La réponse des arbres à l'augmentation des sécheresses liées au réchauffement climatique dépend de leur capacité d'adaptation, c’est-à-dire de la variation génétique et de l’héritabilité de caractères adaptatifs impliqués dans la résistance à la sécheresse. Dans le chapitre I, nous identifions des proxys facilement mesurables de caractères adaptatifs impliqués dans la résistance à la sécheresse en comparant la microdensité du bois d’individus morts et survivants après la sécheresse de 2003. Les variables les plus discriminantes sont les densités moyennes de segments de haute et basse densité, la proportion du segment de haute densité et le coefficient de variation du segment de haute densité. Les arbres survivants ont toujours une densité plus élevée et des profils généralement plus hétérogènes. Si ces traits sont suffisamment variables génétiquement et héritables, alors il est possible de sélectionner des arbres plus résistants à la sécheresse dans les populations d'amélioration et dans les peuplements forestiers destinés à être régénérées naturellement. Nos résultats suggèrent qu’une sélection naturelle directionnelle pour la densité du bois du douglas se produit dans des environnements plus ou moins limités en eau. Cette sélection agit dans des directions différentes selon les caractéristiques des pressions de sélections subies dans les régions étudiées. Dans le chapitre II, nous évaluons le potentiel d’adaptation à la sécheresse du douglas introduit en France. Ce potentiel d’adaptation dépend de la variation génétique et de l'héritabilité des caractères adaptatifs mis en évidence dans le chapitre I. Nous trouvons une grande variabilité des estimations d’héritabilité et d’AGCV entre variables, provenances, sites et, dans une faible mesure, entre cernes annuels. La plupart des variables possèdent des héritabilités élevées à assez élevées au moins pour certaines provenances dans certains sites. Certaines variables tendent à avoir des héritabilités et AGCV généralement plus élevées : ce sont plutôt des variables de la partie de faible densité du cerne. Certaines variables de résistance à la sécheresse possèdent à la fois une héritabilité et une AGCV élevées : ce sont de bons critères de sélection pour la résistance à la sécheresse en amélioration génétique ou en régénération naturelle. Les estimations différentes entre sites suggèrent que les estimations d’héritabilités augmentent avec le caractère favorable du milieu. Les estimations sont significativement différentes entre provenances, avec de forts effets d’interaction avec les sites. En revanche il y a peu ou pas de différences significatives entre cernes. Dans le chapitre III, nous avons tenu compte de la variation associée aux cernes de croissance pour étudier les relations entre les estimations annuelles de paramètres génétiques et des variables climatiques et édaphiques. Les estimations d'héritabilité et de variation génétique de la plupart des variables de microdensité sont corrélées significativement avec la plupart des variables environnementales testées. De rares variables n’ont montré aucune corrélation significative dans aucun cas. Les caractéristiques des relations significatives sont variables entre les caractères, les sites et les provenances. Les prédicteurs climatiques les plus importants sont la température, l'évapotranspiration, la réserve en eau du sol et le déficit en eau. Les précipitations affectent peu l´estimation des paramètres génétiques. De façon générale, meilleures sont les conditions de croissance, plus élevées sont