The Henry Classification System and Pattern Types

Francis Galton had demonstrated by 1892 that fingerprints could be classified into broad types and that the configuration of ridges at specific landmarks could be encoded. But Galton's system was not scalable to an operational criminal bureau: his classification categories were too broad for large collections and his retrieval procedure was cumbersome.

Edward Richard Henry, Inspector General of Police in the Bengal Presidency (present-day West Bengal and Bangladesh), began developing a workable classification system around 1896. Working with two Indian sub-inspectors, Azizul Haque and Hemchandra Bose, who contributed critically to the mathematical design of the classification formula, Henry produced a system that classified each finger individually by pattern type and ridge count or ridge trace, then combined the ten fingers into a two-digit primary classification fraction. This fraction sorted cards into 1024 primary groups, small enough that individual-card comparison within a group became feasible.

The Bengal Criminal Investigation Department adopted the system in 1897. Henry brought it to England, and after a committee chaired by Lord Belper recommended replacing the Bertillon anthropometric system, Scotland Yard established a Fingerprint Bureau using the Henry system in 1901. By 1902 it had its first criminal identification, a burglary case. Within a decade, most English-speaking police forces had adopted variations of the Henry system, and several continental European forces (France, Germany, and eventually the Netherlands) adopted parallel systems with overlapping category structures.

The Indian connection to the Henry system's development is frequently underacknowledged. The contribution of Haque and Bose to the primary classification formula was documented by Henry himself and is reflected in the Bengali fingerprint bureau's operational records from 1897. Their work laid the ground for the friction-ridge biology that modern examinations rest on. The Henry system was therefore not a purely British invention but a collaborative product of the Bengal Police's forensic science infrastructure at the turn of the twentieth century.

The Henry system defines three primary pattern types. These are not arbitrary categories; they reflect genuinely different global ridge flow configurations that a trained examiner can distinguish from a good-quality rolled or scanned print.

A loop is a pattern in which one or more ridges enter from one side of the print (called the open side), curve around to form a recurve shape, and exit on the same side they entered. Loops constitute approximately 60 to 65 per cent of all human fingerprints across populations (the proportion varies slightly by population group and by study methodology). A whorl is a pattern in which ridges form at least one complete circuit around a central area, or form a related configuration (such as a double loop or a spiral). Whorls constitute approximately 30 to 35 per cent of prints. An arch is a pattern in which ridges enter from one side, flow across the print with a gentle curve or an upthrust, and exit on the other side without any recurving and without enclosing a central area. Arches are the least common pattern, appearing in approximately 5 per cent of prints.

These proportions have been verified in population studies across European, East Asian, South Asian, African, and Indigenous American groups, with some variation in whorl and loop proportions between populations (East Asian populations show slightly higher whorl frequencies; European and South Asian populations show slightly higher loop frequencies) but with arches consistently forming the smallest category everywhere.

Within each primary pattern type, the Henry system recognises subtypes based on ridge flow direction, the presence and position of specific landmarks, and the relationship between recurring ridges and those landmarks.

For loops, the critical distinction is between ulnar loops and radial loops. An ulnar loop recurves toward the little-finger side of the hand (the ulnar side, named for the ulna bone). A radial loop recurves toward the thumb side (the radial side, named for the radius). Ulnar loops are far more common than radial loops across all fingers, with radial loops appearing most frequently on the index finger. In the Henry formula, ulnar and radial loops are assigned different values (U or R), which affects the classification fraction. The ridge count of a loop is the number of ridges crossing an imaginary line drawn from the core (the innermost recurving ridge or a rod/dot at the centre) to the delta (the triangular confluence of ridges at the outer margin of the pattern). A loop with zero ridges crossing this line is classified as a whorl in some systems; in others, it becomes a tented arch. Ridge count is central to Henry sub-classification.

For whorls, the Henry system distinguishes four subtypes. A plain whorl has ridges forming a spiral or concentric circles around a central point, with two deltas. A central pocket loop whorl has ridges that recurve like a loop on one side but enclose a separate whorl pattern at the centre, the central pocket; it has two deltas. A double-loop whorl (also called a twin loop or composite) contains two separate loop formations whose recurves interlock or nest together; it has two deltas. An accidental whorl is any pattern that does not fit into the other three whorl subtypes or that combines features of two or more pattern types; it may have two or more deltas, and it is a residual category.

For arches, the distinction is between a plain arch and a tented arch. In a plain arch, the ridges flow across with only a gentle upward curve. In a tented arch, a central ridge or group of ridges makes a sharp upward thrust, forming a tent-like angle at the apex. Unlike a loop, a tented arch does not have a complete recurve, and the ridge that rises sharply does not turn back to exit on the side it entered. Tented arches are classified separately in some systems because their narrow apex can be mistaken by inexperienced examiners for a loop with a very short recurve.

Two landmarks are central to Henry classification and to the Level 1 analysis stage in ACE-V. The delta (also called a triradius) is the point of ridge confluence where three ridge systems meet: the loop or whorl system from above, and the two systems from the left and right lower margins. At a delta, the ridges form a characteristic triangular or Y-shaped confluence. Every loop has exactly one delta; every whorl has exactly two deltas (or more for accidentals); arches have no delta.

The core is the approximate centre of the pattern. In a loop, the core is identified as the innermost recurving ridge's peak (if the ridge is a simple curve) or, where a rod or short ridge stands upright within the recurve, the top of that rod. In a whorl, the core is not a single point but is approximated by the innermost circuit of ridges.

Ridge counting is performed along an imaginary straight line from the core to the delta. Every ridge crossing or touching this line is counted, except the core ridge itself and the delta ridge itself. A loop pattern with a ridge count of 0 through 9 is a "few" loop; 10 through 13 is "ten to thirteen"; and so on. The ridge count, combined with the pattern type (loop/whorl/arch) and the loop direction (ulnar/radial), produces the value assigned to each finger in the Henry formula.

