Experiment: The Exposure Scale, Maximum Density and Contrast of a Cyanotype

Every page of this part so far has quoted the same number at you: a classic cyanotype has an exposure scale of about 0.9, three stops, and needs a negative built for it. That figure is Mike Ware’s, measured on his paper, his citrate, his lamp and his densitometer, and published without the courtesy of a range. It is the best figure anybody has. It is also almost certainly not yours, because the chemistry lesson has already shown you why the number is what it is — most of the image washes off, and how much washes off depends on a sheet of paper Ware never saw.

This session measures it. It costs one coating batch, one step tablet, an afternoon and a following morning, and it produces the one number the rest of the alternative-process cluster is built on.

Purpose

To measure the exposure scale and maximum density of the classic cyanotype on your own materials, and then to find out which of four candidate variables actually moves them.

The hypothesis, stated so that the data can refute it: the classic cyanotype’s short exposure scale is not a property of its photochemistry but of image loss during wet processing, so the variables that change how much Prussian blue stays in the paper — the paper itself and the processing bath — will move the scale and the maximum density more than the sensitiser ratio will. That prediction is Ware’s own mechanism turned into something falsifiable: he lists peptisation of the image substance as the fifth shortcoming of the process and the short scale as the seventh, and connects them explicitly, saying that the loss falls hardest on the high values, truncates the tonal scale and confers an artificially high contrast. If he is right, a doubling of the ferricyanide will do less to your curve than changing from a gelatin-sized paper to an alkyl-ketene-dimer-sized one.

The control is run 0: one mix, one paper, one drying time, one dose, plain-water processing, read dry the following morning. Every other run differs from it in exactly one respect, and run 0 is printed twice — once at the beginning of the session and once at the end — because a control measured once is an assumption and a control measured twice is a measurement of your own drift.

The single variable changes from run to run and never within one:

Run	What changes from run 0	Everything else
1	The ratio of iron salt to hexacyanoferrate in the mix	Identical
2	The relative humidity of the paper at the moment of exposure	Identical
3	The first processing bath: water, weak citric acid, or water with peroxide	Identical
4	The paper	Identical

Four variables is four experiments, not one, and the page says so rather than pretending a single afternoon settles four questions. What holds them together is the control and the instrument: one coating session, one tablet, one operator, one densitometer anchoring, so that the four results are at least comparable with each other even where none of them is comparable with anybody else’s.

Learning objectives

By the end you will be able to:

• state an exposure scale as a number with the two density endpoints that define it, and explain why a figure quoted without them is not a measurement;
• expose a calibrated step wedge onto a hand-coated sheet so that both ends of the process’s scale are on the tablet rather than off it, and predict how many steps will land on the curve before you expose;
• read reflection density on a print whose absorption peaks in the red, say which spectral channel you read it in, and say what that choice does to the numbers;
• design four one-variable comparisons that share a control, and identify the confound that a naive version of each would carry;
• distinguish a change that moves a whole curve sideways (speed) from one that changes its slope (contrast) and from one that only lifts its top (maximum density);
• build an error budget for a printing experiment from terms you have measured rather than terms you have assumed, and say which of your findings survive it;
• state what is genuinely available as a contrast control in the classic process, and why the traditional answer is not on the list in this course.

Prerequisites

• Mixing, coating and printing a classic cyanotype — the bench technique, the two stocks, the coating rod, the wash endpoint and the dose bracket for your own lamp. This page assumes every one of them and repeats none.
• Experimental design for the darkroom — the one-variable rule, the control sample, and writing the prediction down first.
• The characteristic curve — toe, straight line, shoulder, and what a slope means.
• Paper sensitometry — reflection density, and what an exposure range figure such as ILFORD’s ISO Range actually is.
• Calibrating the densitometer — arms 9 to 11, the reflection anchors and the three repeatability figures, which you need before the first reading here rather than after.
• What these processes need — the published exposure scales with their measurement conditions, and the paper chemistry that variable four is about.

Safety classification

Level B, and for the same reason as the coating lab: one substance and two things done to it. The classification rubric takes the level from the highest of the chemicals present, and that is potassium ferricyanide at B. Ammonium iron(III) citrate and Prussian blue are Level A. The two operations that keep the letter are weighing dry solids, which this session does more of than the lab did because it makes three mixes rather than one, and working in a room that also contains acid — here deliberately, because run 3 puts a citric acid bath on the bench.

One step is declared above the baseline and it is a step, not a session. Nothing here is Level C. The chromium ruling removes the one operation that would have been: the traditional contrast additive for this family is a dichromate, and this course does not use chromium(VI) at any level. That exclusion is argued as chemistry rather than asserted, under Analysis.

⚠What is not a hazard here, and the two conditions that would change that

There is no hydrogen cyanide, and the reason is a pair of conditions neither of which occurs. The cyanide is bound to iron in a low-spin complex. Ware’s own qualification of the frequently repeated warning is that the heating has to be very strong and the acid very concentrated, and this session heats nothing and uses citric acid at 1 per cent w/v. That is not a claim that the risk is absent in general; it is a claim that these two conditions create it and this procedure creates neither, which is exactly why the acid rule below is absolute rather than proportionate.

There is no chromium(VI) on the bench, in any mix or any bath, and no dichromate re-oxidation bath — a treatment Ware records as having been recommended in the past and rules out in the strongest terms he uses anywhere in the book.

Nothing is heated and nothing is concentrated. The strongest solution made is a 10 per cent w/v ferricyanide stock. The peroxide arm uses household 3 per cent stock diluted tenfold.

The hazard people forget is the light, which is why this page declares requiresUV, and it is worse here than in the coating lab because a comparative session tempts you to open the box and look.

Hazards

Hazard	Where it arises	Control
Dust of either solid, to eye and airway	Weighing for three mixes instead of one	Weighing SOP, splash goggles, no sweeping motions, a P2 filter if dust cannot be avoided
Eye and skin irritation from solutions	Mixing, coating, every tray, and the humidity box	Nitrile gloves and eye protection throughout; the citrate stains skin yellow and the stain goes blue in the wash
Hydrogen cyanide, only if the acid rule is broken	The waste container, not the tray	The rule below. Ferricyanide and concentrated acid never share a room, a funnel or a bottle
Ultraviolet to eye and skin	Every minute the source is energised, and every minute of a sunlight run	The enclosure and interlock from Part XVI. ICNIRP put personal protection last in the hierarchy of controls and so does this page: you do not open the lid to check
Repeated small ultraviolet exposures	A comparative session has more exposures than a printing session	Plan the runs on paper first so the lid opens once per run, not four times
A cut hand	Cutting one sheet into many strips	Steel rule and mat, done dry, before any tray is filled
Stains to bench, clothing and sink	Coating and processing	Blotting paper under everything, a tray under the trays

⚠The acid rule, restated because this session brings acid to the bench on purpose

Run 3 uses a 1 per cent w/v citric acid first bath. That is a deliberate, dilute, weak acid, and it is the only acid in the room.

It has its own graduate, its own tray and its own waste bottle, and none of them is ever used for anything else. The failure mode is not the tray. It is a part-used waste container topped up by somebody who did not read the label, and that is why the rule is written as an absolute rather than as a concentration threshold. Ware’s own warning about stronger acids on this formulation is separate and also worth obeying: on the classic sensitiser, anything stronger than about 1 per cent citric causes blue chemical fog in the highlights, so the concentration limit is a picture-quality rule as well as a safety one.

Required PPE

Nitrile gloves from the first weighing to the last bottle cap. HSE’s COSHH essentials sheet for manual photographic processing takes single-use nitrile at 0.2 mm as splash protection where the safety data sheet gives nothing more specific; glove selection has the reasoning. A glove that has been in the citrate does not go into the ferricyanide jar.

Chemical splash goggles at the balance and while making up stocks; safety spectacles with side protection at the trays, because the riskiest moment is a powder being tipped rather than a 5 per cent solution being poured. An apron, because the citrate stain on cloth is permanent.

