SPRING AND SUMMER ROOT GROWTH OF SOME TEMPERATE–REGION GRASSES AND SUMMER ROOT GROWTH OF TROPICAL GRASSES

The root growth of 3 temperate–region grasses (perennial ryegrass, Italian ryegrass, and cocksfoot) and of 3 tropical grasses (Dallis grass, Bahia grass, and Rhodes grass) was investigated in large glass–sided root–boxes. The trend of root numbers and the rate of elongation of roots were ascertained to discover how the Japanese climate affected the root growth of these species.
All the temperate–region grasses produced new roots vigorously during spring, but root growth ceased in August, whereas the roots of tropical grasses grew rapidly in this month. The alternate use of these two types of grasses during the growing season may therefore be worthwhile in Japan     

SEED PRODUCTION STUDIES ON PERENNIAL RYEGRASS, TIMOTHY AND PRAIRIE GRASS

A study over two years involving tiller identifica tion according to month of origin enabled the contribution of individual tillers of perennial rye grass Grasslands Ruanui, timothy Grasslands Kathu and a local New Zealand strain of prairie grass (Bromus unioloides) to be determined at seed harvest. In addition, the distribution of seed-heads in different arbitrary emergence groups was detemined so as to allow analysis of individual seedhead components within each group. In all species, tillers formed during the first 4 months following an autumn sowing in the first year, and also in the immediate post-harvest period and through the autumn in the second year, made a major contribution to seedhead numbers and seed yield at harvest. These tillers had a lower vegetative mortality rate than those formed in the spring. Primary tillers formed after sowing were highly persistent and became almost exclusively repro ductive. Many vegetative tillers died before har vest, behaving as annual tillers. Spring-formed tillers made a major contribution to the vegetative growth of the plant over the summer and early autumn. These tillers generally died before the following winter. Tillers formed in winter and early spring by ryegrass and prairie grass showed a high mortality, particularly in the second year. In timothy, mortality was highest amongst tillers formed in late spring. The reduced contribution of spring-formed tillers to the seed head population occurred as a result of fewer heads numerically compared with older tillers. In all species, most seedhead components varied according to time of ear emergence. Early emerged heads were generally longer, bore more spikelets and more florets per head and had a greater culm length than late-emerged heads. The major trend in all species was a general depression in head length, spikelet and floret number and culm length with grazing. These effects were in most cases partly or completely overcome by N application. The major effect of applied N was to increase floret number per head.     

Milk-production potential of different sward types in a temperate southern Australian environment

The milk-production potential of different sward types was measured in each of the four seasons of the year in a replicated experiment in south-west Victoria, Australia. Dairy cows were offered ad libitum allowances of a 'short-term winter' sward, based on Italian ryegrass (treatment STW), a 'long-term winter' sward, based on a winter-active tall fescue (treatment LTW), a 'long-term summer' sward, based on a summer-active tall fescue (treatment LTS) and a Control sward (perennial ryegrass) in four seasons (days in milk in parentheses): spring (November dairy cows, 124), summer (February, 227), autumn (May, 234) and winter (August, sixty-four). A 'short-term summer' sward, based on chicory and white clover (treatment STS), was also included in summer. There was a significant season × treatment interaction caused by a more gradual decline in milk yield from peak for cows grazing treatment LTS compared to the Control treatment in the transition period from spring to summer. In summer, cows grazing treatment STS produced more milk (1·41 kg fat plus protein per cow d⁻¹) than cows grazing all other treatments (0·92 kg per cow d⁻¹) because of the superior nutritive value of herbage of pre-grazing pasture and higher apparent dry-matter intakes. Swards based on alternative species to perennial ryegrass are capable of supporting milk production that is at least comparable over an annual cycle.     

