The Albumen Print as an Object, and How It Decays

Pick up any nineteenth-century photograph in any collection and the odds are about five to one that it is an albumen print, and the odds are about five to one again that its whites have gone yellow. Reilly states the first as a proportion — the albumen print “accounts for approximately 85 % of the total number of surviving 19th-century photographic prints” — and the second as an absolute: “it is probably safe to say that not a single albumen print survives from the 19th century without some degree of staining in non-image areas.”

The interesting question is not whether these prints have deteriorated. It is why, and how much of the answer is settled. That matters practically, because if the yellowing is a processing failure then better processing fixes it, and if it is a property of the material then nothing you do at the bench will prevent it and the only lever left is storage. The evidence points at the second answer, and this page sets out how strong the evidence is claim by claim.

The object as it left the studio

A finished albumen print of the 1870s was a thin, glossy, warm-purple image on a sheet of paper so light it could not stand up on its own, trimmed, pasted onto card, rolled or burnished to a high polish, often retouched, and issued in one of a small number of standard formats.

Thin, and unable to lie flat. The Getty measured an unmounted albumen photograph stored flat at 0.204 mm overall, support included, and records that such prints “had a strong tendency to curl inside, forming tight rolls of unmounted albumen photographs that, when left in such a state after processing, were rather fragile and difficult to handle without special treatment and conditioning.” Reilly’s version is the mechanical one: the force in a thick albumen layer is “quite enough, if a print has been incorrectly mounted, to tear the print in half.”

Mounted, almost always. Reilly’s figure is that “approximately 95 % of all albumen prints were mounted at the time of their production.” That was not a presentation choice; it was structural. He adds a second reason that is aesthetic and is easy to miss: mounting “helps to impart more depth and contrast to the image, and creates a smooth, level surface which can be viewed from different angles without the interference of these uneven surface reflections.”

Polished. Rolling and burnishing were routine, and the reason was optical as well as adhesive: a heavy albumen layer can dry with “a horny, rough surface that may obscure the finest details”, and rolling “tends to restore some of the detail and also provides a glossier surface and increased contrast”. Small prints — cabinet portraits and stereo views — went through a heated burnisher which in skilled hands “could produce a mirror-like gloss”, sold under names like “French Enamel” and “Extra Superior Finish”. You can date a print from the back of its mount: cross-hatched indentations mean a wringer-type machine with a ribbed drive roller, common on small prints of 1875 to 1890, and a uniformly shiny, squeezed-looking back means a flat-bed press.

Retouched, and the retouching is now visible. The Getty’s observation is one of the melancholy ones in the atlas: “the retouching material was usually more resistant against light fading than the image material of the albumen photograph. Today many old, retouched photographs exhibit clearly visible retouching marks superposed over the slightly faded original image.” Work that was invisible in 1875 is the first thing you see now.

And tinted, more often than people expect. From 1863 a very large proportion of albumen paper carried aniline dyes in the albumen — pink, purple and blue, with pink the most popular — and the Getty is explicit that the tints were added “to counteract the yellowing”. The dyes had poor light stability, so most of that paper is unrecognisable today; Reilly notes that a peculiar buff or chamois cast sometimes identifies a print that started tinted, a “dirty” colour already complained of in the 1890s.

!Standard formats: what the course can and cannot source

The mustVerify list for this page asked for the standard formats and their dimensions “from a museum source rather than a general one”. Two are sourced and one is not, and the gap is stated rather than filled.

Carte de visite. Reilly describes it as “an albumen print on a cardboard mount that was approximately 6 x 10 cm in size”, and dates the mania to Paris and the early 1860s. IPI’s Graphics Atlas describes a period example as “an albumen print mounted to a thick, 2 ½ x 4-inch paper card” with rounded corners — 6.4 × 10.2 cm, which agrees with Reilly to within the rounding.

Stereograph. Reilly: “nearly all stereo views before 1890 were made on albumen paper”, and the format’s demand was one of the two things that created the albumen paper industry. No source read for this course gives a mount dimension for it.

Cabinet card. Reilly names the format, notes that cabinet-size portraits were among the prints most often burnished, and the Getty lists cabinet cards among the card formats that “many, but not all” contain albumen prints. No source read for this course publishes a dimension for it. The figures in general circulation may well be right; they are not in the corpus, and this page will not print a number it cannot cite.

Gloss and surface as period choices

The single most useful thing to understand about the surface of an albumen print is that there was never one surface. The material spanned a range, the range was a matter of taste and taste moved, and the whole span is visible in surviving objects.

Choice	What it did	When
Diluting the albumen with water	Less gloss, less detail, less “brilliance”	Especially the 1850s, when the Getty records that “the glossiness of albumen photographs was highly criticized in the photographic literature” and the public preferred the matte character of salt prints
Undiluted albumen, aged or fermented	More gloss and a more even coating	From the mid-1850s, standard in Dresden from the early 1870s
Double coating	Considerably more gloss and depth, harder toning, worse curl	“A large portion of the albumen paper sold after 1880”
Drying warm	Faster and glossier	Factory drying rooms at 30 to 50 °C
Burnishing and rolling	Higher gloss, restored detail, increased contrast — and it suppresses the crack network	Especially 1875 to 1890
Varnishing	Higher gloss and a claimed protection against moisture and pollutants	Recommended through the second half of the century
Matte albumen, albumen mixed with starch	A matte surface resembling platinum and gum bichromate	The Getty’s timeline dates its introduction to 1897

Varnishes are worth knowing about because they are commoner than they look. The Getty identified four kinds on surviving prints by infrared spectroscopy: beeswax, paraffin wax dissolved in kerosene, collodion, and shellac. A thick collodion coat makes an unmounted print very stiff and tends over time “to curl the edges of the photograph and in many instances partially separate the photograph from the mount substrate”; a thin one shows as iridescence. None of them is visible under a microscope as a layer.

