Thermal time requirements for germination of four annual clover species

Abstract

Germination of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers was determined from constant temperature (5–40 °C) from incubator experiments. A ‘broken stick’ regression of germination rates against temperature was used to quantify thermal time (Tt) requirements and cardinal (base, optimum and maximum) temperatures. Tt requirements for 75% germination were higher for ‘Cefalu’ arrowleaf (34 °Cd) than ‘Bolta’ balansa (32 °Cd), ‘Prima’ gland (28 °Cd) and ‘Mihi’ Persian (25 °Cd) clovers. All species had a base temperature of 0 °C. ‘Mihi’ Persian clover had an optimum temperature (T_opt) of 33 °C and a maximum temperature (T_max) of 45 °C, while all the other species had T_opt between 16–20 °C and T_max of c. 34 °C. The maximum final germination percentage was above 60% for ‘Cefalu’ arrowleaf clover and 80% for the other three species.

Keywords:

• growing degree days
• Trifolium glanduliferum
• T. michelianum
• T. resupinatum
• T. vesiculosum

Introduction

Annual clovers are used in dryland pastures (<800 mm annual rainfall) to provide sheep grazing during late winter and spring. They grow vigorously during spring, set seeds, then die in summer. These annuals then re-establish from seeds in the autumn. Thus, annual clovers avoid summer drought as seeds. Subterranean clover (Trifolium subterraneum L.) is the most widely used annual legume species in New Zealand. It is capable of producing large quantities of seed even under hard grazing (Smetham 2003). However, its seed burrs means specialized machinery is required for seed harvesting. This is not a problem for top flowering annual clovers. Arrowleaf clover (T. vesiculosum Savi) grows well during summer (Evans et al. 2003) and can be used to extend the grazing period for lamb finishing and improve soil fertility for future cropping (Evans & Mills 2008). Cultivar ‘Cefalu’, which flowers early, is suitable to fill in a feed gap during early spring in New Zealand. Balansa clover (T. michelianum Savi) produces high dry matter yield and has the capacity to generate a large seed bank that can be managed for regeneration in subsequent years (Monks et al. 2008). This research showed that when cultivar ‘Bolta’ was sown in a mixture with cocksfoot (Dactylis glomerata L.) in lowland Canterbury (<650 mm annual rainfall), it produced an annual dry matter yield of 3 t/ha or 30% of the total herbage. This indicates its ability to thrive in a dryland pasture. ‘Prima’, which is the only cultivar available of gland clover (T. glanduliferum Boiss), is highly resistant to pests such as red-legged earth mite (Halotydeus destructor), blue-green aphids (Acyrthosiphon kondoi) and cowpea aphids (Aphis craccivora) (Dear et al. 2001). It has a delayed pattern of hard-seed breakdown, which prevents the germinating seedlings from a ‘false break’ (Dear et al. 2002). Persian clover (T. resupinatum L.) tolerates water-logging and has high dry matter production during spring (Charlton & Stewart 2003). Cultivar ‘Mihi’, which belongs to subspecies majus, has an erect habit, large leaflets and is soft-seeded. This cultivar matures late and is used primarily as a specialist crop for fodder production. Presently, these top flowering annual clovers are not widely used in New Zealand and little is known about their potential performance in New Zealand pastures.

The persistence of annual pasture species depends on the population of seedlings that regenerate each autumn. Therefore, it is important to know the timing and rate of germination because these factors determine the number of seedlings that emerge from the ground. As long as moisture is adequate, seed germination is driven by the accumulation of thermal time (Tt) or growing degree-days (°Cd). Thermal time is defined as cumulative heat units above a base temperature on a daily basis. Quantification of thermal time for germination requires cardinal temperatures, namely base (T_b), optimum (T_opt) and maximum (T_max) to be calculated (Angus et al. 1981). The base temperature (T_b) is the temperature below which no germination occurs. The optimum temperature (T_opt) is the temperature at which the germination rate occurs at its maximum capacity. The maximum temperature (T_max) is the highest temperature at which seeds will germinate. Given that germination is a function of temperature, then the thermal time requirement for germination can be quantified (Arnold & Monteith 1974).

