Materials and methods 

Study area 

The study was conducted at the Arid Forest Research Institute, Jodhpur, which is a premier research institute in the field of forestry in arid hot region of India。 This region forms part of Indian Thar Desert。 Soils had low organic matter content (0。27%) and available P (10。2 mg⋅kg-1 soil), and pH of 7。8 (Singh et al。 2007)。 Microclimates over the study area (20。82 ha) can be assumed as uniform considering the small size of the area。 Landscaping of the campus area was carried out by planting various kinds of ornamental plant species and turfgrass during 2005−2006。 The choice of these species was broadly based on professional experience of the authors after observing their performance in arid and semi arid regions of India。 An area of about 10,000 m2 was developed as grass lawn around office buildings and the residential complex of the institute。 Ornamental plant species were planted in lawn and open areas of the campus, including Alstonia scholaris L。R。Br。, Bougainvillea glabra, Ervatamia paricata (L。) Burkill, Hibiscus rosa-chinensis Linn。, Nerium odorum Sol。 ex Aiton, Tecoma stans (L。) Juss。 ex Kunth, Ficus benjamina L。‘black’, Ficus benjamina L。‘exotica’, Ficus benjamina L。 ‘starlight’, Ficus panda L。, Ficus retusa L。, Thuja compacta Stand。 ex Gordon, Polyalthia longifolia Sonn。, P。longifolia var。 pendula, Phoenix canariensis Chabaud, Areca catechu L。, Codiaeum variegatum (L。) A。 Juss。, Hyophorbe lagenicaulis (L。 Bailey) H。E。Moore, Caryota urens L。, Murraya paniculata (L。) Jack, Nyctanthes Arbor-tristis L。, Duranta repens L。 ‘Golden form’, Clerodendrom inerme (L。) Gaertn。, and Tecomaria capensis (Thunb。) Lindl。

Experimental design 

Spreading of turfgrass into plant pits was noticed between 15 to 30 days and one plant pit of three feet diameter was covered with turfgrass within 45−50 days。 In order to estimate the optimal tilling interval of soil in plant pits, an experiment was initiated in 2006。 The experiment was designed with three factors i。e。 three soil tilling intervals (15 days, 30 days and 45 days), two plantation models (inside lawn area and outside lawn area) and six plant species, namely Alstonia scholaris, Bougainvillea glabra, Ervatamia paricata, Hibiscus rosa-chinensis, Nerium odorum and Tecoma stans。 Fifteen plants of each species were studied for both areas separately。 Five plants of each species were selected at random for different soil tilling intervals。 In total, there were 180 plants of all six species for the study。 In lawn areas, irrigation of both plants and grass was supplied through one hour operation of sprinklers。 Each plant received 3 L of water on average per day。 Plants were irrigated daily in summer and on alternate days in winters。 Plants outside lawn areas were supplied 30 L of water in summer and 15 L in winter at intervals of 10 days。 Total irrigation was 90 L per plant per month in summer and 45 L per plant per month in winter in both models。 Plants outside lawn areas were irrigated manually。 No irrigation was given to the plants and grasses during the rainy season。 Soil in the plant pits on both areas was tilled to 15。24 cm depths to remove grass and weeds over a 0。91。44 cm diameter。 These practices were continued until 2009。 Initial height data were recorded for all plant species in 2006 before carrying out systematic soil tilling。 Initial crown diameter and stem diameter were not recorded as these were too small to be measured。 Height and crown diameter data were recorded after three years in 2009。 

Data analysis 

General Linear Model (GLM) factorial analysis was performed using Statistical Package for Social Sciences (SPSS, 8。0) software for height and crown diameter of six selected plant species。 The analysis was carried out considering plant species (spp。), soil tilling intervals (STI) and plantation models (PM) as independent factors, and height & crown diameter as dependent variables under the null hypothesis that plant height and crown diameter were not affected significantly by these factors and their interactions。 All the significant main effects and their interactions were studied。 To compare the means for main effects and interactions, we used the Duncan Multiple Range Test (DMRT)。 Initial height of plants was taken as a covariate presuming that variations in height at planting would not influence the results。 Covariates were used to refine the experimental results and reduce the error of the experiment (Dospekhov 1984)。