Dgrees of freedom of error sum of squares

[Clarkson]

For a given metal a experiment was carried out to measure the heat variation with temperature in Celsius.The heat was measured twice at each chosen temperature.

If the chosen temperatures are 30,40,75,85,90
Then there would be total of 10 observations(2 for each temperature).In order to find Error sum of squares I need to find the degrees of freedom.What is the degrees of freedom.
Does the total no.of observations=10

 

[emily.stats]

In this experiment you don't have any independent replicates because the two measurements were taken on the same experiment unit. Therefore you do not have any error degrees of freedom. Anova is not valid.

 

[Clarkson]

Can't a simple linear regression be carried out?