Thermal time has been used to quantify germination of some annual clover species (Moot et al. 2000; Boswell et al. 2003; Lonati et al. 2009; Monks et al. 2009). In most cases, the thermal time was species dependent and unaffected by cultivar. However, the range of temperatures used was inadequate to accurately define all three cardinal temperatures and gland clover was not included in any of these studies. Thus, this study aims to define cardinal temperatures and quantify thermal time requirements for germination of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers through a series of incubator experiments.

Materials and methods

Three replicates of 50 scarified bare seeds of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers were placed on wetted blotting paper in sealed plastic containers and germinated in unlit incubators at constant temperatures of 5, 8, 10, 12, 15, 18, 20, 25 or 30 °C. Additional temperatures of 22.5, 35, 37.5 and 40 °C were added later, to increase the number of data points below and beyond the optimum temperature (T_opt) and to assist determination of T_b and T_max. Distilled water was added as required to ensure moisture was non-limiting for germination. Germinated seeds were counted and removed twice daily during periods of rapid germination and daily at other times until germination ceased (ISTA 2004). Seeds were considered germinated when the radical exceeded the small diameter of the seed. A Gompertz model was fitted to the cumulative percentage germination against days:

(1)

Where CG is the cumulative percentage of seeds germinated at time, t (days), C is the final germination percentage and B and M are constants. The number of days to 75% of the final germination percentage (t₇₅) was calculated using Equation 2 derived from the Gompertz model where CG = 75:

(2)

Data analysis

Data for each species were plotted as the reciprocal of the duration (in days) to 75% germination against the mean temperature (T). The inverse of duration (1/days) represents the germination rate. Least squares regression analysis was used for both the positive (sub-optimal) and the negative (supra-optimal) linear portions of the response whereby:

(3)

(4)

Where, a is the y-axis intercept, b is the slope of the relationship and T is the temperature between the T_b and T_opt (for the positive linear portion) and between T_opt and T_max (for the negative linear portion). The regression coefficients can then be related to T_b, T_max and Tt as:

(5)

(6)

(7)

(8)

The T_opt was interpolated from the intersection of the two regression lines as:

(9)

Where the germination rate deviated from a linear model at low or high temperatures, data were excluded from the analysis on the basis that these were outside the species' optimal thermal range (Angus et al. 1981). Where 95% confidence intervals included 0 °C, an additional regression analysis was performed with T_b set at 0 °C to enable direct comparison of the Tt requirements for each development stage among species and previously published results e.g. Moot et al. (2000).

The standard error of T_b and Tt were calculated according to (Campbell et al. 1974) as:

(10)

(11)

Where s² is the residual mean square of y, i.e. the development rate and is the sample mean.

Data were analysed using the statistical software Genstat 12.2. For each measured variable, maximum standard errors were reported. In most analyses, a single thermal time accumulation target is reported. This assumes the cardinal temperatures are representative of a normal distribution that is symmetrical and centred around T_opt (Bonhomme 2000). However, where the data show a skewed or non-symmetrical relationship, a separate thermal time target may be required for sub- and supra-optimal temperatures.

Results

Final germination percentage

Gompertz functions described the cumulative germination over time for each species (Fig. 1). Specifically, for each species by temperature combination, there was a distinct linear phase before the final germination percentage was determined. The maximum final germination percentage was above 60% for arrowleaf clover and 80% for the other three species. However, at the higher and lower end of the temperature range, the final germination percentage was below these maximum values (Fig. 2). For example, ‘Cefalu’ arrowleaf had a final germination percentage of 58%–71% from 10–25 °C but this dropped to less than 50% at 5 °C and 35 °C with no germination at higher temperatures. ‘Mihi’ Persian had the highest final germination of 90%–95% at 30 °C but this decreased to 80% at 5 °C and 41% at 40 °C. Germination was not observed for ‘Cefalu’ arrowleaf, ‘Bolta’ balansa or ‘Prima’ gland clovers at 40 °C.