A comparison of perennial ryegrass‐ and tall fescue‐based swards with or without a cropping component for dairy production: Animal production,herbage characteristics and financial performance from a 3‐year farmlet trial

Perennial ryegrass/white clover swards have some limitations in temperate grazed dairying systems. This study tested the hypothesis that farmlets based on alternative species would be equally or more profitable than those perennial ryegrass‐based, and would produce more herbage in summer‐dry conditions. Six farmlets were established; three with perennial ryegrass‐ and three with tall fescue‐based swards. For each grass species, one farmlet was solely based on grass‐clover swards while the other two had either chicory‐red clover or lucerne crops planted on 20%–25% of the area. Animal‐ and herbage performance‐related variables were measured for 3 years, and calculated financial performance was evaluated. Using tall fescue improved total annual herbage yield compared with perennial ryegrass, but animal production and operating profit were lower. This was likely due to the reduced yield and nutritive value of tall fescue during spring and an associated decline in daily milksolids production. The deficit in spring milksolids production was never recovered, despite greater herbage production from tall fescue during summer/autumn. Incorporating chicory‐red clover or lucerne crops reduced both annual herbage and milksolids production. This reduced farm income, while increasing operating expenses as the farmlets required crop renewal and more purchased supplementary feed to maintain feed supply. Under the conditions of this study (i.e., partial irrigation, high nitrogen supply), changing the forage base from perennial ryegrass to tall fescue did not improve animal production or profitability, nor did incorporation of crops on 20%–25% of the farmlet area.     

STUDIES ON THE ROOT DEVELOPMENT OF HERBAGE PLANTS

The influence of defoliation on the root, stubble and herbage weights of perennial ryegrass during establishment was studied on spaced plants in the greenhouse and field.
In the greenhouse trial the rate of root elongation was reduced by a single defoliation. In both the field and greenhouse, cutting reduced the number of roots and tillers per plant but increased the number of roots per tiller. A few weeks after defoliating plants in the field there was a lower root weight on the cut plants than on the uncut. Eventually the influence of a single cut disappeared, but if the cutting was in the laie summer or autumn the plants commenced the winter with a smaller amount of root and stubble, and this appeared to have a deleterious effect on the earliest spring growth.
Herbage growth in March and April was positively correlated with both root and stubble weights in the previous November.
As the number of cuts during the establishment period (March-November) was increased from 0–4 the root and stubble weight per plant progressively decreased.
The root and stubble weights decreased during the winter.     

The relationships between stolon characteristics, winter survival and annual yields in white clover (Trifolium repens L.)

Five white clover populations of Swiss origin and three bred varieties were grown in binary mixtures with two perennial ryegrass varieties, Aurora and S23. The seasonal yields of clover and grass plus clover were measured under a cutting regime during the second and third years after establishment. A series of destructive detailed sward measurements was made during the late autumn to spring period preceding each harvest year. In this way changes in the amounts of stolon, leaf plus petiole and numbers of growing points were monitored during the winter.
There were large differences in clover yield between populations in both years. These were evident from the first (spring) harvest in each year. Higher-yielding clovers in spring tended to produce higher annual clover yields. No grass × clover interaction was evident at any harvest. Large differences between clovers were also apparent in the morphological characteristics measured, with the Swiss material generally having greater amounts of stolon, leaf plus petiole and numbers of growing points present in early spring. It is proposed that these factors contribute to the high spring yield in the Swiss populations. Loss of stolon length over each winter was less in the Swiss material, indicating that its good spring growth was not obtained at the expense of winter hardiness. Annual clover yield was found to be significantly positively correlated with the amount of stolon present in spring, exemplifying the importance of stolon survival over the winter.     