Sun protection, not chemical protection, if you take the daylight route. Covered arms, a hat, shade for the frame, eyes off the sheet: four runs outdoors is a long time in the open.

Ventilation

The control at this bench is dust extraction at one step and separation from acids everywhere else. It is not airflow, because nothing in this session evolves a vapour at room temperature — neither salt is volatile, the 1 per cent citric acid bath gives off nothing, and the diluted peroxide decomposes to oxygen and water.

Work in a room with the ordinary through draught HSE ask for wet photographic work, on a lipped wipeable surface, and run the ventilation check first. Shut the window while you weigh — a draught at the balance lifts powder — and open it afterwards.

Materials

Item	Quantity	Note
Unbuffered 100 per cent cotton paper, hot-pressed, 8 × 10 in	6 sheets	The session’s primary paper. One sheet yields the strips for runs 0 to 3 comfortably
A second unbuffered paper of different sizing or surface	2 sheets	Variable four. A gelatin-sized sheet against an alkyl-ketene-dimer-sized one is the most instructive pairing, because that is the comparison Ware himself ran
Calibrated transmission step tablet	1	See Preparation. The whole experiment rests on this piece of film
Clean plate glass, about 200 × 250 mm, edges taped	1	Holds the tablet in contact where the frame is oversized
Card mask with a window the size of the tablet	1	Registers the tablet identically on every strip
Self-healing mat, steel rule, sharp knife	1 each	All cutting done dry and once
Sealable box, about 2 to 5 L, with a rack	1	Variable two. The humidity chamber
Saturated sodium chloride solution with excess solid	about 200 mL	Holds about 76 per cent RH at room temperature. Ammonium chloride gives about 80 per cent if you want the wider separation
Silica gel, indicating, in a second sealed box	about 200 g	The dry arm of variable two
Hygrometer reading to about ±3 per cent RH	1	Records what the boxes achieved, which is not the same as what the table says
Blotting paper, masking tape, a soft pencil	—	Every strip is numbered on the back before it is exposed
Distilled or de-ionised water	about 4 L	Stocks, baths and the final rinse if your tap water is hard or above pH 7
Brown glass bottles, 100 mL	4	Two stocks for the primary mix, two for the ratio arm
Waste containers, labelled, screw-capped	3	Sensitiser residue; wash and bath water; the acid bath alone

Chemicals

Chemical	Quantity	Form
Ammonium iron(III) citrate, green	30 g	Solid. Record supplier and lot. Iron content varies from batch to batch and the whole point of this page is that it is your batch being measured
Potassium ferricyanide	15 g	Solid, GPR grade or better
Citric acid	10 g	Solid, for the 1 per cent w/v bath of run 3
Hydrogen peroxide, 3 per cent household stock	60 mL	Diluted to about 0.3 per cent for the re-oxidation arm
Thymol	a few crystals	Optional, to keep mould off the citrate stock if it is to be kept
Distilled water	about 4 L	Stocks, baths, rinses

No formula is printed here. The mixes are the ones on the classic cyanotype sensitiser page, at the strengths and in the order that page gives, and the ratio arm varies only the volumes in which its two stocks are combined. If you find yourself writing a new sensitiser to see what happens, you are running a different experiment from this one, and it belongs on the variants lesson rather than here.

Equipment

A ultraviolet unit that has been commissioned — leak, interlock, temperature, uniformity map, dose series — or a clear stretch of sky and the outdoor session SOP. A contact printing frame with a hinged back.

A densitometer with a reflection head, which is the one thing this experiment cannot substitute its way around, and which needs the spectral discussion under Preparation before you switch it on.

A balance reading to 0.01 g, graduated syringes or pipettes dedicated one per solution, a coating rod, four trays, a timer, a thermometer, a hygrometer, and a drying rack that will hold twenty numbered strips without their touching.

Estimated cost

Band ££, and every pound of it is capital you already own or will not buy for this page. The consumables are two jars of solid, some paper and a bottle of household peroxide. What lifts the band above the coating lab’s £ is the step tablet and the densitometer: this is the first cyanotype page that cannot be run without both. The planner holds the dated figures and, more usefully, the gaps.

Estimated consumables cost

Every priced row is the planner’s own dated UK figure and every quantity comes from the Materials and Chemicals tables above. Most rows have no number, and that is the finding rather than an omission.

Consumed	This session	Sourced price	Cost this session
Nitrile gloves	3 pairs	£6.64 to £14.99 per box of 50 to 100	£0.40 to £0.90
Unbuffered cotton paper, 8 × 10 in, primary	6 sheets	Not priced. Hot-pressed cotton paper is one of the gaps src/data/prices.json already names	—
Second paper, 8 × 10 in	2 sheets	Not priced, same gap	—
Ammonium iron(III) citrate, green	about 30 g weighed; roughly 6 g leaves in the trays	£17.99 per 230 g of a cyanotype-branded pack; the grade is not stated on the listing, and this run wants the green	£2.35 weighed, about £0.47 of it gone
Potassium ferricyanide	about 15 g weighed; roughly 3 g leaves in the trays	£17.99 per 230 g	£1.17 weighed, about £0.23 of it gone
Citric acid, solid	10 g	£10.00 per 250 g of the monohydrate	£0.40
Hydrogen peroxide, 3 per cent, pharmacy	60 mL	Not priced	—
Distilled water	about 4 L	Not priced	—

The priced rows come to about £4.32 to £4.82, and that is a floor and not a total: four of the eight consumables carry no sourced price, and both papers are among them. What the table does establish is the same ratio the coating lab found, and most of that subtotal is stock rather than consumption — about £3.52 of the two salts is weighed out and roughly seventy pence of it actually leaves in the trays, while the eight sheets of rag paper are gone at the end of the session and still carry no figure.

Equipment is deliberately outside the table: a tablet and a densitometer are not consumed, and a page that counts them has stopped measuring what the consumables calculator needs. The tablet is also the one item here a named gap already covers — the price file could not price a 21-step transmission wedge at all.

Waste streams

Three containers, and the third exists only because of run 3.

• Sensitiser residue — the cups, the blots, the rod rinse. Small in volume and highest in concentration.
• Bath and wash water — dilute unreacted citrate and ferricyanide plus the peptised Prussian blue that clouds off the sheet. On this page that cloud is data rather than dirt: it is the image substance whose loss the hypothesis is about, and run 3 will make you look at it four times.
• The citric acid bath, on its own, in its own labelled bottle, never topped up with anything and never used as the receiver for a rinse.

Label every container with contents and date under the labelling SOP and the general chemical waste SOP. The peroxide arm’s bath goes with the wash water; it is a decomposing oxidiser in a bottle of iron salts, so leave it loosely capped for a day before it is moved.

Alternative route

Without an ultraviolet unit, print in the sun, and it changes one thing about this page. Sunlight is the honest alternative for cyanotype and Part XVI’s unit page argues the case; every cyanotype made before electric light was made this way, and the sun is nearly a point source, so it gives the sharpest contact print of anything available. What it costs here, specifically, is the one thing this session needs most: a repeatable dose. The ultraviolet content of daylight moves with hour, season, latitude and cloud, so two runs exposed an hour apart are not two runs at one exposure.

The workaround is not a longer clock but a different design: expose run 0 and one variable arm side by side in the same frame at the same moment, so that the comparison is internal to a single exposure and the dose cancels out. That costs you the absolute level — you will not be able to say what dose produced your curve — and keeps every difference, which is what the four variables are about. Record the exposure under the daylight SOP with date, clock time, orientation and cloud, and repeat run 0 on each day you work.

Without a densitometer with a reflection head, one half of the experiment survives and one does not, and it is worth being exact about which. The step-count method works: expose to full maximum density, count the steps you can still distinguish from one another, and multiply by the tablet’s increment. That gives you an exposure scale with a resolution of one step — 0.10 log units on a T3110 — which is coarse but is a real measurement of the quantity the negative has to match, and it is essentially the method Reilly gives for the salted-paper family where no figure has ever been published. What it will not give you is a curve: no toe, no shoulder, no slope, no maximum density, and therefore no way to separate a change of contrast from a change of maximum density. Runs 0, 3 and 4 still produce a usable result. Runs 1 and 2 are predicted to be small effects, and a method whose resolution is 0.10 cannot see a small effect.

