terms of DM yield, suggesting perhaps a very low P requirement to reach its growth potential.

Shoot P concentration and uptake

In general, all pasture legume species, with the exception of T. striatum, showed steadily rising shoot P concentrations with increased availability of soil P. In agreement with this, Hart & Jessop (1984) reported leaf P concentration of T. dubium and T. repens cv. Huia rose with increasing P supply. Blair & Cordero (1978) found that the shoot P concentration of T. subterraneum and T. glomeratum increased with increasing rates of applied P. However, shoot P concentration continued to increase beyond the point of maximum yield response for T. glomeratum, T. arvense, T. subterraneum and T. repens (Fig. 1), suggesting ‘luxury’ P uptake by these species. There are no reports suggesting P toxicity in pasture legumes grown under glasshouse conditions, which would cause a reduction in P yield response.

The shoot P concentration at which maximum yield was observed varied among the pasture legume species. Interestingly, T. arvense had a much lower shoot P concentration at maximum yield than T. subterraneum. The magnitude of P uptake varied between the clover species, with T. subterraneum showing the greatest mean level of P uptake from the soil across all P rates. As expected, P uptake reflected the DM yield of the clovers.

P efficiency

The ability to acquire P from the soil and use it efficiently for biomass production is an important characteristic for adaptation to soils low in available P (Pang et al. 2010). In general, at low-medium P levels, T. subterraneum was the most efficient species in utilizing applied P for biomass production (P required g⁻¹ DM grown), followed by T. arvense and T. glomeratum. Beyond very low soil P level, T. repens became the least efficient out of the top three higher-yielding pasture legume species. These results clearly indicate the inefficiency of T. repens to utilize P at very low soil P levels, in contrast to T. subterraneum, T. arvense and T. glomeratum.

Practical implications and conclusions

Gaining an understanding of P requirements of plant species is important for the purposes of introduction, selection and breeding. Trifolium arvense, due to high yield and high P efficiency, showed the most promise as a species to be utilized, and its spread and abundance through grazing management encouraged in extensive high and hill country ecosystems. The critical P requirement for maximum yield for T. arvense was lower than T. repens cv. Nomad and T. subterraneum cv. Mt Barker. Trifolium arvense had almost twice the DM yield and greater P response efficiency at lower soil P levels on this soil than T. repens. In addition, being an annual pasture species, T. arvense is able to complete its life cycle before soil moisture deficits occur in late spring-early summer in summer-dry hill country. Although being less productive than T. subterraneum in this study, the herbage DM production of T. arvense could be improved by breeding and selection of better performing cultivars, and searching to find ecotypic variation within the large areas of New Zealand high and hill country where T. arvense is found to survive and persist (Maxwell et al. 2010). Trifolium glomeratum also shows promise with a comparable yield P response trend and greater P response efficiency than T. repens at very low soil P levels. The results of this 20-week glasshouse study need to be examined further in long-term field studies. Such studies, conducted under field climatic and environmental conditions, should identify how the more P responsive naturalized species such as T. arvense and T. glomeratum, when grown as monocultures or mixtures, perform when subjected to realistic P fertilizer rates in the field.

Acknowledgements

This research was funded by the Miss EL Hellaby Indigenous Grasslands Research Trust. We thank Chas Todhunter of Glenfalloch Station and Brendan Malcolm, Carole Barlow, Qian Liang, Sho Kasuya and Fiona McConville of Lincoln University for technical assistance.