les estimations. Toutes les composantes des essais expérimentaux modifient les estimations des paramètres génétiques. Certaines peuvent être fortement déterminées par le choix du matériel végétal et du site. D'autres, comme la variation climatique temporelle, sont moins contrôlées et peuvent affecter de façon plus ou moins aléatoire les estimations. / Forest response to the drought increase associated to the climatic warming relies on tree adaptive potential, i.e. the genetic variation and the heritability of adaptive traits involved in resistance to drought. In the first chapter, we identify easy-to-measure proxies of adaptive traits for resistance to drought. We compare the wood microdensity of dead and surviving trees after the 2003 heat wave in France. The most discriminating variables are the mean density of high and lowdensity segments, high-density proportion and coefficient of variation of the lowdensity segment. The wood of the surviving trees is always denser and more heterogeneous. If these adaptive traits are variable and heritable, then it is possible to select for improved resistance to drought in the breeding population as well as in natural regeneration. Our results also suggest that directional selection is going on in more or less water-stressed environments. The direction is variable according to the nature of the selection pressure in the different regions. In the Chapter II, we estimate the evolutionary potential to drought of the introduced Douglas-fir in France. This evolutionary potential relies on the magnitude of the genetic variation and of the heritability of the adaptive traits found in the first chapter. The heritability and the genetic variation are highly variable between provenances, sites and, to a much lower extent, between annual rings. Most variables have moderate to high heritability estimates for at least some provenances in some sites. Some traits tend to have generally higher heritability and genetic variation estimates. These are mostly variables of the density part of the annual ring. The variables having at the same time relatively high estimates of heritability and genetic variation are good candidates for becoming efficient selection traits for resistance to drought in tree breeding as well as in natural regeneration. The significant between-site variation suggests that the heritability estimates increase with site quality. The estimates are also significantly different between provenances with a strong provenance × site interaction. Conversely there is little significant between annual-ring variation. The chapter III takes advantage of the annual-ring variation to study the relationships between the genetic parameter estimates and climatic and soil variables. The heritability and genetic variation estimates of most variables significantly relates with most tested environmental variables. Very few variables never correlates with any environmental variable. The significant relationships are very variable between traits, provenances and sites. The most important predictors are temperature, evapotranspiration, and soil water reserve and water deficit. Rainfall marginally influences the genetic parameter estimates. Generally, the better the growing conditions, the higher the estimates. All components of the experimental trials affect the genetic parameters estimates. Thus, the choice of the plant material and of the experimental site strongly determines the genetic parameter estimates. The uncontrolled climatic variation may randomly affect the estimates.