Without a transmission tablet there is no version of this page. A hand-cut set of masked strips does not have known density increments, and the whole measurement is a comparison of exposures whose ratios you know. Borrow one; it is the cheapest item in the session that cannot be improvised.

Where none of that is available, the reading half of this page still teaches the thing it exists to teach — what an exposure scale is, why the classic process’s is short, and which lever moves it — and the assignment and the fault clinic later in the part do not depend on your having run it.

Preparation

The measurement, decided before anything is coated

The manifest for this page flagged the reflection-density method as an open question across the whole course, to be settled before the page was written. It is settled here, and the ruling has three parts.

One: reflection density, on the 45/0 head, relative to your white tile. Part XV’s calibration session already establishes what that scale can support — it is internally consistent to a figure you measured, and its absolute linearity has not been established, because no certified reflection step scale was found at a home-lab price. That statement is sufficient for everything on this page, because every claim here is a difference between two prints read on one instrument in one session.

And the precision is yours, not the course’s. The course publishes no precision figure for a reflection density because it has built no head and measured nothing on yours; the number that goes in your log is the sample-replaced repeatability from arm 10. What a professional instrument achieves is worth having as a benchmark rather than as a target: Ware’s X-Rite 310, in diffuse reflectance mode, read to a precision of 0.002 and was recalibrated against a standard density plaque before and after each set of experiments to an accuracy of ±0.01 over the range 0.06 to 1.74. If your own figure comes out an order of magnitude worse than that, the session still works, because everything on this page is a difference — but it decides how large a difference you are entitled to call a result.

Two: read in the red, and say so. Prussian blue’s absorption is a single strong broad band centred near 700 nm — 690 nm for the “soluble” variety and 730 nm for the “insoluble” — which is why the pigment looks blue and why a red channel is the one that sees it. Ware read every density in his fading study on the red channel of an X-Rite 310 reflectance colour head, for exactly that reason. Part XV’s head, as built, carries matched green LEDs, chosen for silver prints. On a cyanotype a green channel is not wrong, but it is reading the shoulder of the absorption band rather than its peak: it will give you lower densities, a compressed scale, and less separation exactly where the interesting differences are.

!The course's ruling on the channel, and what it costs if you cannot follow it

Fit a matched pair of red LEDs to the reflection head for this page, anchor on the tile with them, and record the dominant wavelength and half-width from the datasheet. ISO 5-3 is cited here only for what a spectral condition is: a density is defined by its geometric and spectral conditions together, the spectral response being the product of the detector’s sensitivity and every filter in the path. What you will have is a red-channel density defined by your own instrument. It is not a status density and must never be described as one.

If you only have the green head, run the session anyway and label every figure “green channel”. Your differences remain valid, your exposure scale will read a little short, and your maximum density will read well below Ware’s. What you must not do is compare a green-channel maximum density with his red- channel 1.55 and conclude that your paper is poor.

Three: read dry, the next morning. A cyanotype is not finished when it leaves the wash. The reversed shadows regain density by aerial re-oxidation over several hours and the print deepens as it dries, so a wet reading is a reading of an intermediate state. Read every strip after a full night, in one room, at one temperature, all in one pass.

📓The wet-to-dry shift for cyanotype: the course has no figure, and neither has anybody else

The course searched for a published magnitude for the wet-to-dry density change of a cyanotype and found none. What the sources give instead is the direction and a way to separate the two causes: Ware states that the reversed shadow tones regain their values during wet processing and recover fully during about 24 hours of drying, and that a 0.3 per cent hydrogen peroxide bath makes no ultimate difference to the print densities, only to how soon you see them. His 2022 sheet turns that into a working instruction and names the reason: add peroxide to the first wash bath if the regain must be complete immediately, for example for densitometry.

So this page measures the shift instead of quoting one. Run 3’s peroxide arm has already finished its re-oxidation before it dries, so the difference between its wet and dry readings is drying alone, while run 0’s is drying plus re-oxidation. Two subtractions and you have both, for your paper.

Report your figure as the course’s first, and expect it to be positive: a cyanotype gains density on drying, which is the opposite sign from the silver gelatin dry-down you learned in Part XIX. Two processes, two signs, one word — say which you mean.

The step tablet, and why this page changes the course’s default

The course’s standard article is the Stouffer T2115: 21 steps at a nominal 0.15, half a stop each, to a maximum density of 3.05 on a piece 12.7 × 127 mm. It has been the reference since Part IX. For this page it is the wrong tablet, and the arithmetic says why.

🧮How many steps land on the curve

A tablet spans a fixed range of log exposure. The process spans its exposure scale. The number of steps that carry information is the ratio:

n = exposure scale ÷ step increment

Steps on the curve

At Ware’s figure of 0.9 for the classic process, a T2115 at 0.15 per step gives

n = 0.9 ÷ 0.15 = 6 steps

with the remaining fifteen pinned at maximum density or at paper white. Six points do not define a toe, a straight line and a shoulder; they barely define a slope.

A T3110 — 31 steps at a nominal 0.10 to the same maximum density of 3.05 — gives

n = 0.9 ÷ 0.10 = 9 steps, and 12 at the top of the published range of 1.2.

Half as many again, and it is the difference between a curve and four dots. Ware himself used a T3110 for the New Cyanotype comparisons in chapter 7, and a calibrated Kodak No. 3 tablet of 21 half-stop steps for the fading study in chapter 9.

So this page’s reference is the T3110, and it comes with a caveat that has to be printed. Stouffer state which of their parts are calibrated: the T2120CC and T1530CC are read step by step against NIST Standard Reference Material 38120C on a densitometer conforming to ANSI PH2.19-1986, and the T5100C is sold for calibrating densitometers. The T3110 is not among them. Calibrated and uncalibrated guides come from the same batches and are made with the same control, so a plain T3110’s increments are made to 0.10 and not certified to it; what calibration buys is the recorded reading of your individual piece.

The consequence is exact and belongs in every conclusion you draw: your log exposure axis is nominal. Its zero is arbitrary — you are measuring relative exposure — and its scale carries the maker’s manufacturing tolerance rather than a certificate. That is a Type B uncertainty in the sense Part XV defines, and no number of repeat readings will reduce it. If you own a calibrated part, use it and say so; if you own a T2115 and nothing else, run the session on it, accept six points, and confine your conclusions to maximum density and to the step-count exposure scale, which is what six points will support.

The dose repeatability pre-test, which is the error budget’s largest unknown

Part XVI’s unit publishes a uniformity target — a quarter of a printing-out step across the printing area — and a dose series. It publishes no repeatability figure, because the course has built no box and measured nothing on yours. Since every comparison on this page is between prints made in different exposures, that unmeasured number is the largest term in the budget, and the session starts by measuring it.

1. Cut four identical strips from one coated sheet, from the middle of the sheet, all four adjacent.
2. Expose them one at a time at the same nominal dose: the first from cold, the second after the warm-up your unit’s commissioning established, the third immediately after the second, and the fourth after a twenty-minute pause with the unit off and then restarted with the same warm-up.
3. Process all four together and read one mid-scale patch on each, dry.

The spread of those four readings, expressed in density and then converted to log exposure through the slope of run 0’s curve, is your dose repeatability. Strips 2 and 3 measure the unit; strip 1 tells you what skipping the warm-up costs; strip 4 tells you whether a cold restart is the same as a warm one. If the spread is larger than the effects you are hoping to detect in runs 1 and 2 — and it may well be — then the honest response is to redesign those runs as side-by-side exposures in one frame, not to report the differences anyway.

The cutting plan

One coating batch, four runs, twenty strips

1. 1One coating session, one rod, one drying time — runs 0, 2 and 3 come from a single sheet, so coating thickness is not a variable between them
2. 2The ratio arm gets its own sheet — because it needs a different mix, and a different mix cannot be coated on the same sheet
3. 3The paper arm gets the same mix as run 0 — the only difference is the substrate, which is the whole point
4. 4One tablet, one orientation — the tablet is a piece of film with its own slight unevenness; turning it round makes that unevenness a variable

The layout is the course's own. It is drawn so that every comparison is either within one coated sheet or between two sheets coated back to back from one mix, which is the most that a hand-coating process allows.