And the same negative reads completely differently across that range. Reilly’s account of why is the argument of the whole part, working in reverse: a glossy binder minimises the scattering of light both in the shadows and in the whites, so a matte version of the same image looks paler in the deep tones, less “brilliant” in the light ones, and — because a matte paper needs relatively more reduced silver to make a black — will take a longer exposure and want a longer-scale negative to go with it. Matte albumen is not a variant of the glossy print; it is a different tonal instrument built from the same chemistry.

The two failures that dominate

Almost everything you will see on a surviving albumen print is one of two things, and they are independent of each other.

Highlight yellowing. A yellow or yellowish-brown stain in the non-image areas. Reilly: “the most common form of deterioration of historical albumen prints”, and it “is so prevalent in albumen prints that it often serves as an important clue in their identification, since the highlight yellowing phenomenon is peculiar to albumen paper and does not occur in quite the same way in otherwise similar gelatin and collodion papers.”

Cracking of the binder. The Getty calls a fine network of surface microcracks “the most typical microscopic signature of albumen photographs”, uniform across the layer.

Then, behind those two, generalised image fading — which Reilly ranks as “the second most prevalent kind of deterioration — and the most serious for the informational and aesthetic value of the photograph.”

!Reilly's 85 per cent figure, and what it is

Reilly writes that “approximately 85 % of extant albumen prints made after 1860 display what might be called ‘moderate to severe’ yellowing, while the remaining 15 % or so seem to have white highlights unless compared side by side with a white sheet of paper”, and that for prints of the 1850s “the figures probably are closer to 95 % and 5 %.”

He then says exactly what those numbers are, and the sentence should always travel with them: “These figures are not based on any formal statistical sampling, merely on the accumulated experience of the author in examining prints and through discussions with curators and collectors.” He goes on to call for the study that would settle it: “A statistical study of a large collection of albumen prints with regard to the type and severity of print deterioration — including highlight yellowing — would be a most welcome addition to the literature of photographic preservation.”

That was 1980. No source read for this course reports such a study having been done. The figure is an expert’s estimate from an expert who told you it was an estimate, which is worth more than most numbers and is not a measurement.

One further observation of his belongs with it, because it explains why the problem is under-noticed. Human vision “has a built-in mechanism that automatically seeks out the lightest area in a photographic print and pegs that as a ‘reference white’”. So a yellowed print looks acceptable until you put a true white beside it — and Reilly’s wry conclusion is that this adaptation “is indeed fortunate for our appreciation of albumen paper photographs, since otherwise a majority of albumen photographs would seem excessively ‘flat’ and lifeless.”

Yellowing: what is established, and what is not

Four explanations are on the table. They are not alternatives in the sense that three must be wrong; the honest position is that at least two are operating and the literature has not apportioned them.

⁇Why albumen prints yellow: four candidates, and how much weight each carries

Candidate one: albumen-bound silver converting to silver sulfide. This is Reilly’s own account and the one the conservation literature has inherited. The chain is: sulfur-bearing side chains in the egg protein bind silver during sensitising; the bond is strong enough that “treatment in hypo is not sufficient to remove it”, so a small amount of silver remains in all areas of a fixed print; labile sulfur from residual fixer or from the atmosphere converts it to silver sulfide, which is yellow.

How strong is it? The retention is measured and old: Davanne and Girard found in 1859 that cyanide removed the silver where hypo did not, Spiller named it in 1868 as “an argentic organic compound, colorless, unalterable by light, and comparatively insoluble in hyposulphites”, and Haddon and Grundy in the mid-1890s measured nearly 5 per cent of the applied silver still present in an unexposed, thoroughly fixed and washed sheet — and then printed an image out of it. Ware supplies the reason it is specific to albumen: egg albumen carries much more cysteine than gelatin does. The sulfide step is a chemically reasonable and universally repeated inference; no source read for this course reports a spectroscopic identification of silver sulfide in an albumen print’s highlights.

Candidate two: the Maillard reaction. Glucose in the albumen combines with amino groups of the egg protein to form a highly coloured, conjugated, insoluble compound. The AIC lists it, and adds that “studies show that this yellowing is accelerated by exposure to high humidity, high relative humidity and alkaline conditions”. Ware attributes it to Reilly’s own later work: Reilly and colleagues incubated test specimens at high temperature and various relative humidities, and “their results show that the Maillard, or ‘protein-sugar’ reaction is also an important factor contributing to highlight yellowing.”

How strong is it? This is the candidate with experimental support behind it — accelerated ageing under controlled temperature and humidity, published in the conservation literature of the early 1980s. The course has read Ware’s and the AIC’s accounts of that work and not the primary papers themselves, which are conference preprints outside the corpus.

Candidate three: oxidation of aromatic amino acid side chains. The AIC’s fourth bullet: “egg albumen contains a number of benzene ring-containing amino acids; upon absorption of light a benzene ring structure may be oxidized to a highly colored substance.”

How strong is it? It is stated as a possibility in one tier-1 source and no evidence is cited for it there. Chemically it is unobjectionable — tyrosine, tryptophan and phenylalanine are ordinary constituents of egg white protein — but a mechanism that is plausible is not a mechanism that has been shown to operate here.

Candidate four: heat and humidity as such. The AIC lists “a combination of high humidity and temperature” first among its causes. How strong is it? As a rate statement this is the best supported thing on the list: Ware reports that Reilly’s studies show the instability “is greatly accelerated at high relative humidities”, and it is why every storage recommendation on this page is a temperature and a humidity. As a mechanism it is not one — it is the accelerator of candidates one, two and three rather than a fourth pathway.