xxi, 299 leaves : ill. ; 25 cm / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.1976) from the Dept. of Plant Physiology, University of Adelaide

The Australian wheat production environments are typically water-limited, and both temperature and vapour pressure deficit increase as the season progresses. As a result, high incidences of small or shriveled wheat kernels (screenings) are commonly generated and can substantially reduce grain value. Previous studies suggest the incorporation of the tiller inhibition (tin) gene can reduce the production of infertile tillers and increase kernel weight (KW). It was hypothesised that the incorporation of the tin gene into wheat germplasm may a) contribute to the maintenance of large KW and reduction in screenings (SCR) in terminal water deficit environments; and b) not be associated with a grain yield (GY) penalty in terminal water deficit environments. Thus, the major objective of this thesis was to evaluate the expression and performance of tin gene in terms of GY and SCR: 1) in different genetic backgrounds and across Australian production environments; 2) in various northern production environments which are particularly prone to terminal water deficit conditions; and 3) to determine the mechanisms that contribute to the maintenance of large KW of tin lines in terminal stress conditions. To address the overall objective, populations of lines were genotyped for the presence/absence of the tin gene and were field tested. Line differences in GY, yield components, SCR and general growth and development attributes were determined in 22 field experiments conducted between 2005-2007. The experiments were grouped into those that evaluated: a large number of sister lines from four genetic backgrounds in multi-location experiments; selected lines from Silverstar population in multi-location experiments; and selected sister lines in detailed agronomic experiments examining the effect of plant density and controlled levels of water supply through the use of a rainout shelter facility. The effect of tin on GY and SCR varied with environment and genetic background. In the Brookton, Wyalkatchem and Chara background, there was no reduction in GY associated with tin in southern production environments. However, a 31% and 10%, advantage of free-tillering over reduced tillering Silverstar lines existed in the 2005 western and 2006 northern experiments respectively, and led to an average 12% reduction in GY of Silverstar tin lines. In northern experiments, tin lines in a Silverstar background produced up to 50% fewer SCR than Silverstar free-tillering lines. Averaged across experiments, KW of Silverstar tin lines was 10% greater than free-tillering lines. Based on stem number per plant, Silverstar lines were classified into three groups; the restricted (R) and semi-restricted (SR) tin and free-tillering lines attained 2.9, 3.4 and 4.8 stems per plant respectively. Expression of tin in terms of maximum stem number production was genetic background and genotype dependent, and unlike free-tillering lines, R tin lines in particular, were not as responsive to plant density. Head number per unit area rather than kernel number per head was strongly associated with KW determination. Therefore, to maximize KW under water limiting conditions it is more beneficial if high kernel number can be achieved via the production of low head number with more kernels per head, as can be achieved with the use of tin lines. The KW advantage of Silverstar R tin was associated with greater anthesis total dry matter, stem water soluble carbohydrate and nitrogen available per head relative to free-tillering lines, and thus R tin lines had more assimilate for translocation during the grain filling period. In a terminal water deficit experiment, individual KW data collected for Silverstar tin and free-tillering lines revealed that KW of tin lines (≈ 25 mg per kernel) was maintained for main stem to fourth tiller heads and across floret positions 1-4 within spikelets. In contrast, free-tillering lines (≈ 18 mg per kernel) generated small kernels across the entire plant, with the largest proportion originating from floret positions 3 and 4. Lines containing the tin gene had a greater mean KW and kernel width, and a higher frequency of wider kernels than free-tillering lines. The high frequency of large kernel widths was associated with significantly less SCR in tin lines. A strong positive association between maturity head number per plant and SCR indicated, for every fertile head produced, SCR increased by 11% in the terminal water deficit experiment. Reduction in SCR in the Silverstar tin material in the north, was associated with high KW and a tendency for lower GY, although tin lines with equivalent GY to Silverstar could be identified in each environment. The incorporation of the tin gene has the potential to significantly reduce the incidence of SCR in commercial wheat crops. The reduction in GY associated with the tin gene was dependent on genetic background, suggesting the potential for selection of higher-yielding tin progeny in commercial line development. However, a tailored agronomic package to optimize yield potential of tin lines needs to be developed for different target environments. With the incorporation of the tin gene into genetic material adapted to the northern wheat belt and optimisation of head density, there exists scope for simultaneous improvements in GY and KW and subsequent reduction in SCR for terminal water deficit environments.

Bibliography: leaves 98-123. This thesis investigates the symptoms of leaf senescence in response to plant shoot water stress and demonstrates the effect of a non-hydraulic root signal in the senescence response of mature leaves of sunflower. The alleviation of the leaf soluble protein loss rate by excision of the root system in drying soil indicates that this signal originates in roots in dry soil and acts to promote protein loss.

Bibliography: leaves 201-228. Widespread deficiencies of mineral nutrients in soils along with limited moisture supply are considered major environmental stresses leading to yield losses in chickpea. This study was conducted to determine the zinc requirement of chickpea and the effect on plant water relations. Critical zinc concentration was estimated. It was found that high and low moisture regimes had no effect on critical zinc concentration and that the value remained almost the same in all chickpea genotypes. Sensitivity of 28 chickpea genotypes were evaluated at two zinc levels. Field studies on zinc fertilization in both Australia and Pakistan showed that the application of zinc increased grain yield in all chickpea genotypes. It was found that plants grown under zinc deficiency could not exploit available soil moisture and water use and water use efficiency was reduced, and concluded that high zinc availability may enhance the ability of plants to endure periods of drought by promoting osmotic adjustment.