THE PRODUCTION OF EARLY SPRING GRASS

The effects of autumn management and nitrogenous manuring on the production of early spring grass were studied from 1954 to 1957 on farms throughout England and Wales.
The most important management factor influencing spring yields was the application of spring nitrogen. In the early spring 6 to 10 lb. of dry matter were produced per lb. of nitrogen applied. Autumn nitrogen increased herbage yields in the auttimn by 8 to 17 lb. of dry matter per lb. of nitrogen applied.
Italian and perennial ryegrass responded well to autumn nitrogen but it was important to graze off the extra growth fairly quickly to prevent the sward becoming too open. The yield of ryegrass swards which received spring nitrogen were very much higher if they were grazed during the previous September and October than if they had been rested during those months.
In two years neither the autumn management nor alumni fertilizer appeared to influence the spring yields of cocksfoot, timothy or meadow fescue swards, in one year, however, when autumn growth was poor, autumn nitrogen and resting resulted in slightly better yields on these swards in the following spring.
It is suggested that resting swards during September and October to build up "reserves" is relatively unimportant in relation to the yields of grass in the following spring.     

The value of timothy (Phleum pratense L.) in ryegrass — timothy mixtures managed to simulate intensive grazing

The performance of timothy in mixtures with perennial ryegrass was assessed under a simulated intensive grazing management over two harvest years in 1974–75. Three seed rates of S23 perennial ryegrass were factorially combined with three rates of Scots timothy and compared with pure stands of each grass. All were sown with Huia white clover. When cut six times at monthly intervals and with an annual N input of 350 kg ha^?1, there were no significant differences in total DM production in either year. The 2-year mean DM yield for the nine mixtures and six pure swards was 9·77 t ha^?1 (range 9·34–10·16). Compared with the pure ryegrass swards, in both years the ryegrass-timothy mixtures produced earlier spring growth but were significantly lower yielding at the second cut. Over the first five cuts the proptortion of timothy in the three mixtures with 22·4 kg ha^?1 ryegrass seed averaged 26% in the first year and 37% in the second. Corresponding calculated mean DM yields of timothy were 2·75 and 3·00 t ha^?1. It is concluded that an early timothy variety is capable of competing with a late-heading perennial ryegrass in frequently cut swards managed to simulate intensive grazing. The strong development of timothy in the dry summer of 1975 suggests that in mixtures of late perennial ryegrass varieties, an early variety of timothy should be beneficial for its spring growth in grazed swards.     

Establishment and production of common sainfoin (Onobrychis viciifolia Scop.) in the UK. 2. Effects of direct sowing and undersowing in spring barley on sainfoin and sainfoin-grass mixtures

The dry matter (DM) yield and herbage quality of swards of sainfoin ( Onobrychis viciifolia ), meadow fescue ( Festuca pratensis ,) and tetraploid perennial ryegrass ( Lolium perenne ) grown in monocultures and in four sainfoin:grass mixtures (0·33 sainfoin:0·66 meadow fescue, 0·66 sainfoin:0·33 meadow fescue, 0·33 sainfoin:0·66 perennial ryegrass and 0·66 sainfoin:0·33 perennial ryegrass), established by direct sowing or undersowing in spring barley, were investigated over 3 years in a field experiment in the UK. Direct sowing produced a mean yield across all species and mixtures of 1·8 t DM ha⁻¹ in the establishment year, whereas undersowing produced no measurable yield except for that of the spring barley. Undersowing reduced the yields of sainfoin and sainfoin-grass mixtures in the first full-harvest year but not in the second. The annual yield of a monoculture of sainfoin was 7·53 t DM ha⁻¹ and that of sainfoin-grass mixtures was 8·33 t DM ha⁻¹ averaged over 3 years. Both sainfoin and the sainfoin-grass mixtures had higher annual DM yields than the grass monocultures. The mixture of 0·66 sainfoin:0·33 meadow fescue gave the highest mean annual yield (9·07 t DM ha⁻¹) over the 3 years. There was a higher proportion of sainfoin maintained in mixtures with perennial ryegrass than with meadow fescue. The proportion of sainfoin in sainfoin–meadow fescue mixtures declined from 0·62 in the first year to 0·32 in the third year, whereas the proportion in sainfoin–perennial ryegrass increased from 0·48 in the first year to 0·67 in the second year and remained stable in the third year.     