What the course states. That the retained silver is measured and undisputed; that both the sulfide route and the Maillard route are supported, the second with accelerated-ageing evidence behind it and the first with a century of chemical reasoning and one nineteenth-century measurement of the precursor; that the aromatic-oxidation route is a hypothesis; and that no source read for this course apportions them. Anyone who tells you albumen prints yellow “because of residual hypo” has collapsed four candidates into the one that happens to blame the printer.

What would distinguish them. Each makes a different prediction and the experiments are conceivable. The sulfide route predicts that yellowing tracks available sulfur — so prints stored in sulfur-scrubbed air should yellow more slowly, and the stain should be identifiable as silver sulfide by elemental or vibrational analysis. The Maillard route predicts that yellowing occurs in a print carrying no silver at all, so an albumenised but never-sensitised sheet, aged alongside a print, separates the two cleanly. And the aromatic-oxidation route is the only one that requires light, so dark storage should switch it off. Pretzel and Martin’s finding that Lady Hawarden’s prints changed nearly as much in dark storage as on exhibition is, on that reading, evidence against light being the dominant driver.

⚗The one thing about the yellowing that is not in doubt

Whatever forms the stain, it is more stable than the image. Reilly: “The reversal of yellowing that has already occurred is an extremely difficult task, principally because the silver sulfide which forms the yellow stain is much more chemically stable than the colloidal silver of the image itself. No presently known treatment can remove the highlight yellowing without completely altering the character of the print.”

That is the cruelty of the material and it is worth sitting with. Silver sulfide is what a sulfide toner makes on purpose, because it is inert and durable. Here the same durability is the problem: any reagent strong enough to attack the stain attacks the picture first.

2 Ag + S → Ag₂S

The reaction that protects a sepia print and disfigures an albumen one — the same product, in the wrong place

Cracking

The surface at 25×: albumen, salted paper, gelatin

1. 1Albumen, unburnished, before about 1870 — the Getty: a fine network of surface microcracks, "rather uniform across the surface layer", found in most albumen prints of this period. Paper fibres visible through the layer between the cracks
2. 2Albumen, burnished or heat-treated, after about 1870 — the Getty: such prints "may not fully exhibit such a microcrack pattern, or surface cracks may not be visible at all". Absence of crazing does not rule out albumen
3. 3Salted paper — no layer, so nothing to craze. The image is among the fibres and the surface is the paper's own
4. 4Fibre-base gelatin — a continuous layer over an opaque baryta floor. No fibres visible through it, which is the Getty's two-layer test running the other way

Drawn from the conservation descriptions rather than traced from a micrograph: the course holds no licensed image of an albumen print surface. The crack pattern is schematic — its density and island size are illustrative and are not measured from any published micrograph.

The mechanism is mechanical, and both tier-1 conservation sources give the same one. The AIC: “It is frequent to find overall cracking of the albumen binder, likely caused by variations in humidity that caused differential dimensional changes between the image layer and the support.” A brittle protein film is glued to a paper sheet that takes up and gives back moisture with the seasons. The paper moves; the film does not move by the same amount; the film is thin and rigid and in tension; it cracks, and goes on cracking into a finer and finer network. Reilly’s description of the dried material — “a brittle, transparent mass” — is the property that makes this inevitable.

Note the word “likely” in the AIC’s sentence. That is a conservation wiki hedging its own mechanism. The literature it points to is a study the course has not read: Vitale and Messier’s ESEM investigation of cracking in albumen photographs, listed in the AIC’s further reading and not in the corpus. Anyone taking this further should start there.

Two consequences for identification, and they point opposite ways. The presence of a uniform microcrack network is strong evidence for albumen, and Reilly notes the surface texture of albumen paper “sometimes possesses a ‘crackled’ or ‘crazed’ appearance” that experienced eyes recognise immediately. But its absence proves nothing, because the Getty found that prints made after 1870 and burnished or heat treated “may not fully exhibit such a microcrack pattern, or surface cracks may not be visible at all” — and those are precisely the glossiest, most heavily coated prints, which naively you would expect to crack most.

And one consequence for treatment. Reilly’s caution on the conservator’s mounting method he publishes is that it “exerts considerable stress on the albumen layer of the print as it dries. While such stress does not usually harm prints, especially newly made prints, some 19th century prints may have a visibly weakened or severely cracked albumen layer. In this event mounting by Kolody’s method is not recommended, because the stress of mounting may worsen the cracking.” A crazed print is structurally different from an intact one, and the difference decides what can be done to it.

Fading, and what toning is worth

Fading is the loss of the image itself, and it has internal and external causes that Reilly keeps apart.

Internal: residual thiosulfate and silver-thiosulfate complexes left by inadequate fixing and washing. This is the printer’s fault and it is preventable.

External: sulfiding by atmospheric sulfur compounds such as sulfur dioxide, and oxidative fading by oxidising gases — ozone, organic solvents. Reilly lists both; Ware adds the reason printing-out images are so much more vulnerable than developed ones, which is geometry rather than chemistry. A printed-out image is colloidal silver in particles the AIC gives as 5 to 25 micrometres, and such particles have “a very large surface area relative to [their] mass; hence a large portion of [their] total mass is on the surface and readily accessible to destructive chemical agents.” A developed image is filamentary silver, much coarser, with far less surface per unit of silver.

