Here is an example of the results for one subject. The subject was presented with 50 flashes at each of the intensities of the open circles. The intensities are in arbitrary units related to settings on the apparatus.
You can see that if the subject was presented with 50 flashes at each of the six intensities, there were 300 stimulus trials for this subject.
At the lower intensities the flash was detected less often and at the higher intensities the flash was detected more often. The lowest intensity was never detected (too dim) and the highest intensity was always detected (bright enough to be always seen). A curve was fit to the data. For now we can disregard how the curve was fit.
An arbitrary threshold was selected at the 60% level. The value of the function corresponding to this threshold value is read off the graph (or the fitted function). This value, 5, is taken as the threshold for detecting the light. At this intensity, the flash is detected 60% of the time. Sometimes an experimenter may choose a higher or lower level for the threshold based on theoretical or experimental reasons.
Now, Hecht, Schlaer, and Pirenne, calculated how much light, on average, is contained in a flash of intensity "5". The did their calculations in quantal units (how many photons) for reasons that will be clear later. Sometimes energy units are used. The number of quanta emitted by a flash at intensity 5 was too small to measure directly and had to be calculated based on the properties of light and the instrument (quantal fluctuations, transmission of filters, etc.) but these issues are beyond our scope.
To make a long story short, the results indicated that a threshold flash corresponds to approximately 90 quanta delivered to the eye. This is a remarkable small number. Compare these 90 quanta to the 2 x 10e15 quanta emitted by a flash bulb.) But the story does not end here.