Procedure

Six stages. Stages 1 and 6 are the reading, and between them they take longer than all the wet work.

Stage 1 — Anchor the instrument and write the three numbers down (20 minutes)

Fit the reflection head with its red LEDs, anchor on the white tile, read the black trap, compute the reflection ceiling, and write your three repeatability figures at the top of the log before the first sample: untouched, sample lifted and replaced, head refitted. Those three numbers decide which of tonight’s differences are results. Do not skip this because you did it last month; a head that has been off and on is a different geometry until proved otherwise.

Stage 2 — Coat, and hold everything constant that is not a variable (45 minutes)

Mix, coat and dry by the coating lab’s method, with three additions that are specific to a measurement session.

1. One mix, made once, for runs 0, 2, 3 and 4. Mixed classic sensitiser has a short life, so the volume is calculated for the whole session and used within the hour. Record the volume and the clock time at which each sheet was coated.
2. Coat by rod, not by brush, and use the same number of passes at the same speed on every sheet. A brush is a variable with a handle on it.
3. Dry every sheet for the same time in the same place, in the dark, and record the room’s temperature and humidity while they dry. Ware’s note for heat-dried siderotype paper is that it should rest half an hour to an hour before exposure; the simplest way to obey it is not to heat-dry at all today.

Then cut, number in pencil on the back, and lay the strips out in run order before the frame is opened.

Stage 3 — The dose, and run 0 (30 minutes)

Establish the dose that just reaches maximum density and no further. This is not the bracket from the coating lab, which was aimed at a good print; it is the sensitometric criterion, and Ware states it in one line: expose until the first two steps of the tablet are indistinguishable, so that the maximum density the sensitiser can reach has certainly been reached. Anything less and you are measuring a truncated curve; much more and the solarised shadows spread up into the mid-tones and change the shape of the top of the curve.

Expose run 0’s two strips at that dose, with the tablet in the mask, dense end towards the number. Then put one of them aside and expose it at the end of the session instead — the second half of the control.

Stage 4 — The four variables (50 minutes)

Run 1, the sensitiser ratio. Mix the same two stocks in three volume ratios: 1:1 as the control, 2:1 iron salt to hexacyanoferrate, and 1:2 the other way. Coat one strip from each.

!The confound in run 1, and how to remove it if you want the cleaner answer

Mixing two stocks in a different volume ratio changes two things at once: the ratio of the two substances, and the total mass of solid laid down per unit area, since the stocks are not the same strength. A difference you measure is then a difference of ratio or of loading, and nothing in your data separates them.

The clean version makes up three fresh pairs of stocks whose concentrations are chosen so that the total per cent w/v in the mixed sensitiser is the same in all three, and only the split between the two substances changes. Ware’s survey gives you the units to work in: he compares about sixty published recipes by converting every one of them to per cent w/v in the final mixed sensitiser, precisely so that recipes mixing their stocks in unequal volumes can be compared at all.

The version this page runs is the confounded one, because it costs one mix rather than three and the predicted effect is small. That is a deliberate trade and it is written into the conclusion: what run 1 tests is the effect of mixing the same two stocks in different proportions, which is the question a practitioner actually faces, and not the effect of the molar ratio at constant loading, which is the question a chemist would ask. Say which one you measured.

Run 2, humidity at exposure. Two strips from the primary sheet. One goes into the sealed box over silica gel; one into the sealed box over saturated sodium chloride, which holds about 76 per cent RH at room temperature. Ware’s humidifying-tank figures are the ones to work from — ammonium chloride about 80 per cent, common salt about 76 per cent, calcium nitrate tetrahydrate about 55 per cent — with a minimum of half an hour to equilibrate. Record what the hygrometer actually read in each box, expose both at the same dose, and note that the damp strip begins drying the moment the lid closes, so a long exposure partly undoes the treatment. Wipe the frame glass afterwards.

Run 3, the processing bath. Three strips exposed identically, then processed three ways:

• plain water, face down, gently running, to the yellow-clearance endpoint;
• 1 per cent w/v citric acid for one minute as a first bath, then the same water wash;
• plain water with peroxide, made by diluting the household 3 per cent stock about tenfold to roughly 0.3 per cent, half a minute, then the same water wash.

Run 4, the paper. Two strips of the second paper, coated from the same mix as run 0, dried alongside it, exposed at the same dose, processed in plain water.

Stage 5 — Process together, dry together, and re-print the control (25 minutes)

Everything except run 3’s arms goes through one wash, one tray, one endpoint: the yellow gone from the highlights, judged by holding the strip to a bluish light or looking through a blue filter, which is Ware’s own refinement and makes a faint interior stain visible when room light will not. Then stop; this is the one wash in the course where longer is not safer, because the wash is also what removes the image.

Drain, lay every strip out on one rack in one room, and expose the second run 0 strip now, process it alone by the same method, and put it on the same rack. Leave everything overnight.

Stage 6 — Read every step on every strip (45 minutes)

Re-anchor, and read every step on every strip in one pass, in a fixed order, without lifting a strip off the port to re-check a reading: re-seating the sample is the largest of the three repeatability terms and a re-check is a new measurement, not a confirmation of the old one.

Read the steps that are pinned at maximum density too. They move together, so a shift repeated across four adjacent pinned steps is a shift, while a shift at one step is a reading.

Expected observations

On the strips themselves, before any instrument. The classic process’s characteristic look at the correct sensitometric exposure is a strip whose top three or four steps have gone past blue into the pale grey-blue of solarisation, whose middle is a firm blue, and whose light steps are distinctly green before washing. Ware’s three criteria for a correct exposure — highlights green, mid-tones blue, deep shadows reversed — are what you are looking for, and the strip should look over-exposed. If it does not, it is under-exposed.

Between run 0’s two prints, agreement. That is the finding you most want and least often get. A systematic difference between the start-of-session and end-of-session control is drift, and it has three usual causes in order of likelihood: the mixed sensitiser aged, the lamp changed, or the coating on the second strip was not the same as on the first.

In run 1, very little. Ware’s survey of about sixty published recipes found them scattered over a wide range of proportions and concluded that the process appears to work satisfactorily almost regardless of the proportions chosen. What Whitaker reported in 1883, after testing everything from 2 per cent against 20 per cent to 20 per cent against 2 per cent, is the shape to look for instead of a density change: excess citrate speeds printing but shortens the storage life of the coated paper, while excess ferricyanide gives a finer quality but greatly lengthens the exposure — the internal filter again. So the prediction is a curve displaced sideways, not a curve of a different shape.

In run 2, also very little, and this is a published prediction rather than a guess: Ware states flatly that the prevailing relative humidity has very little effect on the results of printing in cyanotype, but that it can profoundly affect some of the other iron-based processes. A null result here is the predicted result and worth reporting as such.

In run 3, the largest effects on the page. The acid arm should show a stronger top end — Ware’s words for the classic process are that the tonal scale may be strengthened by a first bath of very dilute weak acid rather than plain water — and, if the acid is too strong, blue fog in the highlights and a degraded paper white. The peroxide arm should reach its final density at once and then not change: the prediction to test is Ware’s, that the peroxide makes no ultimate difference to the print densities.

In run 4, the second largest, and possibly the largest. Ware’s whole account of why the classic process behaves as it does is about absorption and retention by paper fibres.

What is happening chemically

Nothing on this page is new chemistry. What is new is that each variable acts on a different step of a sequence the chemistry lesson sets out in full, and knowing which step tells you which part of the curve to expect the effect in.

The ratio acts on the photochemical step and on nothing else. All the light-sensitivity belongs to the iron(III) citrate complex; the hexacyanoferrate takes almost no part in the photochemistry but absorbs strongly in the same near-ultraviolet band, so it works as an internal filter and slows an exposure it does not drive. More ferricyanide is therefore more filter and a longer exposure. More citrate is more absorber and a shorter one. Neither changes what happens after the light stops, which is why the prediction is a sideways shift — a speed change — rather than a change of slope.

