Running samples of 30 contiguous soil blocks, each 15 cm. long, were taken along 

 one side of each wire. The blocks were 7.5 cm. wide and 7.5 cm. thick and were taken 

 at three depths: to 7.5 cm., 7.5 to 15.0 cm., and 15.0 to 22.5 cm. Roots and other 

 organic material were separated from the soil by flotation, skimming, and decanting. 

 The mixture of water, soil, and roots (and underground litter) was first stirred to 

 permit separation of solid parts. The rapidly settling sand carried some of the lighter 

 material to the bottom; so the process was repeated until virtually all organic material 

 had been removed. Fecal matter and long-dead, partly decomposed plant parts waterlogged 

 readily and lost buoyancy, but such material was recovered in the repeated decanting. 



The samples of organic matter were rinsed over the sieve to remove any silt still 

 attached, then ovendried, and weighed. The material was separated into underground 

 shoot parts of grasses, coarse roots (>2.0 mm. in diameter), medium roots (0.5 to 2 mm. 

 in diameter), and "fine roots and other material," and the four parts were weighed. A 

 small portion of the "fine roots and other material" was sorted into fine roots (<0.5 

 mm. in diameter) and litter by means of a teasing needle and a small brush. Weights of 

 these parts were used to calculate the weights of the respective fractions in the orig- 

 inal sampling units. The material took on moisture during separation, so the sorted 

 weights were readjusted to ovendry weights. Grass bases comprised only 1.25 percent 

 of the weight of organic matter in the upper soil layer; their weight is included 

 with that of coarse roots in the presentation of results. 



The coarse roots were the short taproots of shrubs (only four were encountered in 

 the 1,350- by 7.5-cm. area of the sample) and first-order branch roots. These root 

 types had a pronounced taper. Tlie larger grass roots, essentially cylindrical, mostly 

 fell near the lower limit of the medium-root class. Taproots of annuals, which taper, 

 also were mainly in the middle-sized class. When dry, all feeder roots were in the 

 fine-root class; the fine roots were characteristically much branched, with frequent 

 angular axial deviation. Most were fragile and presumably dead when the sample was 

 excavated . 



The productivity data we present are calculated from estimates of annual yields of 

 total green growth (herbaceous material, and the leaves, fruits, and twigs of shrubs) 

 reported by Hutchings and Stewart (1953) for 20 large pastures lying 1 to 8 km. from our 

 sampling site. Yields were estimated by the method of Pechanec and Pickford (1937). 

 As will be brought out in the discussion, these data typify productivity of a wide area 

 of similar climate, soils, and vegetation, including the site of our study. 



The underground sampling layers (7.5 cm. thick) are given Roman numeric designations 

 for the purpose of discussion. Layer I is the first below the surface. Organic mass 

 below the ground is expressed as g./m.^ for each 7.5-cm. increment, regardless of the 

 area of original sampling unit. 



5 



