It turns out that, on having applied

**logarithmic transformations**, durbin watson throws a similar value, being the different graph

Code:

```
library(lmtest)
tb = data.frame("treat" = factor(x = c(rep(1:5, c(4,4,4,4,4)), labels = c("M1","M2","M3","M4","M5")),
"value" = c(110,157,194,178,1,2,4,18,880,1256,5276,4355,495,7040,5307,10050,7,5,29,2))
fit = aov(formula = value ~ treat, data=tb)
# durbin test
dwtest(fit)
plot(fit$residuals)
# -----------------------------
# transformation -- log
tb$value = log(tb$value)
# durbin test
dwtest(fit)
plot(fit$residuals)
```