Humidity acts on where the sensitiser is and how mobile the ions are. In the print-out noble-metal processes water in the paper is part of the reaction and humidity is a genuine contrast control, which is why Ware’s platinum and palladium tables list exposure range against relative humidity. In cyanotype the second step needs no light and no particular water activity: iron(II) meets hexacyanoferrate(III) and the electron transfers. That asymmetry is the mechanism behind the published claim, and it is the reason this page tests the claim on the classic process rather than importing Ware’s New Cyanotype drying measurements, which are a different formulation with a different iron complex and a different Prussian blue. Do not transfer a number between processes because they share a colour.

The processing bath acts on the last step, and on the classic process that step is where the contrast is set. Two things happen in the tray at once. The unreacted iron and hexacyanoferrate dissolve away, which must happen or the sheet is still light-sensitive; and part of the image peptises and leaves with them. The classic sensitiser’s Prussian blue is the potassium-containing “soluble” variety, and the loss is proportionally greater from thin deposits than thick ones — so it does not merely weaken the print, it shortens the scale from the top, collapsing the high values towards the paper white. Ware’s own summary is that this confers an artificially high contrast. A weak acid first bath reduces the loss; a long wash in slightly alkaline water increases it, and hard water damages the blue outright.

And the paper decides how much of that loss there is, because the pigment is only retained where the sensitiser reached the interfibrillar space of the surface fibres rather than sitting in the coarse pores. Sizing chemistry, surface and porosity therefore set the retention, and retention sets the top of the curve. That is why the hypothesis puts paper and bath ahead of ratio.

The peroxide arm is not a fifth variable but a clock. Aerial re-oxidation of Prussian white to Prussian blue happens over hours; peroxide does the same conversion in half a minute. If Ware is right that the endpoint is identical, then the peroxide strip and its plain-water twin will agree the next morning — and the difference between their wet readings is a measurement of how far the air had got by the time you first looked.

Data to record

Three records. Use the lab notebook format, the batch record SOP and the curve plotting sheet.

The session log

Chemicals: substance, supplier, lot number, per cent w/v of each stock, the volumes mixed, the clock time of mixing and of each coating. Paper: maker, product, surface, weight, sizing if the maker states it, batch, and which side. Coating: rod or brush, passes, volume per sheet, area covered, and the two divided. Drying: method, room temperature, relative humidity, and time from coating to exposure for every strip. Light: unit, array height, warm-up, lamp hours, dose; or, outdoors, date, clock time, orientation and cloud. Baths: composition, temperature, time, and the wash endpoint judged how. Instrument: head, LED dominant wavelength and half-width, tile serial, date of last anchoring, reflection ceiling, and all three repeatability figures. Tablet: maker, part number, serial if it has one, and calibrated or not.

The measurement table

Column	What goes in it
Run and strip	0a, 0b, 1a, 1b, 2dry, 2damp, 3water, 3acid, 3peroxide, 4a, 4b
Step	1 to 31
Nominal relative log exposure	(step − 1) × 0.10, on a T3110. Nominal, and the log says so
D wet	Blotted, straight from the final wash — marked indicative, because a wet surface has a sheen a dry one does not
D dry	The same patch the next morning
Δ	Dry minus wet, signed
Flag	Pinned at Dmax, pinned at base, or on the slope

The error budget, built from your own numbers

Six terms, and this page’s job is to make you write a figure beside each rather than a shrug.

Term	Type	Where your figure comes from
Dose repeatability of the ultraviolet unit	A	The four-strip pre-test in Preparation. The course publishes no figure and cannot: it has measured nothing on your unit
Coating unevenness	A	The spread across strips 1 to 8 of run 0’s dose, if you exposed them all; otherwise a bound from the uniformity map
Drying time before exposure	B	The spread of times in your log, and whatever run 2 tells you about how much moisture matters
Wet-to-dry reading difference	A	Run 3’s peroxide arm against its plain-water twin. No published figure exists
Reading repeatability	A	Part XV, arm 10, sample-replaced figure
Tablet increment	B	Nominal 0.10 unless you bought a calibrated part. Irreducible by repetition

Add them by Part II’s rule and quote the bound; compute the root-sum-square as well if you like, but label which rule produced which number and never print the smaller one bare.

Analysis

First, fix what you may claim, from run 0’s two prints. If the start and end controls agree within your sample-replaced repeatability, use that repeatability as the threshold for everything else. If they differ by a constant offset, subtract it and say so. If they differ by a trend — one control steeper than the other — the session drifted in a way that a single subtraction will not repair, and the honest report is that runs exposed late in the session are not comparable with runs exposed early.

Second, plot every run on one pair of axes, reflection density against nominal relative log exposure, using the curve plotting sheet. Make the two scales equal, so that 0.30 of density is the same length on the page as 0.30 of log exposure: a slope is a ratio of the two axes, and a curve drawn on unequal scales has a gradient you cannot read off it.

What the family should look like, and the three ways a curve can move

• Run 0, control: plain water, primary paper, 1:1
• Run 3, acid first bath: longer scale and higher Dmax
• Run 4, second paper: retention changes the top
• Run 1, more ferricyanide: the same curve, slower

Show the numbers behind this plot

Four curves on one pair of axes, log exposure increasing to the right. The control curve, drawn solid, leaves a base density of about 0.06, stays flat to a log exposure of about 0.35, rises steeply through 0.30 at 0.60 and 0.90 at 0.95, and shoulders over to a maximum of about 1.15 by 1.30, staying flat thereafter. A second curve, dashed and labelled as the acid-developed arm, follows the same path through the toe but continues rising past where the control shouldered, reaching about 1.45 at 1.60 and flattening there, so its shoulder is both higher and further right: its scale is longer and its maximum density greater. A third curve, dotted and labelled as the second paper, is displaced upwards through the mid-tones and reaches a maximum of about 1.40 at a log exposure of about 1.45. A fourth curve, dash-dotted and labelled as the ratio arm, is the same shape as the control but shifted bodily about 0.2 log units to the right, reaching the same maximum density of 1.15 at about 1.50: it is slower, not softer. Two horizontal guide lines are drawn, one at a density of 0.10 labelled base plus 0.04 and one at a density of 1.04 labelled ninety per cent of the control's maximum, and the exposure scale of each curve is the horizontal distance between where it crosses those two lines.