The binder cuts both ways here, and Ware states both directions in one section. It protects: “for a thiosulphate-fixed print in which the colloidal silver is protected by a vehicle such as albumen, the oxygen of the air has much more restricted access to the image silver than in a salt print, which is quite permeable to gases”. And it attacks: “the presence of sulphur-containing protein compounds also greatly facilitates the oxidation of colloidal silver metal in the shadow areas, which consequently fade with time, probably regressing also to yellow colloidal silver sulphide.”

Silver mirroring is the third form and the AIC names it: “the silver particles of the image may have been physically redistributed, resulting in the formation of silver mirroring, frequently appearing along the edges of a print.” The course’s own entry carries the diagnosis.

And bleach-and-redevelop, which rescues a sulfided gelatin print, does not work here. Reilly gives three separate reasons: “the residual silver in the highlight areas redevelops along with the image, the ‘redevelopment’ step does not provide sufficient density overall, and finally, the color of the ‘restored’ image is black and therefore totally out of character with the original color of the print.”

⁇Does gold toning make an albumen print last longer, and what kind of evidence is that?

Everybody says yes. Ware: “the permanence of albumen prints was usually greatly improved by gold toning as recommended by the ‘Fading Committee’ of the Photographic Society; it also has the effect of shifting the image colour to a satisfyingly rich purplish-brown.” Reilly: the switch from the sel d’or method to separate alkaline gold toning “was a great step forward that consolidated the advances in print stability made during the 1850’s and allowed albumen paper to attain a much better record of resistance to fading than had been accomplished with the older plain salted papers. The new alkaline gold toners deposited more gold, and this contributed to the resistance of albumen prints to oxidative fading.” The AIC records the Fading Committee’s recommendation in 1855 as thorough washing and toning with gold chloride.

The question this page asks is what kind of evidence that is, because the design of this part demanded it, and the answer is: three kinds, none of them a controlled comparative trial.

Historical comparative observation. Reilly’s clearest case is not about toning at all but about silver: albumen prints of the 1860s were salted at 2 to 3 per cent chloride against the weaker salting of the 1880s and 1890s, which gave higher average shadow densities and therefore more image silver; “more image silver results in improved resistance to fading, and albumen prints of the 1860’s and 1870’s seem to have accumulated a slightly better average record of durability … than their weakly salted and scantily exposed successors”. Note “seem to have accumulated a slightly better average record” — this is a careful observer’s impression of a surviving population, with all the survivorship problems that implies.

Mechanism plus practitioner inference. Gold and platinum “react much less readily with sulfur and are much more difficult to oxidize. A layer of gold or platinum on a silver particle will tend to shield the silver inside from attack, especially from oxidizing agents.” That is sound and it is a mechanism rather than a measurement.

Accelerated ageing, on the underlying instability rather than on toning. Reilly’s incubation studies are the measured work in this area, and what Ware reports them as showing is that “the instability of albumen prints is inherent, rather than due to inadequate processing, and that it is greatly accelerated at high relative humidities.” That is a finding about the material, not a toned-against- untoned trial.

What the course states. That gold toning is supported by the mechanism, by 170 years of practitioner consensus beginning with a first-class investigating committee, and by comparative observation of surviving objects; that no source read for this course reports a controlled comparison of toned and untoned albumen prints under measured ageing conditions; and that a printer should therefore tone thoroughly — the argument is strong — while knowing exactly which kind of argument it is.

And here is where your untoned control print earns its keep. The printing lab asked you to fix and dry one print with no toning at all. Keep it with its toned twin, in the same enclosure, with the date on both. In twenty years that pair is a better piece of evidence than anything on this page.

🧮How much light an albumen print can take, converted into hours

Ware’s Table 20.3 collects the published threshold exposures — the dose at which a just-noticeable difference appears — in kilolux seconds, and puts albumen prints in his category 3, moderately sensitive.

Process	Threshold exposure (klx·s)	Source and note
Photogenic drawing	600	Ware 1994; Talbot, salt-stabilised
Cyanotype	720 at 4,000 lx	Ware 1999; Herschel’s process. Non-reciprocal — the figure at 50 lx is over 36,000
Salted paper print	54,000	McElhone 1993; thiosulphate fixed
Albumen print	80,000 to 2,880,000	Pretzel 1991; Ware’s comment: “highly variable”
Silver-gelatin print	over 1,800,000	Modern processing

The course’s own unit conversion, so that the range means something. One klx·s is 1,000 lux seconds, and there are 3,600 seconds in an hour:

80,000 klx·s ÷ 3,600 ÷ 50 lx = about 440 hours

Lower bound, at a 50 lux display level

2,880,000 klx·s ÷ 3,600 ÷ 50 lx = about 16,000 hours

Upper bound, at the same level

So at gallery lighting, a first just-noticeable fade somewhere between about eighteen days and about two years of continuous display — a factor of thirty-six across the population. The arithmetic is the course’s; the figures and the units are Ware’s, and the conversion assumes reciprocity holds, which Ware says “may be generally true for silver images” while warning that it certainly does not hold for cyanotypes.

The spread is the finding, not the midpoint. Ware’s own reading: “the conspicuously wide range of values cited here for the threshold exposure of albumen prints is an indication that less stable ‘mavericks’ can lurk in any population of historic photographs, because of uncertainties concerning the quality of processing.” You cannot look at an albumen print and know which end of that range it sits at.

And a separate measurement points the other way about light. Pretzel and Martin studied albumen prints by Lady Hawarden at the Victoria and Albert Museum and found average threshold exposure lives “of the order of one or two years (ca. 10,000 hours), and nearly as much change was observed to occur in dark-stored as in exhibited material, implying that the influence of light exposure alone was not significant.” Two conclusions follow that matter for a collection. Turning the lights down does not stop an albumen print deteriorating. And the thing that does change the rate is the one every storage recommendation below is about: temperature and humidity.