Drought resistance among genotypes of Port-Orford-cedar (Chamaecyparis lawsoniana (A. Murr.) Pan., Cupressaceae) seedlings was evaluated both in the field and in the greenhouse. Field water potentials (��) of 5-year-old seedlings were measured at two high-elevation plantation sites where summer drought occurs. Measurements of �� were compared to survival two years prior at the same site. The north coastal breeding zone 1 had significantly lower mid-day �� and lower survival than the southern interior breeding zone 6. Percent survival at 3 years was significantly correlated with mid-day ��. A second assessment of drought resistance was performed in the greenhouse on 1-0 seedlings. Root growth potential (RGP) was measured in the winter under non-stress conditions and following a drought in the summer. Differences among breeding zones were opposite in pattern from outplanting measurements of survival and water potential; therefore, RGP may not be a good predictor of drought resistance or survival among different genotypes within a single species. Low-elevation, coastal families had greater root growth than inland, high elevation sources. Greater root growth occurred in seedlings with more shoot mass. Little regional specialization in RGP across the species' range was indicated as there was greater variation among families than among breeding zones. In the summer RGP test following drought, only the two extremes of the range were evaluated; both north coastal and southern interior families showed decreased root growth compared to the winter RGP under non-stress conditions. Despite significantly higher predawn �� in inland, high elevation families, lower elevation coastal families had significantly more roots. Change in chlorophyll fluorescence yield measured on foliage of droughted plants was positively correlated with the absolute value of predawn ��; however, it was not a sensitive predictor of predawn �� (R��=0.06) at the levels used in this study. Lower levels of �� may be necessary to produce severe stress to damage Port-Orford-cedar foliage. / Graduation date: 2003

Geohumus is a new, unique hybrid soil amendment made of polyacrylic acid, igneous rock flour, and quartz sand. It is biodegradable and is marketed to increase soil porosity, increase plant yield, and increase plant available water by up to 30%, all while using a low inclusion rate of only 1 - 2% volume/volume (v/v). The objective of this research was to evaluate uses of Geohumus in a turfgrass setting. Two field trials and one greenhouse trial were conducted between the summers of 2010, 2011, and 2012. For the field incorporation study, Geohumus was mixed with USGA (United States Golf Association) specified sand and incorporated at a 1% v/v ratio into an existing perennial ryegrass stand maintained at two heights of cut using various incorporation machinery. Treatments included incorporation of Geohumus via an aerator, dethatcher, Graden, or simulated drill-and-fill. Daily light box pictures were analyzed using SigmaScan Pro software to compare percent green cover among treatments and to establish which plots were to receive irrigation. The readings showed significant differences among treatments in irrigation applications required to keep the turf green. The aeration treatment by itself was as effective as any treatment with Geohumus incorporation. Aeration provided a 35% reduction in irrigation applications on the 61 mm height of cut in 2011 and a 24% reduction in irrigation applications on the same height in 2012. The second field trial evaluated various soil amendments for use in divot repair mixes under full sun and partial shade conditions. In addition to differing microclimates, divots were further evaluated by receiving either one or two irrigation applications per day. Uniform divots with dimensions of 10.16 x 7.62 cm were created using a spring-loaded catapult with a hula-hoe attachment. Treatments applied consisted of combinations of United States Golf Association (USGA) specified sand and perennial ryegrass seed but differed according to soil amendments which were incorporated into the divot mixes based on volume (v/v). Soil amendments tested included Geohumus (1% v/v), Zeba (1% v/v), Axis (10% v/v), Lassenite ATS (10% v/v), Dakota Peat (10% v/v), Scotts EZ Seed (10% v/v), and Wondersoil (10% v/v). Weekly ratings were taken using light box pictures in conjunction with digital analysis software to measure percent green cover. Results indicated that microclimate and irrigation frequency affected the establishment of divot repair mixes more than the presence of soil amendments. A third trial comparing perennial ryegrass root masses grown in rooting boxes in a greenhouse showed major differences as influenced by placement of Geohumus within the soil profile. Geohumus influenced root mass in the upper 7.6 cm of the soil profile but no significant differences were found at lower depths. Geohumus incorporated at a 7.6 - 12.7 cm depth appeared to be the most influential depth for increasing root mass in the upper 7.6 cm of the profile. / Graduation date: 2013

Thesis (Ph.D. (Botany))--North-West University, Potchefstroom Campus, 2008.

