Mechanism of Golgi’s silver stain
Why does Golgi’s silver stain colour only a small, random set of nerve cells, but those in full?
Summary
Why it is open
In Golgi's method, blocks of nervous tissue are kept in potassium dichromate, for anything from a day to 50 days or longer, and then immersed in silver nitrate; silver chromate precipitates and fully impregnates some cells [6]. Dyes stain by chemical affinity for a tissue component, whereas metal impregnation results from the preferential precipitation of metal salts onto certain structures, "for reasons that are not fully understood", Marina Bentivoglio and colleagues write [6]. They call the way the stain fills "at random a limited number of cells" "the still unexplained secret of its revelatory power" [6].
The problem is two questions. The first is chemical: where the precipitate starts, and why it fills a cell once it has started [2, 3]. The second is about selection: whether the stained cells are a random sample, or differ from their unstained neighbours. A 1975 count in mouse cortex found a random sample with respect to cell size and position [1]; a study published in 2026 found that age, time of day and light exposure change how many cells are stained in mice [9].
No body sets a bar for a resolution. On the evidence below, an answer would need to predict which cells will be stained in a given piece of tissue, and to account for the variables already shown to alter staining, such as fixation [5] and the animal's condition before the tissue is taken [9]. The experimental literature on the mechanism is small; this page cites four such studies [2, 3, 8, 9].
What is known
The method and its history (1873–1898). Camillo Golgi devised the black reaction in 1873, while head physician at a hospice in Abbiategrasso, near Milan, working in a laboratory he had set up in his kitchen [4, 6]. It was the first technique to show neurons in their entirety, with all their processes [4]. It passed unnoticed at first; interest outside Italy began in 1885 [4], and Santiago Ramón y Cajal first saw Golgi-impregnated sections in 1887 [6]. Results obtained with the stain were vital for establishing the neuron theory [4], which Cajal defended against Golgi's reticular theory of continuity between nerve cells [6]. In 1898 Golgi used a rapid variant of the impregnation to describe an "internal reticular apparatus" in neurons of spinal ganglia, the organelle later named the Golgi apparatus [6], whose own open problem is transport through the Golgi apparatus.
What the precipitate is. In the original method the deposit is silver chromate [2, 6]. In the Golgi–Cox variant, tissue is impregnated with a mixture of potassium dichromate, mercury chloride and potassium chromate [7], and the black deposits on stained neurons have been attributed to mercuric sulfide [8].
How many cells, and which (1975). Joseph Pasternak and Thomas Woolsey counted every Golgi–Cox-impregnated neuron and every Nissl-stained neuron in one barrel (barrel C-1) in layer IV of mouse somatosensory cortex [1]. The impregnated share varied between preparations from 0.73% to 2.26%, averaging 1.29% [1]. Their size distributions were similar; a small difference in mean size matched the shrinkage seen when the precipitate was removed, and with the means equated the two groups did not differ [1]. Impregnated neurons occurred in the same proportions as Nissl-stained ones in the sides and hollows of the barrel [1]. The authors concluded that this variant "impregnates barrel neurons randomly" [1].
Watching impregnation happen (1989–1992). Josef Špaček mounted aldehyde-fixed, dichromate-treated sections of cerebral cortex under a coverslip, replaced the dichromate with silver nitrate and filmed the reaction [2]. In sections 100 µm thick, shrub-like focal precipitates, which he called nucleation centres, appeared after about 7 minutes; from them, and from the section edges, thread-like growths, usually identified as dendrites, spread into cell bodies, and sometimes impregnation began in a cell body and spread into its dendrites [2]. Deposition typically advanced 1–7 µm per minute and was very slow after 30 minutes, by which time many neurons were more or less fully impregnated [2]. Under the electron microscope the earliest deposits inside cells did not cross the plasma membrane or internal membranes; their granules, at first about 10 nm across, enlarged and merged until they filled processes and cell bodies [2]. Špaček concluded that impregnation "occurred only in those cells a part of which was within a nucleation centre" [2]. In tissue blocks he found two modes: crystals growing into cells through membranes cut at the tissue surface, and precipitate spreading from nucleation centres within the tissue, which grows inside the cell "as in a preformed channel until the cell has been filled" [3].
What changes the stained population (2012–2026). In rat brain, Amit Ranjan and Birendra Mallick found that exposing tissue to paraformaldehyde or formaldehyde at any stage of their modified Golgi–Cox protocol greatly increased the number of stained glia, whereas tissue never treated with fixative showed stained neurons; they described the mechanism as unclear [5]. Melissa Richardson and Lauren Barnwell reported in mice that neuronal maturity, time of day and light exposure change the number of Golgi–Cox-stained cells, with low variability within each group [9]. In the preprint version of that paper they propose that in adults the stain captures "quiescent neurons", because fewer cells were stained in hypothalamic regions at times, or after light at night, when other studies report more neural activity there [9].
Main approaches
Crystal nucleation. Staining is a crystallisation process: a cell is stained when a nucleation centre forms within or next to it, and the growing crystal then fills it [2, 3]. Strongest evidence: time-lapse recordings of nucleation centres and of growth into dendrites and cell bodies [2, 3]. Main difficulty: the account moves the question to where nucleation centres form, and three decades later Bentivoglio and colleagues still described the selection of cells as unexplained [6].
