Precise leveling staffs for first-order geodetic work are graduated on dual invar columns with individual lines applied at 0.5 mm intervals, allowing calibrated readings to support about 0.2 mm/km double-run accuracy and per-reading resolution down to roughly a hundredth of a millimeter [S1]. The graduation increment is the single specification that ties the staff to the level, the sight distance, and the working tolerance, so choosing the wrong spacing quietly inflates field error.
The two families in production are analog optical staffs read by eye through the telescope, and coded bar-code staffs read by digital levels. Across both, graduation step, material, and working tolerance are coupled: a 5 m aluminum E-pattern staff with 10 mm blocks and 1 mm subdivisions typically rates at about ±1.5 mm over its full length and ±0.3 mm per meter, while a 5 m fiberglass survey staff with 5 sections and an E + B1 dual scale lists about ±1.2 mm over its full length [S2][S3].
First-Order Invar Staffs: 0.5 mm Lines on a Calibrated Scale
First-order precise leveling staffs carry the graduation directly on invar tape, in two parallel columns, with individual graduation lines applied at 0.5 mm intervals [S1]. The use of invar, a nickel-iron alloy with a near-zero thermal expansion coefficient, is what lets those 0.5 mm divisions stay dimensionally stable enough to be calibrated at sub-tenth-of-a-millimeter uncertainty on a horizontal comparator with a CCD camera and telecentric lens [S1].
These staffs are calibrated periodically, because the staff defines the scale of the height reference system, not just an auxiliary scale for the operator to read [S1]. Calibration produces a correction certificate that is applied to the raw field data, and a serviced staff plus a calibrated level is the combination that lets double-run leveling reach about 0.2 mm per kilometer of section [S1].
Engineering E-Pattern Staffs: 10 mm Blocks, 1 mm or 5 mm Subdivisions
The E-pattern staff is the workhorse of construction and topographic leveling, so named because the 10 mm blocks read through the telescope look like a chain of letter Es, with each block divided into 10 mm, 5 mm, or 1 mm segments depending on the model [S2][S3]. Standard catalog lines include a 5 m 5-section aluminum staff with 5 mm and 10 mm graduations for laser leveling, a 5 m 5-section aluminum E-pattern staff with 10 mm graduations for automatic dumpy level work, and a 5 m high-precision staff that layers a 1 mm laser scale over the 10 mm E scale for combined optical and laser use [S3].
Typical accuracy figures for this class are about ±1.5 mm over full length and ±0.3 mm per meter, with 2.0 mm tube-wall thickness on the fiberglass variant and a 20 arc-min / 2 mm circular bubble for plumbing the staff vertical [S2]. Front-face scales commonly use F2 (1 ft / 10ths / 10ths), and reverse faces carry B1, B3, or E+1 mm patterns, which gives a quick independent check on each setup [S2].
Coded Bar-Code Staffs for Digital Levels
Digital levels do not use a human-readable E pattern. They read a pseudo-random bar code printed on the staff and resolve height internally, so the apparent graduation increment you can see in the field is no longer the spec that matters. What matters is that the bar code matches the instrument firmware: a 5 m 4-section barcode staff for Topcon and Sokkia digital levels, for example, lists about ±1.2 mm over its full length, with B1 (E+1 mm) on the reverse for fallback optical reading [S2].
Staff length is still fixed at typical 3 m, 4 m, or 5 m for portable use, and the staff is still calibrated in a comparator, because the digital level resolves height by interpreting the bar code against the staff scale, and a damaged or worn code degrades the reading long before the staff length changes. For background on how digital bar-code reading compares against optical E-pattern reading, see automatic level vs digital level bar-code staff reading accuracy.
Matching Graduation Increment to Instrument and Tolerance
The decision rule is simple. Use a 0.5 mm invar staff with a precise optical or digital level when the line of work is first-order geodetic leveling, deformation monitoring, or any task that has to close to 0.2 mm per kilometer or better [S1]. Use a 10 mm E-pattern staff with 1 mm or 5 mm subdivisions on an automatic level for general construction, earthwork, and topographic surveys where the working tolerance is in the few-millimeter range [S3]. Use a coded barcode staff only with the matching digital level family, and verify the staff lists the right firmware class for that instrument [S2].
Working accuracy varies by an order of magnitude across the three families, and that is the real engineering rationale for picking by increment. Invar + 0.5 mm lines supports per-reading uncertainty of a few hundredths of a millimeter, optical E-pattern staffs with 1 mm subdivisions are limited to roughly ±1.5 mm full-length accuracy, and digital barcode staffs cluster near ±1.2 mm full-length with the gain that the operator does not have to interpolate the millimeter [S1][S2][S3]. If your sight distance is above about 40 m to 50 m, expect the optical E-pattern reading to lose its last millimeter, and either shorten the sight or move to a digital level with a matched barcode staff.