Effects of white clover cultivar and companion grass on winter survival of seedlings in autumn-sown swards

The aim was to study the effects of white clover cultivar and combinations with perennial ryegrass cultivars on seedling establishment in autumn‐sown swards and on winter survival of seedlings. Large‐leaved white clover cv. Alice and small‐leaved white clover cv. Gwenda, and an erect and a prostrate perennial ryegrass cultivar were sown in autumn in pure stands and as four binary grass‐clover mixtures. Mixtures of white clover cv. Huia and Aberherald with perennial ryegrass were also sown. Companion grasses had no significant impact on the establishment of white clover. The number of seedlings of white clover cv. Alice in mixtures (335 m^?2) was higher than cv. Gwenda (183 m^?2) and pure swards had similar white clover population densities as mixed swards. White clover cv. Huia tended to have more seedlings than Aberherald (355 and 205 m^?2 respectively). No stolons were produced prior to a severe winter, because of the late sowing date. Winter survival of clover seedlings was 0·56 in mixtures and 0·69 in pure stands, irrespective of white clover or companion grass cultivar. Stolon development of white clover in autumn is often considered essential for overwintering survival and spring growth. In this study, there was considerable survival of the non‐stoloniferous tap‐rooted seedlings of all four clover cultivars despite a severe winter.     

Persistence of tall fescue in a subtropical environment: tiller survival over summer in response to flowering control and nitrogen supply

Enhancing pasture persistence is crucial to achieve more sustainable grass‐based animal production systems. Although it is known that persistence of perennial ryegrass is based on a high turnover of tillers during late spring and summer, little is known about other forage species, particularly in subtropical climates. To address this question, this study evaluated survival of grazed tall fescue tillers growing in a subtropical climate. We hypothesized that hard tactical grazing during winter to remove reproductive stems (designated as ‘flowering control’), and nitrogen fertilization in spring, would both improve tiller survival over summer, and thus enhance tiller density. This was assessed in two experiments. In both experiments, few tillers appeared during late spring and summer and so tiller density depended on the dynamics of vegetative tillers present in the sward in spring. In Experiment 2, flowering control and nitrogen fertilization both enhanced the survival of that critical tiller cohort, but the effects were not additive. Responses were similar but not statistically significant in Experiment 1, which had a warmer, drier summer and lower overall survival rates. Unlike grasses in temperate environments, persistence of tall fescue in this subtropical site appeared to follow a ‘vegetative pathway’; i.e., new tillers were produced largely in autumn, from vegetative tillers that survived the summer.     

Effects of pesticides and fertilizer on invertebrate populations of grass and wheat plots in Kent in relation to productivity and yield

At a site in Kent, perennial ryegrass (Lolium perenne) variety S24 and Italian ryegrass (L. multiflorum) variety RVP, wheat varieties Armada and Norman, and the original mixed grass ley were grown in small plots during 1982–84. Two toxic pesticides (phorate and aldicarb) were applied to half the total number of plots three times each year to eliminate soil invertebrate populations. Fertilizer was also applied to most plots. The yields of the crops, grown with and without pesticide, and the effects of the fertilizer were compared. Grass herbage yield was measured on three occasions during the summers of 1983 and 1984. Wheat grain yields were also determined in 1983. During the first year significant differences were not apparent in grass dry matter yield between pesticide-treated and non-treated plots, but significant differences were found in the second year. The perennial ryegrass was more susceptible to pest damage than the Italian ryegrass or the grass ley. Grass yields varied between cuts and in relation to variety and pesticide treatment, yields tending to be greater in untreated plots. Fertilizer treatment greatly increased grass dry matter yields, particularly with the Italian ryegrass. The effects of pesticide treatment on both wheat varieties varied although some yield enhancement was evident. Invertebrate animal populations in pesticide and fertilizer-treated plots were also assessed in autumn 1982, spring and autumn 1983 and spring 1984. In contrast to pesticide treatment, fertilizer treatment had little effect on soil invertebrate populations. Nematode populations were reduced at each sampling occasion by the pesticide treatment. Slug populations were initially unaffected but were subsequently reduced. Leatherjackets, by far the most abundant pest in both grass and wheat plots, were markedly affected by pesticides on all sampling occasions. Generally, fewer soil-dwelling dipterous larvae were recovered in spring than in autumn. Stem-boring dipterous larvae were virtually absent.     