The mount as an agent of decay

Ninety-five per cent of these prints are pasted to a piece of card, and the card is often the worst thing that ever happened to them.

The construction was the problem. Reilly: “A typical 19th-century mount was composed of a thin top and bottom sheet of relatively good quality paper, with a center filler of poor quality pulp board. This pulp center was often loaded with lignin, the noncellulose component of wood, whose presence leads to the acidification of the entire mount and eventually of the photographic print itself.” The decomposition products migrate up through the top sheet and attack the photograph, “causing staining and brittleness and accelerating and fading and yellowing of the silver image.”

The albumen print is unusually exposed to it, and the reason is the thinness this page opened with: “Danger from the mount board is especially acute for albumen prints because they were made on such thin rawstock — very little barrier exists between the silver image and the potentially destructive substances in the mount.”

The adhesives are the second front. Starch was the usual choice and is the one conservators approve; the period also used gelatin, gum arabic, dextrines and albumen itself, and “ordinary ‘glue’ — which is an impure form of gelatin — was known to be a cause of trouble even in the 1850’s, but it was used as an expedient, along with many other destructive substances such as rancid flour paste and India rubber solution.” The Getty adds a visible failure mode: “sometimes the adhesive resulted in a visible tonality change or bleaching of the albumen prints, causing the lines of applied adhesive to show up in the prints as lighter, well-delineated areas.”

Foxing and mould. Reilly names “the brownish-red flecks and stains known as ‘foxing’”, which “may be the result of mold and fungus growth, or may also be caused by the presence of metallic salts in the mount board.” The AIC lists poor quality mounting materials among the causes of staining, discoloration and fading.

And the one treatment that has actually worked. Reilly is blunt about the state of the art: “the removal of prints from obviously defective mounts and their careful re-mounting onto appropriate mounts with safe adhesives is the only technique for the preservation of albumen prints that has proven itself in practice. Remounting, however, does nothing to reverse the deterioration that has already occurred.” He also notes the one piece of luck in all this: “fortunately albumen is a very durable substance, and can usually withstand the rigors of unmounting and remounting.”

Five routes to a ruined albumen print, and where a maker can intervene

1. 1Residual processing chemistry — thiosulfate and silver-thiosulfate complexes. **Preventable**: two fresh baths, conservative capacity, washing aid, full wash, and a residual-hypo test
2. 2The atmosphere — sulfur compounds and oxidising gases. **Manageable**: enclosure, and the storage environment below
3. 3The binder itself — silver bound to sulfur-bearing side chains that no fixer reaches, and the Maillard reaction between the albumen's own sugar and its own protein. **No intervention exists at the bench.** Ware, reporting Reilly: the instability is inherent rather than a processing failure
4. 4The image silver — colloidal particles of very high specific surface area, oxidising and sulfiding. **Reducible**: gold toning, and more image silver from stronger salting and fuller exposure
5. 5The mount — lignin-cored board, decomposing adhesive, foxing. **Wholly a choice**, and it is made at the moment of mounting

Assembled from Reilly's chapter eleven, the AIC's conservation notes and Ware's section 17.5. The asymmetry is the point: four of the five routes are open to a maker's decision and the fifth is a property of egg white.

How a conservator identifies one

Reilly’s own warning comes first and it governs the rest: these are handmade objects, “process identification of 19th-century print materials depends almost wholly on experience and judgement”, and the guidelines exist “to assist persons who may be unfamiliar with these materials in narrowing the range of possible choices, not in making positive identification of individual prints.”

Test	Albumen	Salted paper	Matte collodion	Gelatin or collodion POP
Surface	Smooth, with a greater or lesser gloss; sometimes a horny texture under a heavy coat	Rough or matte; the surface is the paper’s own	Matte	Generally smoother than albumen, because they have a baryta layer
Layer structure at 80×	Two layers: albumen over paper, with paper fibres visible through the layer in the highlights	One: no layer at all	Layer over baryta	Layer over baryta; fibres not visible
Crack pattern	A fine uniform microcrack network in most prints before 1870; largely absent on burnished prints after 1870	None — nothing to crack	—	Not this pattern
Colour, in original condition	Warm brown, purplish-brown, purple or purplish-black. Seldom black; never green, though severely faded and yellowed examples can have a faint greenish tinge	The same family, and Reilly treats the two together on colour	Often olive-black, from combined gold and platinum toning	Matte gelatin POP often brown from platinum toning alone
Highlights	Yellowed in about 85 per cent of surviving post-1860 examples, and the yellowing is “peculiar to albumen paper” in the way it presents	Can yellow, but not in the same way	—	Can yellow, especially where combined toning-and-fixing baths were used
Support	Very smooth, lightweight rag stock, especially after 1860	Both smooth and porous stocks, various weights	—	—
Format and mount	Overwhelmingly mounted. Carte de visite, cabinet card, stereograph, album page	Often unmounted or matted	Card formats of 1895–1920	Card formats of 1895–1920
Date	1855–1895 dominant; 1860–1885 a period of “reasonable certainty”	1840–1855 dominant	1895–1925	1895–1905

Three of those rows carry more weight than the others. Reilly names two: “the presence of highlight yellowing and the characteristic surface texture of albumen are two of the most readily apparent and reliable indicators that a given print is an albumen print.” The Getty adds the third and it is the sharpest: the two-layer structure, with paper fibres clearly visible under the albumen layer in the Dmin area at about 80× magnification. Albumen paper has no baryta, so the fibres show through; a gelatin or collodion printing-out paper does have one, so they do not.

