Hematoxylin

Hematoxylin is extracted from the heartwood of the logwood tree, Haematoxylum campechianum, and it is not itself the dye. It must be oxidized to hematein, slowly in air or quickly with an oxidant such as sodium iodate. Hematein then needs a mordant, usually aluminum, to form a positively charged dye-metal complex that binds the phosphate groups of nucleic acids. That is why chromatin stains strongly.

Freshly applied, the color is reddish purple, and a brief alkaline rinse, the bluing step, turns it the familiar blue. Progressive formulations such as Mayer's are applied until the right depth is reached. Regressive ones such as Harris's are overstained, then partly removed with dilute acid.

The blue is never quite the same twice. It depends on the formulation, the staining time, the bluing solution and its time, the age and oxidation of the solution, and section thickness. The same tissue stained in two laboratories can differ visibly, and so can two runs in one laboratory weeks apart.

That matters for image analysis because color deconvolutionLoading... needs a stain vector: the exact color of pure hematoxylin. A published default may not fit your slides. Estimate the vector from a hematoxylin-only control section stained in the same run, and re-estimate it when the protocol or reagent lot changes.

In brightfield IHCLoading..., a nuclear marker puts DABLoading... and hematoxylin in the same compartment. A heavy counterstain then overlaps weak nuclear DAB and partly masks it, both to the eye and to deconvolution.

A lighter counterstain makes weak nuclear positivity easier to detect and separate, at the cost of fainter nuclei for segmentation. Settle the balance on control slides at the start of a study and keep it fixed. The depth of the blue decides which weakly positive nuclei are called positive, so changing it midway changes the results without any change in the marker.