The effect of height and frequency of mowing on the yield and composition of perennial ryegrass—white clover swards in the autumn to spring period

Ninety-six plots (3 × 2 m) of well-established perennial rye grass/white clover pasture were mown to heights of 2·7 (Low) or 3·96 (High) cm (rising plate meter) at 14-, 28-, 84- or 112-d intervals in autumn-winter. A 7-, 14- and 28-d mowing interval was superimposed in spring on each autumn–winter mowing interval treatment with the low and high mowing heights altered to 2·92 and 4·80 cm, respectively.
With the low cutting height, accumulated herbage DM was more than doubled (1806 ± 79 kg DM ha^-1) compared to a 'high' (754 ± 49 kg DM ha^-1) cutting height in autumn–winter and this was due to increased harvesting efficiency rather than growth as estimated by leaf extension. Although defoliation interval had no effect on DM yield, the grass component increased and clover decreased. The composition effect carried over into spring. On average, 3·5 tillers were produced over winter for each ryegrass tiller present in autumn and tiller densities were higher in spring. Tillers produced over autumn–winter contributed more than 60% of ryegrass growth by early spring.
In early spring (16–30 September), the low cutting height increased herbage DM yield, in mid-spring (1–14 October) it reduced DM yields particularly in combination with short defoliation intervals, while in late spring (14 October to 11 November) cutting height had no effect on DM yields.
Over the entire spring period there was a very marked effect of defoliation interval on DM yields.     

Tillering and leaf area index in grasses in the vegetative phase

The relationship between tillering of a vegetative grass stand and Leaf Area Index (LAI) has been studied during sward establishment of perennial ryegrass and Italian ryegrass with different seedling densities and nitrogen fertilizer rates and on established swards of tall fescue during winter and early spring after different dates of last cut in autumn.
The different experiments on different species led to the conclusion that in every instance the tillering rate slows down as soon as the LAI reaches a value of 3 and terminates rapidly at higher LAI, which corresponds with the almost complete extinction of light at the level of the tiller buds. The results confirm the direct role of light on morphogenesis previously demonstrated by several authors. The existence of such a limit of LAI helps to explain some contradictory effects of temperature or nitrogen on tillering.     

Sward structure changes within a perennial ryegrass swara: leaf appearance and death

The appearance of successive leaves on labelled tillers of S24 perennial ryegrass, and their subsequent death, were monitored for 1 year on cut plots on which four levels of N fertilizer were applied.
A new leaf appeared about once every 15 d from the end of April until the beginning of September, and about once every 52 d from the end of October to the middle of February. Rate of appearance changed rapidly during September and October, and again from mid-February until the end of April. The response of leaf appearance rate to temperature changed early in the year; the effect of overwintering is discussed.
Leaf lifespan was 50–60 d in summer and 120–140 d in autumn and early winter, and was influenced by conditions either during expansion of the leaf or shortly afterwards.
The results are discussed in relation to the construction of a model to describe changes in the herbage available to animals grazing a grass sward.     

Impact of deferred grazing and fertilizer on herbage production,soil seed reserve and nutritive value of native pastures in steep hill country of southern Australia