Figure 1 Cumulative germination of (a) ‘Cefalu’ arrowleaf, (b) ‘Bolta’ balansa, (c) ‘Prima’ gland and (d) ‘Mihi’ Persian clovers at 13 different constant temperatures. (•) 5 °C, (○) 8 °C, (▾) 10 °C, (▵) 12 °C, (▪) 15 °C, (□) 18 °C, (♦) 20 °C, (⋄) 22.5 °C, (▴) 25 °C, (▿) 30 °C, (⬢) 35 °C, (⬡) 37.5 °C, (×) 40 °C. Error bars represent the maximum standard error of the mean for final germination percentage.

Figure 2 Maximum germination (%) for ‘Cefalu’ arrowleaf (•), ‘Bolta’ balansa (○), ‘Prima’ gland (▾) and ‘Mihi’ Persian (▵) clovers at different constant temperatures.

Time to 75% germination and germination rate

The number of days to reach 75% of final germination decreased as temperature increased up to 20 °C for both ‘Cefalu’ arrowleaf and ‘Bolta’ balansa, 15 °C for ‘Prima’ gland and 30 °C for ‘Mihi’ Persian clovers (Fig. 3). There was then an increase in the duration until the maximum effective temperature was reached: 35 °C for ‘Cefalu’ arrowleaf and ‘Bolta’ balansa, and 40 °C for ‘Mihi’ Persian clover. For ‘Prima’ gland clover, the time to germination remained the same from 15–20 °C before it increased up to the maximum effective temperature of 30 °C. The reciprocals of these data were used to define the cardinal temperatures for each species.

Figure 3 Number of days to 75% of final germination for ‘Cefalu’ arrowleaf (•), ‘Bolta’ balansa (○), ‘Prima’ gland (▾) and ‘Mihi’ Persian (▵) clovers at different constant temperatures.

In all species, the germination rate showed a positive linear trend from the minimum temperature (T_b) up to the optimum temperature (T_opt), and then a negative linear decline until no germination occurred at the maximum temperature (T_max) (Fig. 4). Split-line regression analysis estimated a T_b of 1.4 °C for ‘Cefalu’ arrowleaf, 2.6 °C for ‘Bolta’ balansa, −0.1 °C for ‘Prima’ gland and 4.5 °C for ‘Mihi’ Persian clover. For all species, the 95% confidence interval showed T_b estimates for germination included 0 °C. ‘Cefalu’ arrowleaf, ‘Bolta’ balansa and ‘Prima’ gland clovers had a maximum germination rate of between 0.6–0.7 per day across a T_opt range of 16.2–19.7 °C and T_max of c. 34 °C. ‘Mihi’ Persian clover had the highest T_opt of 32.6 °C (1.5 per day) and a T_max estimate of 45.4 °C.

Figure 4 Germination rate of ‘Cefalu’ arrowleaf (•), ‘Bolta’ balansa (○), ‘Prima’ gland (▾) and ‘Mihi’ Persian (▵) clover at different constant temperatures.

The Tt requirement for 75% of final germination for temperatures in the sub-optimal range was 25( ±1.0) °Cd for ‘Mihi’ Persian and 34 (±0.9) °Cd for ‘Cefalu’ arrowleaf above a T_b of 0 °C (Table 1). In the supra-optimal range, Tt requirement (from T_opt to T_max) was similar to the sub-optimal range except for ‘Mihi’ Persian clover (9 [±1.7] °Cd).

Table 1 Base (T_b), optimum (T_opt) and maximum (T_max) temperatures and thermal time (Tt) requirements for 75% germination of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers.

