rtgamma {BART}  R Documentation 
Truncated Gamma draws are needed for the standard deviation of the random effects Gibbs conditional.
rtgamma(n, shape, rate, a)
n 
Number of samples. 
shape 
Sampling from a truncated Gamma where

rate 
This parameter is the inverse of the scale which is an alternative representation for the Gamma distribution. 
a 
The truncation point, i.e., 
Returns n
truncated Gamma, i.e., Gam(shape, rate)I(a, infinity)
.
Gentle J. (2013) Random number generation and Monte Carlo methods. Springer, New York, NY.
set.seed(12)
rtgamma(1, 3, 1, 4)
rtgamma(1, 3, 1, 4)
a=rtgamma(10000, 10, 2, 1)
mean(a)
min(a)