Ridge tracing applies to whorls. The examiner locates the left delta and follows the ridge that issues from the inner side of the left delta in the direction of the right delta, tracing it until it meets or passes the right delta. If the traced ridge passes inside (above) the right delta, the whorl is "inner" (I); if it passes outside (below) the right delta, it is "outer" (O); if it meets the right delta or passes through a position within two ridges of it, it is "meeting" (M). Whorl tracing results (I, O, or M) feed into the Henry secondary classification alongside ridge counts.

1. Identify the pattern type
   Determine whether the print is an arch, loop, or whorl by examining the overall ridge flow. Look for the presence of recurving ridges (loop or whorl) and the number of deltas (arch: 0, loop: 1, whorl: 2 or more).
2. Locate the delta(s)
   Find the triradius or Y-junction where three ridge systems meet. For loops, one delta; for whorls, two deltas. Mark each delta's position.
3. Identify the core
   For loops: find the innermost recurving ridge's peak or the tip of any rod within the recurve. For whorls: identify the innermost circuit.
4. Count ridges (loops) or trace ridges (whorls)
   Loops: draw a line from core to delta and count crossing ridges, excluding core and delta ridges themselves. Whorls: trace the ridge from the inner side of the left delta toward the right delta and record the result as I, O, or M.
5. Identify the loop direction
   For loops only: determine whether the open side faces the little finger (ulnar) or the thumb (radial). Ulnar loops are far more common.
6. Assign the Henry formula value
   Each finger position (right thumb through right little, left thumb through left little) receives a value based on pattern type, ridge count, and loop direction. The ten values combine into the Henry primary classification fraction and the secondary, sub-secondary, and special loop and whorl sub-classes.

The transition from manual Henry-filing systems to Automated Fingerprint Identification Systems did not eliminate the need for pattern type classification; it changed its purpose and its position in the workflow.

In a manual Henry bureau, the pattern formula was the primary search tool: a new arrest fingerprint card was assigned its formula, and only cards with the same or similar formula were physically retrieved for comparison. The formula carried the entire retrieval burden. In an AFIS system, the retrieval burden is carried by minutia coordinates extracted algorithmically from a scanned print and matched against a database of similarly encoded reference prints using a similarity score and rank-ordered candidate list.

Pattern type in an AFIS context serves as a filter, not a retrieval key. Most modern AFIS implementations allow pattern type to be entered as a search constraint: searching only among loops, for example, if the examiner is confident the latent print is a loop. This can reduce false candidates and speed the rank-ordering. But the AFIS does not require a pattern type entry to function: a latent print with no determinable pattern type can still be searched on minutia alone.

Where Henry's vocabulary became more important post-AFIS is in the ACE-V method's Analysis stage. When a latent print examiner analyses an unknown latent print before any comparison, they document Level 1 detail, which includes the pattern type, using the Henry vocabulary. A latent loop cannot have come from a reference whorl. This exclusionary function of pattern type is used in the Analysis stage before the examiner ever looks at a candidate reference print, which is an important procedural safeguard against confirmation bias.

In the UK, the Fingerprint Bureau's standard documentation protocol requires the examiner to record pattern type as part of the Level 1 analysis before the comparison target is loaded. The FBI's fingerprint quality standards (OSAC Friction Ridge Subcommittee guidelines) and the Scientific Working Group for Friction Ridge Analysis, Study and Technology (SWGFAST, now superseded by the OSAC fingerprint subcommittee) similarly formalised Level 1 documentation. In India, the Bureau of Police Research and Development (BPR&D) guidelines for fingerprint examination reference pattern type recording as a pre-comparison step, though the level of documentation formality varies between central and state-level fingerprint bureaus.

The ACE-V method structures the examination of a latent print into hierarchical levels of detail, a framework articulated by David Ashbaugh in his 1999 monograph "Quantitative-Qualitative Friction Ridge Analysis" and now used in some form by every accredited forensic fingerprint laboratory.

Level 1 detail is the overall ridge flow and pattern type: is this a loop, whorl, or arch? It is assessed from the print as a whole, without reference to individual ridge events, and it corresponds directly to the Henry pattern vocabulary. Level 1 detail is visible even in low-quality latent prints where individual ridges cannot be resolved: a smeared latent on a glass surface may not permit minutia comparison but may clearly show the overall spiral or loop structure.

Level 2 detail is the minutia: ridge endings, bifurcations, dots, islands, and other individual ridge events that carry most of the individualisation weight in a fingerprint comparison. The full catalogue of Galton features and their statistical models is examined in the companion module topic. The biological basis for why these features are both individual and persistent is covered in the friction ridge anatomy and in-utero development topic. The number of corresponding minutiae and their positional relationships constitute the core of any positive identification opinion.

Level 3 detail is the finest resolution: the position and shape of individual sweat pores along a ridge crest, the precise contour of a ridge's edges, and any micro-features such as incipient ridges or scars. Level 3 detail is reliably interpretable only in high-quality prints (inked rolled prints, high-resolution scans at 1000 ppi or above) and is used in some jurisdictions to supplement Level 2 comparisons, particularly in the Netherlands and in some UK fingerprint unit workflows.

Pattern type (Level 1) is applied in ACE-V as an exclusionary gate before Level 2 comparison begins. If the latent print's Level 1 detail is clearly a plain arch, and the reference print from the suspect is clearly a double-loop whorl, no further comparison is required: the prints are excluded on Level 1 alone. This exclusionary use of Henry's pattern vocabulary means that the classification framework established at Scotland Yard in 1901 is a structural component of the modern ACE-V examination, not merely a historical artefact.