Cultivar	Tb (°C)	Topt (°C)	Tmax (°C)	Ttsub (°Cd)	Ttsup (°Cd)	^aTt (Tb=0°C) (°Cd)	^bExcluded temperatures
‘Cefalu’	1.4	19.7	34.8	30	25	34	37.5, 40
‘Bolta’	2.6	19.7	33.9	26	21	32	37.5, 40
‘Prima’	0.1	16.2	34.1	29	32	28	37.5, 40
‘Mihi’	4.5	32.6	45.4	19	9	25	_
Max s.e.	0.48	1.22	0.88	1.8	4.9	1.3
95% c.i.	−5.0, 10.5	13.8, 34.7	25.7, 53.6

^a Analysis assumes a base temperature of 0 °C.

^b Excluded temperatures because germination rate deviated from the linear model.

Tt_sub, from T_b to T_opt

Tt_sup, from T_opt to T_max

s.e. Standard error.

c.i. Confidence interval.

Discussion

In all species, the germination rate increased linearly with increased temperatures up to an optimum. Further increases in temperature then decreased the germination rate until there was no further germination (Fig. 4). The relationship between germination rate and temperature enabled the cardinal temperatures (base, optimum and maximum) to be defined. A two-piece ‘broken stick’ regression model was appropriate for describing the germination response of these annual clovers at the sub-optimal and supra-optimal range of temperatures. In most cases, a single thermal time target of a phenophase is appropriate, based on the assumption that the slopes of the two regression lines are symmetrical. In this case, a single thermal time target can be used to quantify germination of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa and ‘Prima’ gland clovers. However, for ‘Mihi’ Persian clover, the asymmetric relationship meant thermal time requirements differed at sub-optimal and supra-optimal temperatures. Also, in the supra-optimal range, germination rate at 37.5 and 40 °C for ‘Cefalu’ arrowleaf, ‘Bolta’ balansa and ‘Prima’ gland clovers deviated from the negative linear model (Fig. 4) and were excluded from analysis. These were outside the species' optimal thermal range (Angus et al. 1981). The germination rate deviated because at these high temperatures, seeds either germinated rapidly or died and the germination rate was overestimated, based on the 75% germination duration of the surviving seeds. At 37.5 °C, the seeds of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa and ‘Prima’ gland clovers took six, three and five days, respectively, to germinate. However, the seed population that survived at 37.5 °C was <20% for both ‘Cefalu’ arrowleaf and ‘Prima’ gland clovers and <45% for ‘Bolta’ balansa clover (Fig. 2). Except for ‘Mihi’ Persian clover, no germination occurred at 40 °C. ‘Mihi’ Persian clover had greater tolerance to high temperatures with its fastest germination rate at 33 °C and 40% of its seed population germinated at 40 °C. The ability of ‘Mihi’ Persian clover to germinate at high temperature is unknown. ‘Mihi’, a soft seeded cultivar that belongs to a subspecies majus, is developed for use in irrigated farm for forage production. Cultivars from subspecies majus may have been selected for early germination under warm conditions, as they are often sown under irrigation in late summer (February), when temperatures can still be high (Anonymous 2003).

In the moist temperate climate of New Zealand, the soil temperatures that seeds would experience were within the sub-optimal range. However, during dry summer, near soil surface temperature has been reported to exceed 30 °C in the Bay of Plenty (Watson et al. 1996) and >40 °C at the ‘Ashley Dene’ dryland farm, Canterbury (R. Sim, unpubl. data). The ability of ‘Mihi’ Persian clover to germinate at 40 °C suggests that in any given population, there is a small percentage of seeds that can germinate outside the optimum temperature range. From an ecological perspective, this may be an adaptive strategy for the survival of an annual species under extreme (>37.5 °C) soil temperature conditions.