Developing sustainable grazing management systems based on perennial species is critical to preventing land degradation in marginal land classes. A field study was conducted from 2002 to 2006 to identify the impacts of deferred grazing (no defoliation of pastures for a period generally from spring to autumn) and fertilizer application on herbage accumulation, soil seed reserve and nutritive value in a hill pasture in western Victoria, Australia. Three deferred grazing strategies were used: short‐term deferred grazing (no defoliation between October and January), long‐term deferred grazing (no defoliation from October to the autumn break) and optimized deferred grazing (withholding time from grazing commenced between annual grass stem elongation and seed head emergence and concluded in February/March). These treatments were applied with two fertilizer levels (with or without fertilizer at 50 kg phosphorus ha^?1 and 2000 kg lime ha^?1 applied in year 1 only) in a factorial arrangement and two additional treatments: continuous grazing (CG) and no grazing (NG) in year 1. The deferred grazing treatments on average produced herbage dry matter of 4773 kg ha^?1, the NG produced 4583 kg ha^?1 and the CG produced 3183 kg ha^?1 in year 4 (2005–06) of the experiment. Deferred grazing treatments with and without fertilizer application produced an average of 5135 and 4411 kg DM ha^?1 respectively. Averaged over 4 years, deferred grazing increased the germinable seed pool of perennial grasses by 200% and annual grasses by 50% (except optimized deferred grazing that considerably decreased the annual grass seed pool) compared with the CG. The best of the deferred grazing strategies increased the digestibility of pastures by 7% compared with the CG. The results demonstrated that deferred grazing from spring to autumn followed by rotational grazing could be an effective tool to increase herbage production and soil seed pool and improve the digestibility of native pastures in the steep hill country of southern Australia.     

De-hardening in contrasting cultivars of timothy and perennial ryegrass during winter and spring

Scenarios of climate changes indicate longer and more frequent spells of mild weather during winter in northern latitudes. De-hardening in perennial grasses could increase the risk of frost kill. In this study, the resistance to de-hardening of different grass species and cultivars was examined, and whether the resistance changes during winter or between years, was tested. In Experiment 1, two cultivars of timothy ( Phleum pratense L.) and perennial ryegrass ( Lolium perenne L.) of contrasting winter hardiness were grown under ambient winter conditions, transferred from the field in January and April 2006 to the laboratory for 9 d with controlled de-hardening conditions of 3°C, 9°C and 15°C. The timothy cultivars were tested at 3°C, 6°C and 9°C in a similar experiment (Experiment 2) in January 2007. De-hardening, measured as decrease in frost tolerance (LT₅₀), was less in timothy than in perennial ryegrass and increased with increasing temperatures. The northern winter-hardy cultivar Engmo of timothy de-hardened more rapidly than the less-hardy cultivar Grindstad, but had higher initial frost tolerance in both experiments, whereas there was less difference between cultivars of perennial ryegrass in Experiment 1. Cultivar Grindstad of timothy lost all hardiness in early spring at all temperatures, whereas cultivar Engmo maintained some hardiness at 3°C. Cultivar Engmo de-hardened at a lower rate in 2007 than in 2006, in spite of similar frost tolerance at the start of de-hardening treatment in both years. This indicates that the rate of de-hardening was controlled by factors additional to the initial frost tolerance and that autumn weather conditions might be important for the resistance to de-hardening.     

Impact of deferred grazing and fertilizer on plant population density,ground cover and soil moisture of native pastures in steep hill country of southern Australia

The impact of deferred grazing (no defoliation of pastures for a period generally from spring to autumn) and fertilizer application on plant population density, ground cover and soil moisture in a hill pasture (annual grass dominated, with Australian native grasses being the major perennial species) were studied in a large‐scale field experiment from 2002 to 2006 in southern Australia. Three deferred grazing strategies were used: short‐term deferred grazing (no defoliation between October and January each year), long‐term deferred grazing (no defoliation from October to the autumn break, that is the first significant rainfall event of the winter growing season) and optimized deferred grazing (withholding time from grazing depends on morphological development of the plants). These treatments were applied with two fertilizer levels (nil fertilizer and 50 kg P ha^?1 plus lime) and two additional treatments [continuous grazing (control) and no grazing for year 1]. Deferred grazing increased (P < 0·05) perennial grass tiller density compared with the control. On average, the tiller density of the three deferred grazing treatments was 27–88% higher than the control. There was a negative (P < 0·01) relationship between perennial and annual grass tiller density. Fertilizer application increased (P < 0·05) legume plant density. The densities of annual grasses, legumes, onion grass (Romulea rosea) and broadleaf weeds varied between years, but perennial grass density and moss cover did not. The ground cover of the deferred grazing treatments in autumn was on average 27% higher than the control. Soil moisture differed between treatments at 15–30 cm depth, but not at 0–15 depth over autumn and winter. The results imply that deferred grazing can be an effective tool for rejuvenating degraded native pastures through increases in native grass tiller density and population and through improving farm productivity and sustainability.     

