nls error says "singular gradient"

#1
I cannot get this code to work and I'm sure there is a very simple solution

i<-seq(1,length(rhino[,2])-1)
j<-seq(2,length(rhino[,2]),1)
Increase=c((Population[j]-Population)/Population)
Years<-c(Year[j]-Year)
Years2<-c(0,Population[j]-1960) # Cumulative Years

R.est<-Increase/Years # Estimate "r"
R.av<-mean(R.est)


Rhi.mod <- nls(Population ~ No*exp(r*Years2),
start=list(No=2230, r=-0.1587006), data=rhino)

It all works up to the very last line where it says "Error in nls(Population ~ No * exp(r * Years2), start = list(No = 2230, : singular gradient"

Here is the data:
Year Population
1 1960 2230
2 1971 650
3 1976 500
4 1981 350
5 1983 70
6 1984 15
7 1991 30
8 1995 31
9 1998 25