from the tall- and the short-height zones of seven locations when the grass 

 was mature. Since -the total amount of nutrients incorporated in the plant 

 tissue is highly dependent on yields, the dry weight is a good predictor 

 for standing crops of the mineral nutrients. The R 2 values for standing 

 crop of nitrogen, phosphorus, potassium, sulfur, calcium, magnesium, and 

 sodium are quite high (Tab. 33), and these equations should be reliable 

 and useful for estimating the amount of nutrients contained in the mature 

 shoots of S. alterniflora salt marshes. However, the fate of the nutrients 

 as the plants decompose is more difficult to determine. Because of greater 

 variation, the predictions for iron, zinc, manganese and copper standing 

 crops would be less accurate. 



g. Summary . The results indicate that tissue concentrations of 

 several nutrients and several soil properties were significantly associated 

 with variations in yield and height of S. alterniflova. It is important 

 to keep in mind the purpose of multiple regression analyses. Predictive 

 models are not necessarily functional, but can lead to insight into a 

 problem. According to Draper and Smith (1966) , construction of this type 

 of model from problems where much intercorrelation of data exists is where 

 regression techniques can make their greatest contribution. It provides 

 guidelines for further investigation, pinpoints important variables, and 

 is useful in screening variables. 



Several variables selected by the multiple regression procedure in 

 this study seem to warrant further investigation to determine their 

 relationship to productivity of S. alterniflova. Such variables negatively 

 associated with yield include salinity of the soil solution, manganese 

 concentration in the plants and soil, and sulfur concentrations in the 

 plants. Important variables positively associated with yield include 

 phosphorus concentration in the plant tissue and in the soil. 



The reduction of yield of S. alterniflora with increasing soil salinity 

 has long been recognized. This study reconfirms this, but there was not 

 a striking difference in salinity between height zones. This is indicated 

 by the fact that the simple correlation of height with soil salinity is 

 near (Tab. 27) . 



The importance of manganese and sulfur in both the yield and height 

 equations and iron in the height equation suggests investigation into the 

 influence of the chemical effects of waterlogging of the marsh soil on 

 S. alterniflora growth. Undoubtedly there are different degrees of 

 aeration both within and between Spartina marshes which affect soil chemi- 

 cal properties. There is an extensive literature on waterlogged soils in 

 connection with rice culture (Redman and Patrick, 1965; Black, 1968) . 

 Useful reviews on the chemistry of phosphorus and nitrogen in sediment- 

 water systems are presented by Syers, Harris, and Armstrong (1973) and 

 Keeney (1973) . 



The positive influence of phosphorus concentration in the plant tissue 

 would be expected, since in most natural plant-soil systems phosphorus is 

 second only to nitrogen as a limiting factor in plant growth. Nitrogen 



122 



