populate vector (sampling from various rnorms w/various x=, sd=,) conditionally based on two other vectors

i need to create a vector (CBD) and populate it using 12 different normal distributions with various means and so's.

for example:

for all cells that treatment == "wellwaterd" & plantpart =="flower" I would pull 15 samples from a vector x<-rnorm(100,mean=3.6, sd=0.3). I would then pull another 15 samples from y<-rnorm(100,mean=2.2, sd=0.2) to populate treatment=="wellwaterd" & plantpart=="leaves" I would pull another 15 samples from z<-rnorm(100,mean=2.0, sd=0.8) to populate treatment=="wellwaterd" & plantpart=="stems" and so on until all CBD cells were filled based on treatment and plantpart then i need to do the same for another vector named CBDA and do the same process.

rm(list=ls())
plantID<-c(seq(1,45,1))
plantID<-rep(1:45,each=12)
strain<-rep(c("CHPF-01"), 540)        
length(strain)
head(strain)
treatment<-rep(c("wellwatered","drought","waterlogged"), each=180)
treatment
plantpart<-rep(c("flower", "leaves", "stem", "seed"),135)
location<-rep(c("high", "mid", "low"), each=180)

CBD<-vector("numeric", length=540)
data<-cbind(plantID, strain, treatment, plantpart, location, CBD

x<-rnorm(100,mean=3.6, sd=0.3)

My attempts:

if (treatment=="wellwatered" & plantpart=="flower")
  CBD<-sample(c(x, 15, replace = TRUE))

while (treatment=="wellwatered" & plantpart=="flower") {
  CBD<-sample(c(x, 15, replace = TRUE))
}  

cbd<-function(n) {
  when treatment=="wellwatered"|plantpart=="flower"
  sample(x, 15, replace = TRUE)
}