Now let us take a piece of carbon steel as before, but this time containing .15% carbon, and again proceed with our observations. Again the needle of the pyrometer records the point of recalescence and also the point designating the second range of critical temperature, but this time strange to say, as the test piece continues to absorb heat, a third critical range is registered, all of which when added to our former picture gives a result something as follows: [Illustration: Graph showing different behavior of samples containing different rates of carbon] By repeating the operations as outlined above, with pieces of steel containing various percentages of carbon from zero to 1.25% and by plotting the different critical temperatures so obtained, we finally obtain a chart which graphically expresses the critical ranges of iron and steels due to the variation of the carbon content. With very low carbon steel it is interesting to note that the first critical point would not occur until 1395 degrees Fahrenheit was reached. Metallurgists have long designated the lines so obtained by letters, “r”, standing for, “refroidissement”, which is the French word meaning “cooling”, the suffixes 1-2-3 simply standing for the lines in the order drawn. From the completed chart it is further evident that our first piece containing 0.9% carbon in one way is the most interesting of all since it is the only case where only one point of critical temperature occurs. It will be noticed from the chart that steels containing less than .10% carbon have no point Ar1 and it is therefore undoubtedly due to the carbon content that this, the point of recalescence, occurs. From tests which we made with the magnet we would also find that the temperatures at which loss of magnetism occurs are those designated by the line Ar2, whereas the loss of ability to conduct an electric current occurs at the point designated Ar3. In steels containing .45% carbon to .75% carbon loss of magnetism and loss of ability to conduct an electric current occur at the same points designated on our chart by the line Ar3-2; whereas in the steel containing .90% carbon—all these changes take place at the same time. Now, as we concluded before, it is evident that some internal change must have taken place in the steel itself, and as we know that the chemical content does not vary, it is further evident that the change must be of a physical nature, or as in the language of the Metallurgist, an “allotropic change”. Therefore, another conclusion which we can draw at this point is that a very much more thorough investigation is required for the proper handling of steel at high temperatures than a mere knowledge of the chemical analysis of the same.