Membranes as containers. The cell membrane confines the precipitate, which would explain why a stained cell is filled to the tips of its branches [2, 3]. Strongest evidence: electron micrographs of early deposits that do not cross plasma or internal membranes, and growth "as in a preformed channel" [2, 3]. Main difficulty: precipitate that nucleates outside cells also extends into the narrow gaps between them [3], and confinement explains complete filling rather than which cells are entered [2, 6].
Cell chemistry. Something in the stained cells binds the metal. Strongest evidence: Min Ai and colleagues observed fluorescence from Golgi–Cox deposits on mouse neurons and, using infrared and X-ray photoelectron spectroscopy, identified mercury-binding proteins in them, which they say helps to explain how Golgi–Cox staining works [8]. Main difficulty: the work concerns the mercury-based Golgi–Cox variant, and its abstract reports deposits on stained neurons only [8].
Cell state. Whether a neuron is stained depends on its condition when the tissue is taken [9]. Strongest evidence: changes in the number of stained cells with age, time of day and light at night in mice [9], and the shift from neurons to glia with fixation in rats [5]. Main difficulty: the link to activity rests on comparing the counts with activity measured in other studies, and comes from one laboratory [9]; Richardson and Barnwell write that the pattern "is not random" [9], whereas the 1975 count found a random sample by size and position [1], and the two studies measured different things.
Open sub-questions
- Where nucleation starts. In time-lapse recordings like Špaček's, does the first nucleation centre form preferentially in a particular compartment or cell type, and can its position be predicted before silver nitrate is added [2, 3]?
- Random with respect to activity. Where Golgi–Cox labels about 1% of neurons, are the stained cells random with respect to markers of recent activity, as they are with respect to size and position [1, 9]?
- Silencing and stimulation. Does silencing or stimulating a defined population before tissue collection change its share of stained cells, as the quiescent-neuron proposal predicts [9]?
- Fixation and glia. Which step of aldehyde fixation shifts Golgi–Cox staining from neurons to glia, and does the same shift occur with the silver chromate method [5, 7]?
- Unstained neighbours. Are the mercury-binding proteins found in Golgi–Cox deposits also present in unstained neighbouring neurons [8]?
Work on this in Papex
If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.
Key references
- Pasternak, J. F. & Woolsey, T. A. (1975). On the "selectivity" of the Golgi-Cox method. Journal of Comparative Neurology 160(3), 307–312. 10.1002/cne.901600304 · Checked 16 Oct 2026
- Špaček, J. (1989). Dynamics of the Golgi method: a time-lapse study of the early stages of impregnation in single sections. Journal of Neurocytology 18(1), 27–38. 10.1007/BF01188421 · Checked 16 Oct 2026
- Špaček, J. (1992). Dynamics of Golgi impregnation in neurons. Microscopy Research and Technique 23(4), 264–274. 10.1002/jemt.1070230403 · Checked 16 Oct 2026
- Pannese, E. (1996). The black reaction. Brain Research Bulletin 41(6), 343–349. 10.1016/S0361-9230(96)00177-3 · Checked 16 Oct 2026
- Ranjan, A. & Mallick, B. N. (2012). Differential staining of glia and neurons by modified Golgi-Cox method. Journal of Neuroscience Methods 209(2), 269–279. 10.1016/j.jneumeth.2012.06.023 · Checked 16 Oct 2026
- Bentivoglio, M., Cotrufo, T., Ferrari, S., Tesoriero, C., Mariotto, S., Bertini, G. et al. (2019). The original histological slides of Camillo Golgi and his discoveries on neuronal structure. Frontiers in Neuroanatomy 13, 3. 10.3389/fnana.2019.00003 · Checked 16 Oct 2026
- Vints, K., Vandael, D., Baatsen, P., Pavie, B., Vernaillen, F., Corthout, N. et al. (2019). Modernization of Golgi staining techniques for high-resolution, 3-dimensional imaging of individual neurons. Scientific Reports 9, 130. 10.1038/s41598-018-37377-x · Checked 16 Oct 2026
- Ai, M., Huang, K., Ji, Z., Wang, Y., Liu, Y., Xiao, L. et al. (2021). Unveiling Hg-binding protein within black deposit formed on Golgi-Cox-stained brain neuron. Neuroscience Letters 742, 135537. 10.1016/j.neulet.2020.135537 · Checked 16 Oct 2026
- Richardson, M. E. S. & Barnwell, L. M. (2026). Developmental, time-of-day, and stimulus-specific Golgi-Cox staining patterns detected in the mouse brain. Journal of Neuroscience Methods 425, 110618 (preprint posted July 2025, bioRxiv 10.1101/2025.07.03.663022). 10.1016/j.jneumeth.2025.110618 · Checked 16 Oct 2026
Source
Open, as listed by Wikipedia · Unsolved problems in biology. Checked 08 Oct 2026.
The status is the source’s. Papex does not decide whether a problem is solved.
Records on this problem
No public record addresses this problem yet.