Reading the Graduation at the Telescope
Three hairs in the telescope reticle do the actual reading: an upper stadia hair, a central cross hair, and a lower stadia hair, and the reading is the height of the central cross hair on the staff plus an intercept-based distance to the nearest centimeter or millimeter, depending on the staff increment [S4]. The stadia hairs also give a one-shot distance, multiplying the intercept by the instrument stadia constant, typically 100, for a quick horizontal distance to the staff without a tape [S4].
To pull the most out of a 10 mm E-pattern staff, train the operator to read to one tenth of the smallest subdivision, not to the graduation block, so a 1 mm subdivision is read to 0.1 mm in the notebook, and a 5 mm subdivision to 0.5 mm. The bubble on top of the staff and a 20 arc-min / 2 mm circular bubble both have to be centered before the reading, because a 1 degree staff tilt at 3 m height introduces about 50 mm of vertical error, which is an order of magnitude worse than any graduation increment on the staff itself.
Materials, Length, and the Practical Trade-Offs
Aluminum staffs are the lightest, typically 5 m telescopic in 4 or 5 sections, and they are adequate for general construction work where the working temperature range is moderate and the staff is not left wet in storage. Fiberglass staffs add corrosion resistance, non-conductivity, and a 2.0 mm tube wall that survives rougher site handling, which is why they are specified for utility, railway, and electrical work, and they still list about ±1.5 mm full-length accuracy on the E-pattern scale [S2][S3]. Invar staffs are precision instruments, not site tools, and are stored, transported, and calibrated under controlled conditions, not leaned against a fence post.
Length choice is also a tolerance choice, because absolute thermal and mechanical error accumulates along the staff. A 5 m staff is the practical maximum for handheld work, a 3 m staff is the everyday choice for interior and tight-site work, and longer invar staffs in geodetic work are handled as a separate calibrated line scale, not as a longer hand rod. The trade-off is consistent across the catalog: shorter staffs are easier to plumb and read, longer staffs halve the number of setups per kilometer, and the graduation increment and material set the accuracy floor on either choice.
Failure Modes and Common Reading Errors
Four failure modes show up repeatedly in the field. First, reading a 10 mm E-pattern staff to the block, not to the millimeter, throws away a factor of 10 in resolution. Second, using a worn aluminum staff whose 5 mm or 1 mm subdivisions have been polished flat by site abrasion, which silently degrades the reading long before the staff length changes. Third, mismatching a barcode staff to a digital level from the wrong family, which gives a reading that looks plausible but is biased by a millimeter or more. Fourth, ignoring the staff plumb bubble, which combines a vertical error with a horizontal error and is the most common source of millimetric drift on a 3 km traverse. [S2]
To control all four, the operating practice is to calibrate optical E-pattern staffs at least annually against a horizontal comparator, replace any staff with polished graduations, barcode-staff match the staff part number to the instrument firmware list, and use a 20 arc-min / 2 mm circular bubble for every setup [S1][S2][S3]. Calibration certificates come from laboratory comparators, since no commercial turnkey comparator exists, and each institution builds its own, so the traceability chain has to be checked before relying on a published accuracy figure [S1].
Standards, Calibration, and What the Numbers Actually Mean
There is no single global standard that pins every graduation increment, but two principles are standard practice and worth quoting. First, calibration of a leveling staff on a horizontal comparator is what defines the staff scale, and the result is a calibration certificate applied to raw field data; without it, the staff graduation is a nominal value, not a metrological one [S1]. Second, the stated full-length accuracy of an E-pattern staff, around ±1.5 mm over its full length and ±0.3 mm per meter for typical 5 m aluminum or fiberglass models, is a working accuracy figure, not a calibration certificate, and it is met only when the staff is plumb, the level is within compensator range, and the sight distance is inside the manufacturer limit [S2][S3].
For first-order work, the working figure to design to is about 0.2 mm per kilometer for double-run leveling with calibrated invar staffs and a precise level, and per-reading uncertainty down to a hundredth of a millimeter [S1]. For everyday construction work with an E-pattern staff and an automatic level, design to roughly ±1.5 mm per setup and roughly ±0.3 mm/m, and verify in the field by closing a short loop at the start of each day. For a digital level with a matched barcode staff, design to about ±1.2 mm full length and trust the instrument to interpolate, but still plumb the staff and still respect the sight distance limit, because the bar code does not fix tilt or air shimmer.
Looking ahead, the practical signal to track is whether digital barcode staffs continue to displace optical E-pattern staffs on mid-tier construction sites, and whether invar-staff calibration comparators become commercially available, which would let more surveyors tie their field work to a metrological certificate rather than a nominal spec. Either development would shift which graduation increment the average site actually reads, but the underlying rule will not change: pick the staff increment to match the instrument, the sight distance, and the working tolerance, and verify with a closed loop before you trust the numbers.