And one row is a trap. Reilly’s first question is whether the print is a silver photograph at all: “a perfectly intact, unblemished image usually indicates that a print is not an albumen and salted paper print”, but a photomechanical process — a woodburytype, a carbon print, a collotype — could and did mimic the colour and appearance of an albumen print convincingly. Perfect condition on a print that looks nineteenth century is a reason to look harder, not a reason to be pleased.

The date chart is a prior, not a proof. Reilly’s own: 1840–1855 salted papers, 1855–1895 albumen, 1895–1905 gelatin and collodion printing-out papers, and the transitions are genuinely uncertain. He singles out 1860–1885 as the period where “there is a great likelihood that any given silver photographic print of that period was made on albumen paper”, with studio portraits and stereo views especially likely. The Library of Congress’s own chart gives albumen prints as 1850 to about 1900, so the two differ at both ends by about five years, and neither is a boundary you can stand on.

Storage, and where the guidance disagrees

Five conservation sources give a recommended environment for these objects and no two of them give the same numbers. That is not a scandal; it is what a young and honest field looks like. Read them as a band rather than as a specification.

Source	Temperature	Relative humidity	Notes
Reilly, chapter ten	18 to 20 °C	35 to 45 %	“The single most important consideration in print storage is that temperature and humidity should not undergo sudden drastic changes”
Reilly’s later study, as Ware reports it	not exceeding 18 °C	30 to 40 %	Ware adds that these “would seem to hold good for salted paper prints also”
AIC conservation wiki	68 °F (20 °C) ± 2 °C of drift	50 % ± 5 % over 24 hours	Stated as an ideal, with the tolerance on the variation rather than on the level
Library of Congress	18 °C is the highest recommended extended-term temperature for black-and-white prints; daily fluctuation greater than ± 2 °C to be avoided	30 to 50 %	General guidance for black-and-white prints; the LOC’s albumen entry is a date range and carries no separate environment
Image Permanence Institute	Room, cool, cold or frozen, chosen by medium	30 to 50 %, for all four of those environments	The quick reference gives no albumen-specific recommendation at all: its media list does not separate nineteenth-century printing-out papers from photographic paper prints in general

What all five agree on is more useful than where they differ: cool, on the dry side of the middle, and above all stable. Reilly’s reasoning is the one to keep, because it explains why the level matters less than the constancy: temperature and humidity “govern the rate of all possible destructive reactions that prints might undergo. The higher the temperature and humidity, the faster all the various mechanisms of deterioration will operate.” And the humidity number carries extra weight for albumen specifically, because Ware reports Reilly’s finding that the instability is “greatly accelerated at high relative humidities” — so the one variable that acts on the material’s own inherent failure mode is the one you can control with a dehumidifier.

Two general rules from the Library of Congress apply here without modification. Store prints larger than 8 × 10 in horizontally. And do not put a print with a flaking or sensitive surface into plastic, “because damage that can occur from the static charge lifting media or binders off the photographic support” — which for a crazed albumen print is a real risk rather than a theoretical one, since the Getty’s own ELISA sampling method works precisely by dry-swabbing loose material off the raised edges of the cracks.

⁇Buffered or unbuffered enclosures: an argument that turned around

This is the one place where storage guidance for albumen genuinely differs from guidance for other prints, and the reason it differs is a chain of reasoning that has since been partly retracted.

The position photograph conservators adopted. Enclosures for photographs should be unbuffered, close to pH 7, and alkaline-buffered papers should not be used. Ware records that this view “appears to stem from a statement in 1980 that ‘Papers with a pH of over 8 are detrimental to photos, both black and white and color …’, but no references to experimental evidence are cited.” An unbuffered wrapping paper, Atlantis Silversafe Photostore, was developed for the purpose and introduced in 1982.

Why albumen looked like the strongest case for it. Reilly showed that the Maillard reaction degrading the albumen binder is accelerated by alkali. That is a specific, mechanistic reason to keep alkali away from an albumen print, and Ware records that it “gave rise to the tentative recommendation in 1982 against the use of buffered enclosures with albumen prints.”

And then Reilly tested it. Ware: “subsequent experiments by Reilly on albumen prints incubated in contact with calcium carbonate buffered paper, yielded no detectable deterioration within the duration of the experiment; as the author remarks, ‘More work should be done, but it seems clear that calcium carbonate buffering is not by itself a major threat to albumen prints’.” Ware’s own chemical note explains why the effect is smaller than the mechanism suggests: calcium carbonate “is not a strong alkali — its saturated solution has a pH of only 9.4 owing to the low solubility.”

Where the guidance stands now. The Library of Congress’s enclosure section recommends buffered paper at pH 7 to 9.5 for “acetate and nitrate films, platinum prints, silver prints, chromogenic prints, and prints mounted on acidic boards”, and reserves unbuffered paper for “cyanotypes and architectural drawings” — citing Ware’s own 1999 book. An albumen print is a silver print, and one that is very probably mounted on an acidic board.

What the course says. The specific case against buffered enclosures for albumen is weaker than it was, on the originator’s own later evidence, and the case for them on a mounted print is real — the mount is acidifying, and an alkaline reserve is what neutralises acid migrating out of it. Where a reader wants the conservative choice, an unbuffered enclosure is not a mistake and has been the field’s default for forty years. What the course will not do is repeat the 1982 recommendation without mentioning that its author’s own subsequent experiment did not support it.

And the contrast that makes the whole argument intelligible. Ware sets three processes side by side. A cyanotype must have an unbuffered enclosure, because Prussian blue is hydrolysed by alkali and the image is destroyed — that one is settled, and it is why Part XXI makes such a point of unbuffered paper. A platinotype can only benefit from a buffered enclosure, “because their besetting problem is acid embrittlement of the paper substrate; there is no protein binder layer to decompose, and the imaging substance is quite inert.” An albumen print sits between them, and that is the honest position rather than a failure to decide.

