Above the loom in Julia Maudlin’s photograph, a gridded floral drawing sits where the weaver can see it. The finished pattern is already visible in the knots below, but there is a translation happening between the two. A carpet cartoon is not simply a picture of the rug. It has to tell the weaver which colour belongs at a particular point and where one shape ends and another begins.

A drawing has to become instructions

On a gridded cartoon, a small square can correspond to a knot position. That makes the grid useful, but it does not solve every problem. A painted flower may contain several close shades; its outline may pass through a corner of a square; and the drawing may include marks that help the eye without calling for a knot of that colour. The weaver needs a legible system for separating design, yarn choice and execution.

Carpet designer Abdollah Mirzaei compared two ways that workshops in Isfahan and Tabriz made such plans readable. His study draws on interviews and professional practice in those cities. It describes common approaches in the workshops he studied, not rules that every rug from either city must follow.

The handoff starts before a weaver chooses yarn. A design has to be divided into usable sections, secured so it can be handled at the loom, and marked clearly enough to distinguish a colour boundary from a decorative flourish. Mirzaei describes several people who may take part, including the designer, a specialist who paints or dots the plan, and the weaver. The cartoon is therefore a working document that travels between decisions. Its beauty on paper is useful only if the person tying knots can read the intended sequence.

That is why a photographed cartoon cannot be understood as an enlarged picture of the finished pile. Some marks are instructions for knots. Others organize the drawing or help distinguish neighbouring areas. The two methods below make different choices about which information sits directly on the grid and which is carried on a separate colour card.

When the colour is painted onto the grid

In the Isfahan method Mirzaei describes, the design is mounted in workable sections and its areas are painted in the colours intended for weaving. The grid remains visible through the paint. An experienced painter then marks boundaries at knot scale, often with darker, carefully squared dots. The drawing can be read rather directly: look at the painted field, locate its boundary and follow the chart as the work advances.

The order of preparation matters. In the documented process, sections of the drawing are fixed to boards. A protective gum layer helps keep paint from soaking into the paper and lets the painter correct a mistake. Colour is applied lightly enough for the grid lines to remain countable. Larger parts of the composition, such as the field, main border and medallion, are treated before smaller areas. These are details of a particular way to prepare a durable, readable plan, not a materials list for identifying an Isfahan rug.

The colour filling and the later dots do different work. A broad painted shape tells the reader which colour occupies an area; the dots settle where adjacent shapes meet at the scale of individual knots. Mirzaei stresses that poorly placed boundary dots can break the apparent curve or detail of a motif. A flower that looks smooth in a freehand drawing must still be resolved into a sequence of small squares. The person making that translation is interpreting the design, not just colouring inside lines.

This is skilled preparation, not a mechanical transfer. A line that looks graceful on paper can turn jagged in pile if its sequence of knots is poorly resolved. Mirzaei describes protective coatings, carefully prepared pigments and final checks as part of making a working cartoon durable enough for repeated handling. Those details concern the paper plan. They are not evidence that the rug’s wool was dyed with the same mineral pigments used on the drawing.

Even a directly coloured plan need not show every possible shade. The study describes relatively restrained shading in this example, sometimes limited to a lighter line around a motif. That helps explain why it can remain comparatively easy to read a painted field as a weaving colour. It does not mean that every Isfahan carpet has a small palette or that its designer never sought depth. The paper itself notes other subsidiary methods, and the comparison is about the preparations it documents.

When a number stands for a yarn shade

Mirzaei’s Tabriz example uses another layer of translation. The cartoon may be shaded to make a flower or leaf read convincingly, while numbered areas refer to a separate card of actual yarn samples. A weaver matches the written number to a yarn on that card. The plan can therefore carry subtle visual modelling without relying on every painted tint to be the final weaving instruction.

The sequence in the study starts with outlining a design, then adding painted or drawn shading. It can use watercolour, gouache or coloured pencil to give a naturalistic flower a sense of volume. Those drawing colours may be chosen for their effect on the paper and need not match the yarn ultimately tied into the rug. Once the design's areas and boundaries are legible, the specialist selects real yarn samples, fixes them to a card and assigns each a number. Corresponding numbers are written into the plan. This creates a link between an area of the drawing and a physical yarn shade rather than asking the weaver to guess from the painter's wash.

