One part of evidence that the color of the flame created is
from the metal ion and not from the chemical is that not any of the flames with
dissimilar metals had the similar color (for each metal had its own flame
color). Even if most of the metals tested had chloride, the colors of the flames
were all dissimilar. The two flames that both had copper (one had copper (II)
chloride and the other had copper (II) sulfate) were exactly close in color. The
one was green-blue, and the other was a bright green. This displays that they
were nearly the same, and the minor difference could be credited to error.