THE PRODUCTION OF WINTER GRASS FROM A ROW CROP OF COCKSFOOT AND LUCERNE

A crop of cocksfoot and lucerne in alternate rows was subjected to various autumn treatments for winter pasture. The yield of grass in winter, and the effect of autumn and winter treatments upon yield and composition in the following spring, are reported and discussed.
Maximum yield of winter grass was achieved by resting from an early date and applying nitrogenous fertilizer. The response to fertilizer applied in August and mid-September for winter grass production was as good as that obtained on ordinary swards at other times of the year. To ensure an adequate yield in any year the crop would need to be rested from early August and to receive about 70 lb N per acre.
The spring growth of cocksfoot was distinctly retarded by cutting twice in winter, i.e. in November and February. Cutting once only had no such adverse effect.
Both cocksfoot and lucerne persisted satisfactorily. The use of nitrogenous fertilizer to promote autumn growth of grass did not greatly reduce the overall productivity of the lucerne.
The two species were complementary in the sense that, in combination, they maintained a high aggregate yield of herbage under a range of weather conditions in midsummer and in winter.     

The impact of defoliation frequency and nitrogen fertilizer application in spring on summer survival of perennial ryegrass under grazing in subtropical Australia

Abstract This field study investigated the effect of timing of nitrogen (N) fertilizer application in spring on the survival of grazed perennial ryegrass (Lolium perenne cv. Dobson and Yatsyn) over summer in a subtropical environment. There were five N fertilizer treatments: no applied N, 46 kg N ha^?1 on 22 October or 22 November or 22 December, or on 22 October and again on 22 December. Water‐soluble carbohydrate (WSC) concentration of perennial ryegrass plants entering the summer was altered by varying defoliation frequency, with defoliation interval based on the number of leaves per tiller. Frequent defoliation was set at a regrowth level of one leaf per tiller and less frequent defoliation at a regrowth level of three leaves per tiller, over a total of two by three‐leaf per tiller regrowth periods. Application of N fertilizer was found to have no significant effect (P > 0·05) on survival of perennial ryegrass plants over summer. On the other hand, defoliation had a marked effect on perennial ryegrass persistence, with frequent defoliation decreasing ryegrass plant density (51 vs. 88 plants m^?2; P < 0·001) and increasing the density of tropical weed grasses (99 vs. 73 plants m^?2; P < 0·001) by autumn. Frequently defoliated plants had a lower stubble WSC content on a per plant basis than less frequently defoliated plants in spring (103 vs. 201 mg per plant; P < 0·001) and summer (59 vs. 101 mg per plant; P < 0·001). The lower WSC content was associated with a smaller root system in spring (1·50 vs. 2·14 g per plant; P < 0·001) and autumn (1·79 vs. 2·66 g per plant; P < 0·01), and this was reflected in 0·29 more plants being pulled from the soil by livestock between November 1996 and April 1997. Rhizoctonia fungus was associated with roots of pulled plants, but not with roots of seemingly healthy plants, indicating that this fungus may have a role in a weakened root system, which was more prone to sod pulling. Nitrogen applied in October and November resulted in a reduced WSC concentration, although the effect was restricted to 1 month after N application. The present study indicates that survival of perennial ryegrass plants over the summer in a subtropical region is prejudiced by frequent defoliation, which is associated with a lower WSC concentration and a shallower root system. Under grazing, sod pulling is a reflection of this weaker root system and contributes to plant mortality.