Cardinal temperatures for germination of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers showed that these species are adapted to a broader range of temperatures compared with subterranean clover and the adventive annual clovers namely, haresfoot (Trifolium arvense), suckling (T. dubium), cluster (T. glomeratum) and striated (T. striatum) clovers (Boswell et al. 2003; Lonati et al. 2009). This means that ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers could inhabit the same ecological niche as the subterranean clover and adventive clovers, and also germinate in warmer areas where some other species cannot. In addition, ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers germinate quicker than subterranean and white clovers (Moot et al. 2000). The thermal time requirements for germination of ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers were similar (Table 1) to the adventive clovers (Boswell et al. 2003; Lonati et al. 2009). Their similarity in temperature response and germination rate suggests that ‘Cefalu’ arrowleaf, ‘Bolta’ balansa, ‘Prima’ gland and ‘Mihi’ Persian clovers may be suitable for over-sowing in hill and high country where the adventive clovers are abundant (Maxwell et al. 2010).

Within the range of sub-optimal temperatures, the final germination of each species was above 50%. The concept of thermal time summarized the range of individual temperature responses (Fig. 1) into a single coefficient that can be used for all temperatures in the sub-optimal range. For example, thermal time calculated for germination of ‘Cefalu’ arrowleaf clover was 30 °Cd, hence at 5 °C ‘Cefalu’ arrowleaf could be expected to germinate in six days compared with two days at 15 °C. In all species, the T_bs calculated were ≤4.5 °C and were not different from 0 °C (Table 1). These values were consistent with those reported by Monks (2009) and Lonati (2009) and therefore suggest that future work in annual clover species could assume a T_b of 0 °C for germination.

The use of 95% confidence interval to determine whether T_b was different from 0 °C (Table 1) highlights the need to understand bases and limits of using linear relationships to describe development rate with temperature. This is because the linearity between development rate and temperature is only limited for a short range of temperatures. This relationship is practically exponential for temperatures approaching the T_b (Angus et al. 1981; Bonhomme 2000). In this study, the lowest temperature used for germination tests was 5 °C. At too low a temperature (<5 °C), the enzyme is insufficiently flexible to carry the reaction; hence seeds either took too long to germinate or eventually rotted and died (Toole et al. 1956). Given that the T_b is calculated from an extrapolation of the linear relationship between development rate and temperature, this method excludes the exponential relationship near the physiological T_b. Therefore, calculated T_b from extrapolation is often higher than the actual (physiological) T_b for which development is zero. For example, the T_b extrapolated for ‘Mihi’ Persian clover based on the linear relationship was 4.5 °C. In reality, there is likely to be some germination occurring below the calculated T_b because germination was still above 80% at 5 °C (Fig. 1).

Non-linear models such as Lactin (Lactin et al. 1995), Beta (Yan & Hunt 1999), line plus exponential and critical exponential (standard curves from Genstat) were also tested to define cardinal temperatures for germination in this study but none of these models gave a better fit to the data points (based on the R² value) compared with the ‘broken stick’ linear model. Therefore, the use of the linear model was considered appropriate to describe development rate with temperature. It emphasizes the idea that T_b is only an approximation of the actual temperature threshold (Angus et al. 1981; Bonhomme 2000). Determining T_b is crucial to calculate thermal time accumulation between phenophases and when 95% confidence interval for T_b includes 0 °C, it is possible to re-analyse thermal time using a T_b of 0 °C to allow comparisons among species (Moot et al. 2000).

Conclusions

The use of a linear model was appropriate to define the cardinal temperatures and thermal time requirements for germination of annual clovers. The thermal time requirements for germination were highest for ‘Cefalu’ arrowleaf (34 °Cd) and lowest for ‘Mihi’ Persian (25 °Cd) clovers. ‘Mihi’ Persian clover had an optimum temperature (T_opt) of 32.6 °C while all the other species had a T_opt between 16–20 °C. The effect of temperature on germination showed that ‘Mihi’ Persian clover seeds can survive under extreme temperature beyond 35 °C.

Acknowledgements

H. Nori acknowledges the Ministry of Higher Education Malaysia for a scholarship. Part of this work was funded through the ‘Pastoral 21’ programme, supported by Beef + Lamb New Zealand.