Series	Nominal relative log exposure	Reflection density, red channel
Run 0, control: plain water, primary paper, 1:1	0.00	0.06
Run 0, control: plain water, primary paper, 1:1	0.20	0.06
Run 0, control: plain water, primary paper, 1:1	0.35	0.08
Run 0, control: plain water, primary paper, 1:1	0.50	0.17
Run 0, control: plain water, primary paper, 1:1	0.65	0.38
Run 0, control: plain water, primary paper, 1:1	0.80	0.64
Run 0, control: plain water, primary paper, 1:1	0.95	0.90
Run 0, control: plain water, primary paper, 1:1	1.10	1.06
Run 0, control: plain water, primary paper, 1:1	1.30	1.14
Run 0, control: plain water, primary paper, 1:1	1.60	1.15
Run 0, control: plain water, primary paper, 1:1	2.00	1.15
Run 3, acid first bath: longer scale and higher Dmax	0.00	0.06
Run 3, acid first bath: longer scale and higher Dmax	0.20	0.06
Run 3, acid first bath: longer scale and higher Dmax	0.35	0.08
Run 3, acid first bath: longer scale and higher Dmax	0.50	0.16
Run 3, acid first bath: longer scale and higher Dmax	0.65	0.34
Run 3, acid first bath: longer scale and higher Dmax	0.80	0.57
Run 3, acid first bath: longer scale and higher Dmax	0.95	0.82
Run 3, acid first bath: longer scale and higher Dmax	1.10	1.05
Run 3, acid first bath: longer scale and higher Dmax	1.30	1.26
Run 3, acid first bath: longer scale and higher Dmax	1.60	1.44
Run 3, acid first bath: longer scale and higher Dmax	2.00	1.45
Run 4, second paper: retention changes the top	0.00	0.07
Run 4, second paper: retention changes the top	0.20	0.07
Run 4, second paper: retention changes the top	0.35	0.10
Run 4, second paper: retention changes the top	0.50	0.22
Run 4, second paper: retention changes the top	0.65	0.45
Run 4, second paper: retention changes the top	0.80	0.72
Run 4, second paper: retention changes the top	0.95	0.99
Run 4, second paper: retention changes the top	1.10	1.21
Run 4, second paper: retention changes the top	1.30	1.36
Run 4, second paper: retention changes the top	1.60	1.40
Run 4, second paper: retention changes the top	2.00	1.40
Run 1, more ferricyanide: the same curve, slower	0.00	0.06
Run 1, more ferricyanide: the same curve, slower	0.20	0.06
Run 1, more ferricyanide: the same curve, slower	0.55	0.08
Run 1, more ferricyanide: the same curve, slower	0.70	0.17
Run 1, more ferricyanide: the same curve, slower	0.85	0.38
Run 1, more ferricyanide: the same curve, slower	1.00	0.64
Run 1, more ferricyanide: the same curve, slower	1.15	0.90
Run 1, more ferricyanide: the same curve, slower	1.30	1.06
Run 1, more ferricyanide: the same curve, slower	1.50	1.14
Run 1, more ferricyanide: the same curve, slower	1.80	1.15
Run 1, more ferricyanide: the same curve, slower	2.00	1.15

Constructed to show the three ways a result can look, not measured. No source publishes numerical characteristic-curve data for a classic cyanotype: Ware's Figs 9.1a to 9.1f are printed as pictures on axes running to 1.6 in reflectance density, and the numbers that belong on these axes are the ones you measure. Read the drawing for the distinction it makes and for nothing else — a curve that moves sideways is a speed change, a curve whose shoulder rises is a maximum-density change, and a curve whose shoulder moves right is an exposure-scale change. The curve is drawn to show the shape, not measured from a real material. Your own materials will differ, and measuring them is what the sensitometry part of the course is for.

Third, read the four numbers off each curve, and define them before you read them, because a figure without its endpoints is not a measurement.

🧮The four numbers, worked on constructed figures

Every number in this box is invented to demonstrate the arithmetic. None is a measurement, none comes from a source, and none should be copied into your notebook. The course has not run this session; it publishes the method, and your paper supplies the numbers.

Maximum density. Take the mean of the steps flagged as pinned at the top. Suppose four adjacent pinned steps read 1.13, 1.16, 1.15 and 1.14 — mean 1.15, agreeing to ±0.02, which is a result. One step reading 1.15 would not have been.

The endpoints of the exposure scale. This course uses base + 0.04 at the bottom and 0.9 × Dmax at the top, and says that it chose them rather than pretending they are given. The pair is Ware’s own: his platinum and palladium tables state their exposure ranges as Δ log H from fog + 0.04 to 0.9 Dmax, read in reflectance, and using one convention everywhere is worth more than using a better one on each page. It is not ILFORD’s ISO 6846 definition, so a cyanotype figure measured this way and an ISO Range figure are close cousins rather than the same quantity — which is exactly why the endpoints are printed beside every number. Here they are 0.06 + 0.04 = 0.10 and 0.9 × 1.15 = 1.04.

The exposure scale. Read where the curve crosses each endpoint and subtract:

ES = logH(0.9 D_max) − logH(base + 0.04) = 1.28 − 0.42 = 0.86

Exposure scale from the curve

Interpolate between steps rather than rounding to the nearest one; a step is 0.10 and you can do better than that from a plot.

The slope. Take two points on the straight portion, four steps apart so 0.40 in log exposure. If they read 0.38 and 0.90:

G = (0.90 − 0.38) ÷ 0.40 = 1.30

Average gradient over the straight portion

And now compare, not to Ware, but to a paper you have printed on. A slope of 1.30 with a scale of 0.86 is a hard material: ILFORD’s MULTIGRADE RC DELUXE gives an ISO Range of 90 at filter 2 and 70 at filter 3, so a log exposure range of 0.90 and 0.70. Your cyanotype sits between them. That is the sentence the whole session exists to let you write — and note that it is a comparison of exposure scales, not of maximum densities, because that paper reaches 2.15 and no cyanotype comes near it.

Where the toe and shoulder fall, and one figure that will confuse you

The toe is long and the shoulder is abrupt, and both are diagnostic. The long flat foot of a printing-out curve is the region where a little pigment has formed and most of it washes away, so density stays near the paper white long after the light started working — which is why the first four or five steps of your tablet will read within a few hundredths of each other and carry no information. The shoulder, by contrast, arrives suddenly, because solarisation takes over: the blue starts being reduced to Prussian white as fast as more is made, and the curve stops rising rather than tapering.

⁇Seven stops or three? Two figures in one book, and what the difference is

Ware’s own work carries two numbers that look like the same quantity and are not, and a reader who meets them a chapter apart will think one of them is wrong.

In the photochemistry chapter, deriving how long a cyanotype needs in sunlight, he writes that the exposure scale for a proto-photographic process such as cyanotype is 7 to 8 stops between a just-perceptible image and maximum density — a factor of 128 to 256 — and gets 2 to 4 minutes of average sunlight from it. In the formulary chapter he writes that the classic process has an exposure scale of only about 0.9, three stops.

The course’s reading, offered as interpretation and not as Ware’s statement: the two figures measure between different endpoints. The 7-to-8-stop range runs from the very first perceptible tone — deep in that long flat toe, where the separation between adjacent steps is far below anything you could print with — up to maximum density, and it is the right quantity for the question he was asking, which was how much total dose it takes to darken a sheet. The 0.9 runs between useful endpoints, and it is the right quantity for the question this page asks, which is what density range a negative must hold. Both are correct measurements of different things, and neither may be substituted for the other.

Your own data will show you the gap directly. Find the first step whose density is distinguishable from the paper white on your instrument, and read the log exposure difference between it and the step at base + 0.04. That difference is the part of the toe your exposure scale throws away, and it is the whole of the discrepancy. It is also the best argument on this page for stating endpoints beside every figure, because a scale quoted without them is not a number anyone can use.

Interpreting the family: which changes move what

Three motions, and telling them apart is the analytical skill this page teaches.

• A curve displaced sideways with its shape unchanged is a speed change. Same scale, same slope, same maximum density, different exposure. Run 1 is predicted to do this, and only this.
• A curve whose shoulder rises is a maximum-density change, and on this process that almost always means a change in how much pigment survived the wash. Runs 3 and 4 are predicted to do this.
• A curve whose shoulder moves right while its toe stays put is an exposure-scale change, and it is the only one of the three that changes what negative you need. It usually comes with the second.
• A change in the toe alone, with the rest of the curve unmoved, is fog — chemical fog from paper impurities or from an over-strong acid bath, not a property of the sensitiser. Ware’s diagnostic for it is the coated-but-unexposed border against the uncoated margin, and every strip you cut has both.

What is actually a contrast control, and what is not

The classic cyanotype has no contrast control worth the name, and saying so plainly is more useful than listing four things that move the scale by a few hundredths.

What genuinely moves it, in descending order of effect: the paper, because retention sets the top of the curve; the processing bath, within the narrow band the formulation tolerates; and the choice of green against brown ammonium iron(III) citrate, which Ware’s comparison table credits with a stronger solarisation effect and therefore a higher maximum density, along with about a stop more speed in the mid-tones. What does not move it: exposure, which moves you along the curve rather than changing its shape; and the sensitiser ratio, on the evidence of about sixty published recipes.

