Error in correlog: Bad selection of class breaks, try another one

I am trying to test my glm model residuals for spatial autocorrelation by using a correlogram using the function correlog from the package pgirmess. I have read here that reducing the number of classes could work (however it didn't) and that it could possibly be a result of too small a data set.

My data:

data <-read.csv('LandscapeFireMetrics.csv')
    structure(list(CircleID = c(158L, 201L, 235L, 237L, 254L, 287L
), Lat = c(-22.648623, -22.5945, -22.65447, -22.769486, -22.745345, 
-22.824849), Lon = c(122.609423, 122.44868, 122.10026, 122.776987, 
122.10434, 122.77037), Richness = c(40L, 50L, 68L, 46L, 58L, 
56L), FFDiv = c(0.6973625, 1.5231012, 1.440973, 1.3545885, 1.4448913, 
1.5282626), TSFDiv = c(0.560789, 1.3154271, 1.7571675, 1.42653, 
1.5950494, 2.0779354), MaxArea = c(15.86722306, 9.24703964, 16.03766892, 
19.08628445, 8.516517773, 8.460637152), AverageRain = c(234.184, 
227.566, 232.387, 241.896, 230.983, 221.319), spinsandplain = c(0.81, 
0.91, 0.88, 0.88, 0.98, 0.82), KunarkaProp = c(0.17, 0.08, 0.17, 
0.41, 0.16, 0.3)), row.names = c(NA, 6L), class = "data.frame")

Loaded packages:

library(ape);library(sp);library(lme4);library(MASS);library(spdep);library(pgirmess)

Model:

m1 = glm(Richness~TSFDiv+FFDiv+(MaxArea)+(spinsandplain)+(AverageRain)+(KunarkaProp),family=poisson,data=data)

Correlogram code:

rs <-residuals(m1)
spatialstuff <- data.frame(x = data$Lat, y = data$Lon)
fit2 <- correlog(coords = spatialstuff, z=rsd,increment = 1, resamp = 10, quiet = TRUE,nbclass =1)
plot(fit2)

From that I just keep getting this error:

Error in correlog(coords = spatialstuff, z = rsd, increment = 1, resamp = 10,  : 
  Bad selection of class breaks, try another one...