Water Costs and Water Revenues for Established and Rain-Induced Roots of Agave deserti

1 Production of new roots leads to increased water uptake and increased respiration. Water uptake can be considered as the revenue from root production, and root respiration divided by the wateruse-efficiency as the water cost. Using an analogy to economics, the optimal root mass occurs when the marginal cost of root production equals the marginal revenue of root production. 2 Agave deserti Engelm. has two types of roots, established and rain-induced; the latter were shed three times as the soil dried during the year considered. Per unit dry mass, respiration required for growth was 90-fold higher and respiration required for maintaining living tissue was 40-fold higher for rain roots compared with established roots. Conductance to water flow per unit dry mass was 11-fold higher for rain roots compared with established roots. 3 Previous excavations showed that A. deserti can have an established-root length of 44 ? 8 m (mean ? SD) and a rain-root length of 13 ? 3m. Optimal root lengths predicted for this plant were 51 m of established roots and 31 m of rain roots, which were within 1 SD of the measured lengths for established roots but greater than 6 SD for rain roots. Thus, water may be the currency for the production of established roots but not for rain roots. Key-words: Respiration, root growth, root length, water relations * Present address: Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109, USA. Introduction Perennial plants continually re-explore the adjacent soil by producing new roots, especially shortlived fine roots, which may increase water and nutrient uptake (Caldwell, 1976, 1979). Optimization models using analogies to economic theory are useful for generating predictions of plant response to the environment (Bloom, Chapin & Mooney, 1985; Givnish, 1986). Bloom, Chapin & Mooney (1985) predicted that plants will continue to produce roots until the marginal cost of root production equals the marginal revenue. To compare costs and revenues, the same units or currency must be used, either carbon, water or nutrients (Bloom, Chapin & Mooney, 1985); the revenue from root production can be increased water uptake, and the cost can be defined as the water loss from transpiration associated with the photosynthetic CO2 uptake that is required to provide the carbon for growth and maintenance respiration of roots. For the Sonoran Desert perennial, Agave deserti Engelm. (Agavaceae), which has succulent leaves, water is the main limiting factor for shoot growth (Nobel, 1976, 1984). Moreover, A. deserti has both established roots and rain-induced roots, the latter being finer, produced on established roots within 12 h of a rainfall and generally shed as the soil dries (Nobel & Sanderson, 1984; Jordan & Nobel, 1984). In this study, respiration rates of established roots and rain roots were measured to determine the carbon requirement and hence the water cost for root growth. Water revenue was estimated from a water uptake model for desert succulents (Hunt & Nobel, 1987a, b) to determine the optimal root mass at which the marginal water revenue equals the marginal water cost. Materials and methods Plants of Agave deserti were started from seed collected in the field at the Philip L. Boyd Deep Canyon Desert Research Center, located near Palm Desert, California. Three-year-old plants were transplanted into aerated hydroponic solutions (Nobel & Hartsock, 1986) or porous-membrane This content downloaded from 157.55.39.120 on Mon, 05 Sep 2016 05:16:32 UTC All use subject to http://about.jstor.org/terms 126 culture (Brown & Haq, 1984) and placed for 3 E. R. Hunt et al. months in a controlled environment chamber with 12-h days and 12-h nights, day:night air temperatures of 25?C:18'C, day:night relative humidities of 40%:60% and photosynthetically active radiation of 600 [mol m-2 s-1. Porousmembrane culture involved placing the complete root system into an envelope (100 mm X 100 mm) of Versapor-3000 (Gelman Sciences, Ann Arbor, Michigan), which is a nylon-acrylic copolymer sheet with an average pore size of 3 [m and a high hydraulic conductivity of 16 [m s-1 Pa-1. The envelopes containing the root systems were placed into 250-mm diameter pots filled with sand and watered twice weekly with 0 1-strength Hoagland solution number 1 (Hoagland & Arnon, 1950). Relative growth rate (RGR, g kg-' day-') of both established and rain roots was defined as: RGR = (InW2 InW1)/AT Equation 1 where W is the dry root mass (kg) based on the lengths of established roots and rain roots and AT is a time interval of 30 days (an RGR of log kg-' day-' is equal to 1% day-'). After relative growth rates were determined, the complete root system was surface sterilized in a 5 kg m-3 NaOCl solution for 1 min, rinsed with water and placed in a 0 333 x 10-3 m3 cuvette filled with 0 02-strength Hoagland solution and maintained at 250C. A Clarktype oxygen electrode (type YSI-4004, Yellow Springs Instrument Co., Yellow Springs, Ohio) was used to measure total-root respiration for 15-min periods. Then the rain roots were excised and the respiration rate of established roots was determined; rain-root respiration was defined as total-root respiration minus established-root respiration. Respiration rate (R, [Lmol kg-' s'1) was expressed on a dry mass basis. Because starch is the primary energy storage compound of A. deserti (R. F. Denison, D. T. Tissue & P. S. Nobel, unpublished), glucose was assumed to be the substrate for respiration so that 1 mol of CO2 would be produced for every mol of 02 consumed during respiration. A two-dimensional, radially symmetric model of water uptake by the entire root system (Hunt & Nobel, 1987a, b), developed from a one-dimensional soil water model (Young & Nobel, 1986), was used to predict the amount of water taken up from the soil over a given time period. The ground area around the plant was divided into a central core and eight concentric rings; the ground area of the innermost five rings and core was set equal to the ground area occupied by its roots. Soil depth was divided into two 25-mm-thick upper layers and nine 50-mm-thick lower layers. For 1-h intervals, water movement between soil layers was calculated using Darcy's law and water evaporation from the uppermost 25-mm-thick soil layer was calculated using Fick's law. The year 1984 was used for microclimatic inputs (total precipitation was about average, 159mm); a similar year (1985, in which precipitation was 164mm) was used to validate the model (Hunt & Nobel, 1987a). Root systems of A. deserti at the field site were excavated in February and March 1985 to determine root length distributions; the plant used for validation of the water uptake model in the field had 42m of established roots, 12m of rain roots, and an explored ground area of 0.32 m2 (Hunt & Nobel, 1987a). The volumetric flux density (JVroot, m s-) of water into the roots was calculated for 1-h inter-