What the tradition offers, and this course excludes. The additive that appears whenever contrast is discussed in this family is a dichromate, and the manifest asked whether it belongs to the classic process or the New Cyanotype. The answer is both, in different roles, and it is worth having exactly:

Where it appears	Amount	What its own author says it does
The improved Classic sensitiser	0.2 g of ammonium dichromate in 200 cc, with oxalic acid	Claimed to keep the highlights whiter. Ware prints the formula and argues against it in the same paragraph, calling the oxalate and dichromate poisonous and saying the sensitiser should not be placed in the hands of children
The New Cyanotype	0.1 g in 100 cc, marked optional	Primarily a preservative: a strong oxidising agent that stops impurities forming Prussian blue in the bottle and lifts the shelf life to several years. Ware adds that it “will also tend to increase the contrast of the sensitizer significantly”, citing his own characteristic curves, and that omitting it shortens the shelf life to a few months
A re-oxidation bath	—	Ruled out by its own advocate. Ware records that a dilute dichromate bath re-oxidises solarised prints rapidly and has been recommended for it, and instructs that under no circumstances should it be used

So the contrast claim is real, it is Ware’s own, and the course still does not use it. The chromium ruling is that chromium(VI) is used at no level in this course, and PubChem’s aggregated classification for potassium dichromate is the reason: carcinogenicity, germ cell mutagenicity and reproductive toxicity, in a domestic setting with a domestic waste stream. Ware notes that the optional use “does introduce a toxic substance into the working environment” and reaches a different conclusion about whether that is acceptable at 0.1 per cent — but he also writes, in the chapter on disposal, that the dichromate is the only chemical in the whole work recognised as hazardous, calls it a listed human carcinogen, and adds in the next sentence that it does not need to be used in traditional cyanotype at all. On that point the course and its main source agree completely, and the exclusion costs the classic process nothing it had.

The practical consequence is worth stating without hedging: if you need a cyanotype whose contrast you can set, you do not add anything to the classic sensitiser — you use a different sensitiser. The Simple Cyanotype sets its exposure scale at 2.7, 2.3 or 1.8 by the volume of ammonia alone, with a fourth grade at about 1.3 from processing in water, and it contains no dichromate and no oxalate at all. The New Cyanotype offers about 2.6 down to 2.1 by the strength of its development bath. Both belong to the variants lesson, and both are the answer to the question this section asks.

Feeding the result forward

Your measured exposure scale is the specification for the standard negative. In contact printing there is no Callier effect, so the ultraviolet density range the negative must hold is not merely related to the process’s exposure scale — it is the same number. Write it on the negative’s envelope and in the exposure log.

And it is the first row of the comparison atlas. The printing processes matrix currently says of itself that its rows come from different photographers, papers and decades, so it describes what the processes are like rather than what they do to one picture. This measurement, on your paper with your negative, is the first row that does not have that caveat — and the reason Parts XXII to XXV keep asking you to print the same negative is that eight such rows make a table nobody has published.

!What this session does not establish, at any confidence

Nothing about a second batch of citrate. The iron content of ammonium iron(III) citrate varies from 14 to 28 per cent w/w depending on form and supplier, and Ware calls it a notoriously ill-characterised chemical. Your figure belongs to your lot number, which is why the lot number is in the log.

Nothing about another paper. Variable four measured two papers and found a difference; it did not establish a rule about sizing, because two papers differ in more than one way.

Nothing about the New or Simple formulations. Every number here is the classic sensitiser. Ware’s measurements on the New Cyanotype — the one-stop shortening from heat drying, the one-stop lengthening from darktime — are for a different iron complex making a different Prussian blue, and importing them here would be exactly the error this page warns against in run 2.

Nothing about permanence. A curve measured the morning after is not an ageing test, and Prussian blue is a pigment that fades in light and regains in the dark.

Troubleshooting

What you see	Where it came from	What to do
The two run 0 controls disagree by a constant offset	Mixed sensitiser aged between the first and last coating, or the lamp drifted	Subtract the offset and say so. Next time, coat everything in one pass and expose the controls first and last as here — which is how you found out
Only four or five steps land between the endpoints	The tablet’s increment is too coarse for this process	Expected on a T2115 and the reason this page prefers a T3110. Report the exposure scale from the step count and do not claim a slope
Every strip’s toe is lifted and the coated border is blue	Chemical fog from the paper, or an acid bath that was stronger than 1 per cent	Blue fog in cyanotype highlights. Compare the coated-unexposed border with the uncoated margin before anything else
Maximum density lower than expected on every strip	Under-exposure, image loss in the wash, or a green-channel reading on a red-absorbing pigment	Weak cyanotype maximum density. Check first that the top two steps were indistinguishable before washing
The whole scale collapses and the wash tray runs blue	Peptisation, which is the mechanism under test	Not a fault. It is run 3 and run 4’s subject. Record how blue the tray went, per arm
Highlights stained yellow after drying	The wash stopped before the sensitiser cleared	Yellow iron stain. The endpoint is the disappearance of yellow under blue light, not a time
The image bleached during the wash	Alkaline or hard water	Bleached by an alkaline wash. Test your tap water; use distilled for the final rinse
The damp strip in run 2 is unevenly dense	The sheet dried unevenly during the exposure, or it stuck to the negative	A real risk of the treatment, not a result. Shorten the exposure or re-run side by side
Densities creep between the first and last strip read	The instrument drifted, or the strips were not equally dry	Part XV’s warm-up rule, and read again in the reverse order to see whether the trend reverses

Clean-up

Bottle the leftover mixed sensitiser only if you intend to use it within the hour; otherwise it goes to the residue container, because a mixed classic sensitiser has a short life and a bottle of it is a source of tomorrow’s confusion.

Rinse everything the sensitiser touched — rod, cups, syringes, trays — into the collected stream, three times with a small volume rather than once with a large one. The acid tray, its graduate and its funnel are washed separately and last, and their rinse goes into the acid bottle. Wash gloved hands, remove gloves inside out, wash hands, wipe the bench.

Storage

The stocks. Brown glass, dark cupboard, labelled and dated, with the lot numbers on the label rather than only in the notebook. A few crystals of thymol will keep the mould off the citrate solution.

The strips are the experiment and they keep. All of them in one labelled envelope, interleaved with unbuffered paper — never a buffered archival sleeve, which is alkaline and will attack the image. Re-reading them in six months against a re-anchored instrument tests your uncertainty statement better than anything else on this page.

The tablet goes back flat, dark, in its sleeve, handled by the edges. It is part of the instrument now, and a calibrated part’s certificate lives with it.

Disposal considerations

The chemistry first. The collected liquid is dilute unreacted ammonium iron(III) citrate and potassium ferricyanide, suspended peptised Prussian blue, and — in one bottle — 1 per cent citric acid. The cyanide is complexed to iron and stays complexed at every condition this session creates.

Where the sources disagree. Ware argues that cyanotype effluent need not be treated as chemical waste, naming the citrate as food additive E381, the ferrocyanide that ferricyanide reduces to in the environment as E536, and Prussian blue as so insoluble that it is given clinically as an antidote to thallium and radiocaesium poisoning. Princeton’s environmental health guidance treats ferricyanide solutions as hazardous waste without qualification by dilution. The coating lab sets both positions out in full and neither is settled here.

What the course does is publish no jurisdiction-specific instruction and make no exception: the containers are collected and labelled, and the acid bottle stays separate. The disposal ruling has the reasoning. Check your local regulations — they govern, they differ between authorities inside one country, and they change. What does not change is the composition of the liquid in the bottle, which is why the label carries it.

Questions

1. Your run 0 curve gives a maximum density of 0.95 and an exposure scale of 0.72, both well below Ware’s figures. Give three explanations that are about your measurement rather than your chemistry, and say which one the coated border and the uncoated margin would distinguish.
2. Run 1’s two arms produce curves of identical shape, one displaced 0.25 log units to the right of the other. Which arm had more ferricyanide, what is the mechanism, and why is this a speed result rather than a contrast result?
3. Explain why the exposure scale you measure is also the specification for the negative, and name the optical effect whose absence makes that equality true.
4. Ware writes that a proto-photographic process spans 7 to 8 stops between a just-perceptible image and maximum density, and also that the classic cyanotype’s exposure scale is about 0.9, which is three stops. Both statements are in the same book. Reconcile them, and say what your own endpoints have to do with the answer.
5. You have only the green reflection head. Which of the four runs can you still report, which become unreportable, and what single sentence must appear beside every density you publish?
6. Your dose repeatability pre-test gives a spread equivalent to 0.15 log units. What does that do to run 2, and what redesign would rescue it?

