Uses of error: A word of advice

In research, chance observations are not uncommon. In 1958, when studying the effect of endotoxin on haemostasis in the dog, I happened to notice that in the presence of fibrinolysis a decline in the fibrinogen didn't always mean it was consumed or broken down. If I repeated the procedure on plasma samples that by the clotting technique contained little or no fibrinogen I would find, to my great surprise, that it was present if I used the salting-out technique instead. I was mystified. I pursued this strange finding for months. Since I was new to the field, I tried to learn all I could but couldn't find anything on it.