The card changes the kind of decision a workshop can make. If a group of close blues is difficult to separate on paper, samples and numbers can preserve their intended distinctions. Mirzaei also notes that adjusting the chosen yarns or their assigned numbers could produce another colour arrangement from a design. That does not make the process effortless: neighbouring yarns may affect how one another appears only once woven together. The card supports communication, but someone still has to judge the palette and the weaver still has to execute it.

The study also distinguishes main dots that guide the woven line from smaller guide marks used to separate nearby shaded areas. Those small marks help the person reading the chart but are not themselves tied as knots. The distinction matters when several close tones meet: a useful mark on the paper does not necessarily become a visible feature in the rug.

Here the question is not simply whether a plan has dots. Both documented approaches use them. The crucial issue is what each mark means in that system. In Mirzaei's Isfahan example, the boundary dots are squared and coloured to indicate knots along the line. In the Tabriz example, larger boundary dots coexist with finer marks that merely keep adjacent shaded passages visually distinct. Reading every tiny mark as a knot would turn a helpful notation into a false account of the rug.

Read the diagram as a handoff, not a finished rug

The left half of the schematic makes its colour decision directly in the grid. Red, blue and green are filled into cells, so the coloured plan carries much of the answer a weaver needs. The right half separates the drawing from the selected yarns. A written number points to a shade on the card; a dot can clarify a boundary. The shapes in this illustration are deliberately simple. They show where information is stored, not how a historical workshop rendered a real flower.

Imagine that a leaf in a design passes from dark to light. A direct-colour plan can paint those passages on the grid and define their edges at knot scale. In a numbered plan, the drawing can suggest the leaf's volume while separate numbers identify the yarns assigned to its passages. This is a hypothetical reading of the two communication systems, not a reconstruction of any particular carpet. In either case, the drawn transition becomes a sequence of knots only when someone interprets it at the loom.

The photograph at the start shows why these abstractions matter: the weaver can consult a plan while the textile grows beneath it. It also has a clear evidential limit. It shows a loom and cartoon in Isfahan, but it does not establish which exact preparation method the pictured workshop used. We should not infer a yarn card, a paint recipe or a particular knot sequence from a room photograph alone.

Why the difference matters

These systems respond to a practical question: how can a designer’s decisions survive the handoff to another person at the loom? Painting the intended colours directly can make a restricted palette quick to read. Numbered yarn samples can keep many neighbouring shades distinct, even when the drawing uses visual shading to communicate depth. In either case, a skilled weaver still has to interpret the chart and control the material under hand.

The two plans also carry different kinds of labour. In the direct-colour example, considerable care goes into preparing the paper surface, filling its fields without hiding the grid and squaring the colour boundaries. In the numbered example, drawing, yarn selection, coding and reading the code must stay aligned. A mistake in either system could separate the designer's intention from the woven passage, but the study does not measure which method produces fewer errors or better rugs. It describes how the observed workshops made their information readable.

Mirzaei links the practices he observed to different design preferences and workshop relationships. That is an explanation of his cases, not a reason to call one city traditional and the other inventive by nature. Both systems require design judgement. Neither can be identified with confidence merely by looking at a finished rug. Workshop methods vary, designs travel, and one carpet can involve several people whose decisions are no longer documented.

For a reader looking at a finished carpet, the distinction is valuable chiefly because it changes the question. Subtle colour changes in a woven flower show that design and yarn choices were coordinated; they do not reveal whether a painter shaded a cartoon, a card assigned the yarns, or a different working method was used. A cartoon displayed beside a rug may offer stronger evidence, but it still needs a secure connection to that specific weaving before it can explain its production. Treat the plan as evidence of a process, not as an automatic certificate of origin.

The next time you notice a flower with several shades in a hand-knotted rug, look beyond the number of knots. Someone had to decide where each shade should turn, how to make that instruction readable, and how to keep the passage coherent as it moved from a drawing into wool. For a broader account of Tabriz design and weaving, see our Tabriz rug guide.

Source

• Abdollah Mirzaei, “An Explanation Study of the Causes of Difference in Coloring and Dotting Methods of Isfahan and Tabriz Carpet Patterns” (2020).