What a maker today should do differently

Eight decisions, each with the reason attached and each traceable to something on this page.

Tone thoroughly, and know what kind of argument that is. The mechanism is sound, the practitioner consensus is 170 years old, and the comparative evidence from surviving objects supports it — and nobody read for this course has run the controlled trial. Reilly’s practical target is to tone until, by transmitted light, the last traces of warm colour have almost gone.

Put more silver in the print. This is Reilly’s own comparative finding and it is the least-quoted piece of preservation advice in his book: prints of the 1860s were more heavily salted, carried more image silver, and have survived better than the weakly salted, scantily exposed prints of the 1890s. So keep the chloride at 1.5 per cent or above, keep the bath at 10 per cent or above, and print deep rather than thin.

Fix in two fresh baths, at a conservative capacity, and wash properly. Residual thiosulfate is the one decay route that is entirely yours. Reilly’s conservative figure is 10 to 15 8 × 10 prints per litre against a literature suggestion of 150, and the reason is that printing-out papers carry an unusually high silver load. Then run the residual hypo and silver tests rather than assuming.

Keep acid and hardener out of the sequence. No acid stop bath, no acid hardening fixer. The mechanism is the sel d’or failure: acid decomposes thiosulfate and liberates sulfur, and this material has more silver waiting for sulfur than any other print in the course.

Control the coating thickness, and understand the trade. A double coat buys gloss and depth, and costs permeability in toning and fixing, and adds curl. Reilly notes that heavily coated papers are “brittle and hard to tone”, and thick coatings need a longer final wash — 40 to 50 minutes rather than 30.

Choose the mount as if it were part of the print, because it is. Museum board of 100 per cent cellulose, acid-free; cream or ivory rather than bright white, which is what museums choose “in order to harmonize with the warm colors of print-out silver images”; wheat or rice starch paste and nothing else if you paste; hinges or corners in preference to pasting; never dry mounting. And consider not mounting at all: a polyester sleeve with a two-ply rag board, or a window mat, both hold the print without being attached to it.

Store it cool, dry and stable, in the band the four sources above agree on, and treat the stability as more important than the exact numbers.

And record what you did, on the object. This is the recommendation that costs nothing and is worth most, and it comes straight out of the identification section above. A conservator in 2126 looking at your print will face exactly the questions this page has been unable to answer about 1876 prints: what chloride, what bath, single or double coated, which toner and for how long, what fixer and what wash, what mount and what adhesive. Write it in pencil on the back near the edge, which is where Reilly says any identifying information belongs, or on the mount. Any facsimile or reproduction print should also be indelibly marked as such — Reilly’s instruction, and it is about honesty rather than permanence: “any and all facsimile prints, however, should always be indelibly marked as such, with an embossed stamp or other method of identification.”

The atlas entry

Filling the albumen row from your own print

1. 1Measure, do not estimateExposure scale and maximum density from a step tablet printed alongside the picture, read on the densitometer, on the dry print. Both are meaningless unless the test reached maximum density: check that the step under the densest patch matches the print margin
2. 2Describe the surface in raking light, against your salt printGloss is the property this process exists for, and the comparison is more informative than any adjective on its own
3. 3Describe the colour twice — toned and untonedThe untoned control print you fixed and dried alongside is the only thing that tells you what the gold did
4. 4Record the cost honestly, as a floorSilver dominates and most of the other rows have no sourced price. Say so; a floor that names its gaps is a datum and a made-up total is not
5. 5Compare each field with a period exampleThen go and look at a real carte de visite under a loupe, and fill a second column: highlight colour, crack network present or absent, fibres visible through the layer, mount and its condition
6. 6And say where your print differs, and whyA modern albumen print made on modern paper with a fresh two-bath fixer is not the same object as an 1875 studio portrait. The differences you can name are what you have actually learned

The row feeds Part XXV's comparison atlas assignment, which is the only place in the course where one negative is printed across every process by one person in one room.

The assignment for this page is that second column. Complete the process atlas entry for albumen from your own print, and compare every field with the conservation description of a period example — from a real object if you can reach one, and from the descriptions on this page if you cannot. The fields where a modern print and a period one will differ most are the ones worth writing about: your highlights are white and theirs are not, your layer is uncracked and theirs probably is, your mount is museum board and theirs is a lignin-cored card, and your print has been fixed in two fresh baths by somebody who knew why.

Recap

An albumen print is a very thin, glossy, warm-toned image in a protein layer on a very thin sheet of paper, almost always mounted, often burnished and frequently retouched. Almost every surviving example has yellowed in the highlights, and the best available evidence says that is inherent to the material rather than a processing failure: the protein binds silver at sulfur-bearing side chains that no fixer reaches, and the albumen’s own sugar attacks its own protein by the Maillard route. Both are accelerated by heat and humidity. Neither can be prevented at the bench. What can be prevented is the other half — residual thiosulfate, an acid sequence, a thin silver deposit, an untoned image and a lignin-cored mount — and those are the maker’s decisions.

The layer cracks because a brittle film is glued to a support that moves with the humidity, and the crack network is the process’s clearest identification signature except on the burnished prints made after 1870, where it may be absent altogether. Gold toning improves survival on mechanism, on consensus and on comparative observation of surviving objects, and not yet on a controlled trial. Storage guidance from five conservation sources spans 18 to 20 °C and 30 to 50 per cent relative humidity, and what they agree on is that stability matters more than the level. And the argument about buffered enclosures — the one place where albumen guidance differs from the guidance for other silver prints — turns out to rest on a mechanism its own author later tested and could not confirm.