Hemiepiphytes are important components of tropical forests and are attractive to scientists due to their unique epiphytic growth habit during some period of their life cycle. Unique characteristics in plant water relations and carbon economy have been found in hemiepiphytic plants; however, to further understand this group of species on an evolutionary basis it is necessary to carry out comparative studies between hemiepiphytes and their close relatives. In this dissertation I conduced a comparative study in a suite of functional traits related to plant water relations and photosynthesis between hemiepiphytic and non-hemiepiphytic tree species from a single genus-Ficus. Great differentiation in functional traits has been found between species of the two growth forms both during juvenile and adult stages. Seedlings of hemiepiphytic Ficus species (H) had significantly lower xylem hydraulic conductivity, stomatal conductance, net light saturated CO2 assimilation, and higher water use efficiency than congeneric non-hemiepiphytic species (NH), which are adaptive to a drought-prone epiphytic growth conditions under natural conditions. The conservative water use adaptation in H species is likely crucial to the drought tolerance and survival in the forest canopy but is related to much lower growth rates than NH species. Species of the two growth forms both showed relatively large plasticity in responding to variation in light level as in typical light-demanding species. Surprisingly, the NH species showed characteristics related to higher light demand than H species, which is opposite from the prediction that H species are more light-demanding than NH species. Thus, although commonly accepted, it is likely that light was not the selective pressure for the evolution of hemiepiphytism in Ficus. Using adult trees grown in a common garden, I found that H species showed characteristics of more conservative water use even after they established connections to the soil. Moreover, H species showed significantly different traits in photochemistry compared to NH species due to hydraulic-photosynthetic coordination. The evolution of an epiphytic growth habit during the juvenile stage of a life cycle in the hemiepiphytic Ficus species thus involved changes in a suite of functional traits that persist during their terrestrial growth stages.

A stochastic, compartmental Model for Adaptive Livestock Management (MALM) was developed for cow-calf enterprise for Rolling Plains of Texas from an existing model, Simple Ecological Sustainability Simulator (SESS). The model simulates forage and animal production. It runs on a monthly time step. Two stocking strategies, flexible and fixed, were evaluated at seven stocking levels for effects on forage and animal production, range condition, and net ranch income. Evaluation data were obtained from published and unpublished data from Texas A&M Agricultural Experimental Station at Vernon for Throckmorton. The model adequately simulated forage and animal production. Light fixed stocking rates and flexible stocking strategies resulted in cows of median body condition score (BCS) 5, compared to low BCS of 4 under moderate fixed stocking rate, and BCS of 3 under heavy fixed stocking. BCS declined from autumn to early spring and peaked in summer. Cows under light fixed stocking rates and under flexible stocking were heavier (460 kg) compared to those under heavy fixed stocking (439 kg). Replacement rates were lower under light stocking (22 %), compared to flexible (37 %) and heavy stocking (56 %). Calf crops were all above the reported 90 % expected for bred heifers because of the replacement policy. Flexible stocking strategy resulted in higher net income ($19.62 ha-1), compared to fixed light ($5.93 ha-1) or fixed heavy ($-17.35 ha-1) stocking strategies. Coefficient of variation (CV) in net income was highest under heavy stocking (90%) compared to light stocking (60%) and flexible stocking (50%). Maximum net income was obtained between 0.05 AUM·ha-1 and 0.13 AUM·ha-1 when fixed stocking strategy was used but when flexible stocking strategy was used maximum net income was obtained between 0.1 AUM·ha-1 and 0.17 AUM·ha-1. Range condition rapidly declined under fixed heavy stocking, increased under fixed and light flexible stocking, and remained constant under moderate flexible stocking. Heavy fixed stocking decreased range condition rapidly over a 20-year period. MALM was an effective tool to demonstrate effects of different management strategies. The model can function as a strategic or a tactical decision aid. It is concluded that there is potential for this model to assist managers in improving the sustainability of agriculture.