Further experiments

Repeat run 3 with the acid concentration as a series — 0.25, 0.5 and 1 per cent w/v citric — and find where your paper’s highlights begin to blue. Ware gives the upper bound as a rule (“below 1 per cent citric” for this family) rather than as a measurement, and where the boundary falls is a property of the paper.

Print the same tablet on a bought pre-coated paper and put its curve on the same axes. The kits do not publish their strengths, so you cannot say what the difference is of — but you can say how large it is.

Run the brown against the green salt, if you can obtain both. Ware’s Table 4.2 makes five predictions in one column — about a stop slower, weaker solarisation and therefore lower maximum density, a distinctive edge etch, a duller blue and better fade resistance — and a step tablet tests three of them in one afternoon.

Do the fifth run this page left out: double coating. Ware names it as one of the two remedies practitioners adopt for the classic process’s image loss. Coat, dry, coat again, dry, expose. If the hypothesis is right, the second coating should raise the maximum density more than it lengthens the scale, because it adds pigment where pigment already survives.

And measure the fade. Cut one strip in half, put one half in a light-proof box and the other in a north window for a month, and re-read both. Ware’s finding is that the maximum fade sits in the mid-tones between densities of about 0.5 and 1.0 rather than at maximum density, which is a testable claim and a surprising one.

Recap

The classic cyanotype’s exposure scale is short — about 0.9 to 1.2 in Ware’s published figures — and it is short because the image washes away, not because the photochemistry is contrasty. That one mechanism predicts everything this session measures: the paper and the processing bath move the curve because they change how much Prussian blue stays in the sheet, while the sensitiser ratio moves it sideways at most and humidity, on Ware’s own statement, barely moves it at all.

The measurement is a step tablet, a calibrated dose, plain water, a night on the rack and a reflection densitometer read in the red, because Prussian blue absorbs near 700 nm and a green channel is looking at the wrong part of the band. Out of it come four numbers with your name on them — a maximum density, an exposure scale between stated endpoints, a slope, and an error budget that says which of them you may quote — and the second is what the rest of the alternative-process cluster will ask you for.

The honest closing note is what the session cannot give you. There is no contrast control in this process. The traditional one is a chromium(VI) salt this course does not use at any level and which its own advocate says the traditional process does not need; the modern answer is not an additive but a different sensitiser, which is the next lesson.

Check your understanding

Question 1. Your control curve has a maximum density of 1.20 and a base density of 0.07. Using this course's endpoints, at what two densities do you read the exposure scale?

• 0.07 and 1.20
• 0.11 and 1.08
• 0.00 and 1.20
• 0.07 and 1.08

Show the answer and why

Answer: 0.11 and 1.08

Base + 0.04 gives 0.07 + 0.04 = 0.11 at the bottom, and 0.9 x Dmax gives 0.9 x 1.20 = 1.08 at the top. The pair is Ware's own convention for his platinum tables, and the course uses it everywhere so that its figures are comparable with each other. The point of defining endpoints at all is that an exposure scale quoted without them is not a measurement: read the same curve between base and Dmax and you would get a longer number for the same material.

Question 2. Why does this page prefer a Stouffer T3110 over the course's usual T2115, and what does it give up?

• The T3110 has a higher maximum density, so more of the curve is covered
• The T3110's 0.10 steps put about nine points on a 0.9 exposure scale where the T2115's 0.15 steps put six, but no calibrated version of the T3110 is offered
• The T3110 is calibrated against a NIST reference and the T2115 is not
• The T3110 is smaller, so it fits more strips on a sheet

Show the answer and why

Answer: The T3110's 0.10 steps put about nine points on a 0.9 exposure scale where the T2115's 0.15 steps put six, but no calibrated version of the T3110 is offered

Both tablets reach a maximum density of 3.05; the difference is the increment. The number of steps that land between the endpoints is the exposure scale divided by the increment, so a short-scale process needs a fine tablet or it gets four or five usable points. What you give up is certification: Stouffer name the T2120CC and T1530CC as their calibrated parts, so a T3110's increments are nominal and your log exposure axis carries the maker's tolerance as a Type B uncertainty that no amount of repeated reading will reduce.

Question 3. Run 1 produces two curves of identical shape, one displaced to the right of the other. What have you measured?

• A contrast change, because the curves are separated
• A speed change, because the shape, slope and maximum density are unaltered
• A fog change, because the toe has moved
• Nothing: identical shapes mean the arms were the same

Show the answer and why

Answer: A speed change, because the shape, slope and maximum density are unaltered

A sideways displacement with the shape preserved is a change in how much exposure the material needs, which is speed. The mechanism is the internal filter: the hexacyanoferrate takes almost no part in the photochemistry but absorbs strongly in the same near-ultraviolet band, so more of it lengthens the exposure without changing what happens after the light stops. Whitaker reported the same thing in 1883 from a very wide range of ratios, and Ware's survey of about sixty recipes found the visible outcome varies little with proportion.

Question 4. Ware measured that 19 hours of darktime lengthens the New Cyanotype's tonal scale from 2.3 to 2.6. What may you conclude about the classic cyanotype?

• The same, since both are cyanotypes making Prussian blue
• The opposite, since the classic process is slower
• Nothing, because the measurement is for a different iron complex and a different Prussian blue
• That darktime is worth about one stop in any iron process

Show the answer and why

Answer: Nothing, because the measurement is for a different iron complex and a different Prussian blue

The two formulations differ in the photosensitive complex, in the cation of the pigment and in how much of the image survives the wash, which is exactly why their exposure scales differ by a factor of more than two. Transferring a measured figure between them because they share a colour is the error this page warns against in run 2 and the reason the humidity claim is tested on the classic process rather than imported. Ware's general statement that humidity has very little effect on cyanotype printing is a claim about the process; his darktime figures are a measurement on one formulation.

Question 5. You read your strips on a green reflection head because you have no red LEDs. What is the effect on your numbers?

• None, since density is density
• They are useless, and the session should not be run
• Densities read low and the scale reads short, so differences remain valid but no figure may be compared with a red-channel one
• Densities read high, because green light is scattered more by paper

Show the answer and why

Answer: Densities read low and the scale reads short, so differences remain valid but no figure may be compared with a red-channel one

Prussian blue absorbs in a single broad band centred near 700 nm, so a red channel sees the peak and a green channel sees the shoulder of it. A green reading therefore understates the density and compresses the scale, which is why Ware read his fading study on the red channel of a reflectance colour head. The differences between your own strips are still real because every one of them was read the same way; what fails is any comparison with a figure measured in another band. ISO 5-3 is the reason the channel has to be stated at all: a density is defined by its geometric and spectral conditions together.

Question 6. A dichromate is the traditional contrast additive in this family. Where does it actually belong, and why does this course exclude it?

• Only to the New Cyanotype, and it is excluded because it shortens shelf life
• To both the improved Classic sensitiser and the New Cyanotype, and it is excluded because chromium(VI) is used at no level in this course
• Only to the classic process, and it is excluded because Ware found it ineffective
• To neither: it is a re-oxidation bath and nothing else

Show the answer and why

Answer: To both the improved Classic sensitiser and the New Cyanotype, and it is excluded because chromium(VI) is used at no level in this course

Ammonium dichromate appears at 0.2 g in 200 cc in the improved Classic sensitiser and at 0.1 g in 100 cc, marked optional, in the New Cyanotype, where its primary role is as a preservative and its contrast effect is secondary but real on Ware's own curves. It also has a third historical use as a re-oxidation bath, which Ware rules out in the strongest terms he uses. The course excludes all three because the chromium ruling permits chromium(VI) at no level, and the practical consequence is that a reader who wants a settable contrast changes sensitiser rather than adding anything.

Check my answersClear