Check your understanding

Question 1. Name the two most common forms of deterioration in surviving albumen prints and give the physical or chemical cause of each.

• Highlight yellowing, from silver bound to the protein converting to silver sulfide and from the Maillard reaction; and cracking of the binder, from humidity-driven dimensional mismatch between a brittle layer and a moving support
• Fading and silver mirroring, both from residual thiosulfate
• Cracking of the mount and separation of the print, both from adhesive failure
• Highlight yellowing from residual hypo, and fading from light exposure

Show the answer and why

Answer: Highlight yellowing, from silver bound to the protein converting to silver sulfide and from the Maillard reaction; and cracking of the binder, from humidity-driven dimensional mismatch between a brittle layer and a moving support

The fourth option is the popular answer and it is wrong twice over. Residual hypo is one contributor to yellowing and not the main one: Reilly and the AIC name silver bound to sulfur-containing side chains, and Reilly's own accelerated ageing added the Maillard protein-sugar reaction, both of which occur in a perfectly processed print. And Pretzel and Martin found nearly as much change in dark-stored albumen prints as in exhibited ones, so light is not the dominant driver of fading either.

Question 2. A print you are examining has a smooth glossy surface, warm purple-brown image colour, slightly yellow highlights, is mounted on a card — and shows no crack network at all under magnification. What does the absence of cracking tell you?

• It is not an albumen print; the crack network is diagnostic
• It is probably a gelatin printing-out paper print
• Very little on its own: the Getty found that albumen prints made after about 1870 and burnished or heat treated may show no visible surface cracks
• The print has been treated by a conservator

Show the answer and why

Answer: Very little on its own: the Getty found that albumen prints made after about 1870 and burnished or heat treated may show no visible surface cracks

Presence of a uniform microcrack network is strong evidence for albumen; absence is weak evidence against it, because burnishing and heat treatment suppress it and those were applied to exactly the glossiest and most heavily coated prints. The test that does discriminate here is the Getty's two-layer check: examine the highlights at about 80× and look for paper fibres visible through the layer. Albumen has no baryta, so they show; a gelatin or collodion printing-out paper has one, so they do not.

Question 3. Which statements about the evidence for gold toning improving the survival of albumen prints are accurate? Select all that apply.

• There is a sound mechanism: gold and platinum resist sulfur and oxidation, and a layer of gold shields the silver beneath
• The Photographic Society of London's Fading Committee recommended thorough washing and gold toning in 1855
• Reilly observed that more heavily salted prints of the 1860s and 1870s have on average survived better than weakly salted prints of the 1890s
• A controlled comparison of toned and untoned albumen prints under measured ageing conditions has established the size of the effect

Show the answer and why

Answer: There is a sound mechanism: gold and platinum resist sulfur and oxidation, and a layer of gold shields the silver beneath, The Photographic Society of London's Fading Committee recommended thorough washing and gold toning in 1855, Reilly observed that more heavily salted prints of the 1860s and 1870s have on average survived better than weakly salted prints of the 1890s

The fourth is the one no source read for this course reports. That does not make toning a bad idea — the mechanism is sound, the consensus is old and well founded, and Reilly's observation of surviving populations supports it. It makes it an idea supported by mechanism, consensus and observation rather than by a trial, and knowing which is which is the difference between using evidence and quoting it. Note also that Reilly's third observation is about silver quantity rather than about gold: more image silver, from stronger salting and fuller exposure, is a separate lever.

Question 4. You store your own albumen print unmounted in a calcium-carbonate-buffered enclosure for twenty years. What do you expect, and how should each part of your answer be classified?

• Severe yellowing from alkaline acceleration of the Maillard reaction — published
• The print will curl badly (published); buffer-driven damage is not established, because Reilly's own incubation of albumen prints against carbonate-buffered paper showed no detectable deterioration (published); and any specific twenty-year outcome is inferred
• No change at all, because calcium carbonate is insoluble — measured
• Complete fading, because unmounted prints have no protection — inferred

Show the answer and why

Answer: The print will curl badly (published); buffer-driven damage is not established, because Reilly's own incubation of albumen prints against carbonate-buffered paper showed no detectable deterioration (published); and any specific twenty-year outcome is inferred

Three separate claims with three different warrants, which is what the question is testing. The curl is published and is a property of a thin sheet under a thick albumen layer — the Getty describes unmounted albumen prints curling into tight rolls. The buffer question is published on both sides: alkali does accelerate the Maillard reaction, and Reilly's subsequent incubation test against carbonate-buffered paper found no detectable deterioration, with his own conclusion that carbonate buffering "is not by itself a major threat to albumen prints". And the twenty-year outcome for your specific print in your specific room is an inference, because nobody has run that experiment on a print like yours.

Question 5. Ware's table gives the threshold exposure of albumen prints as 80,000 to 2,880,000 klx·s — a factor of thirty-six. What does he say that spread means?

• That the measurement technique is unreliable
• That albumen prints vary between makers, so an average is meaningless
• That less stable "mavericks" can lurk in any population of historic photographs, because of uncertainties about the quality of processing
• That reciprocity fails for albumen prints as it does for cyanotypes

Show the answer and why

Answer: That less stable "mavericks" can lurk in any population of historic photographs, because of uncertainties about the quality of processing

Ware's own gloss on the range. The practical consequence for a collection is uncomfortable: you cannot look at an albumen print and tell which end of that range it sits at, so a display policy set by the average is a policy that will fade some fraction of the collection. The fourth option inverts what he says — he warns that reciprocity certainly does not hold for cyanotypes, while it "may be generally true for silver images", which is why the hours-of-display conversion on this page is offered for albumen at all.

Check my answersClear