Minimum Number of Individuals
Minimum Number of Individuals (MNI) is a primary quantitative metric in zooarchaeology, paleontology, and bioarchaeology that defines the smallest possible number of distinct individual organisms necessary to account for all identified skeletal and dental elements in a taxonomic assemblage.
The Minimum Number of Individuals (abbreviated MNI) is a foundational derived quantitative unit in zooarchaeology, vertebrate paleontology, and human osteological paleodemography that represents the smallest number of distinct individual organisms required to account for all recovered skeletal and dental specimens assigned to a given biological taxon within a defined archaeological context. Originally applied in vertebrate paleontology by Chester Stock (1929) and Hildegarde Howard (1930) to quantify fossil bird and mammal tar-pit assemblages and introduced into North American archaeology by Theodore E. White in 1953, MNI was designed to overcome a major taphonomy bias inherent in raw fragment counts (NISP, the Number of Identified Specimens): namely, that different species possess different numbers of bones in a living skeleton and that human marrow fracturing, pot-sizing butchery, and sediment compaction shatter single bones into dozens of interdependent fragments.
The Fundamental Quantification Triad: NISP, MNI, and MNE
To reconstruct prehistoric subsistence economies, herd management, or commingled mortuary populations, osteologists balance three hierarchical counting units—NISP, MNE, and MNI—alongside Binford's MAU (Minimal Animal Units):
- Number of Identified Specimens (NISP): The direct observational tally of all individual bone, tooth, and antler fragments in an assemblage identified to a specific taxon (sometimes augmented by biomolecular collagen fingerprinting via Zooarchaeology by Mass Spectrometry). While NISP is additive across stratigraphic layers and requires no mathematical derivation, it suffers from specimen interdependence (counting 5 fragments of the same shattered femur as 5 units) and skeletal multiplicity (a pig has more foot phalanges and teeth than a bovid, and a teleost fish has hundreds of vertebrae and scales).
- Minimum Number of Elements (MNE): Introduced by Lewis R. Binford (1984) and Richard G. Klein & Kathryn Cruz-Uribe (1984), MNE estimates the minimum number of complete anatomical skeletal elements (for example, total left plus right humeri, or total mandibles) required to produce the observed fragments of that element, regardless of right/left body side. Analysts determine MNE by recording diagnostic anatomical zones (such as the deltoid tuberosity, nutrient foramen, or >50% of the distal trochlea) or by manual overlap and mechanical conjoining analogous to lithic refitting.
- Minimum Number of Individuals (MNI): Derived from the sided MNE counts for a taxon. In White's (1953) baseline definition, all specimens of a species are sorted by skeletal element and anatomical side (Left [L], Right [R], or Axial [A]), and the basic MNI equals the single most abundant sided diagnostic element:
MNIstandard = max(MNEL, MNER, MNEA / k) where k is the number of times that axial or repetitive element occurs in a single living skeleton (for instance, k = 1 for the atlas or axis vertebra, k = 7 for mammalian cervical vertebrae). For example, if a Magdalenian reindeer (Rangifer tarandus) assemblage contains 19 left distal tibiae and 14 right distal tibiae, the standard unadjusted MNI is 19 reindeer—because all 14 right tibiae could theoretically pair with 14 of the 19 left tibiae.
- Minimal Animal Units (MAU and %MAU): To evaluate differential skeletal part transport (the "Schlepp effect," where hunters abandon heavy axial bones at kill sites and carry high-utility limb quarters back to base camp), Binford (1978, 1984) defined MAU = MNEelement / k, where k is the frequency of that element in one complete skeleton (k = 2 for paired limbs, k = 1 for cranium/sacrum, k = 8 for proximal phalanges in artiodactyls). Normalizing each element's MAU by the highest MAU in the assemblage yields %MAU, which correlates against meat/marrow utility indices and bone mineral density.
| Quantification Metric | Abbreviation & Primary Pioneers | Mathematical / Operational Formula | Additivity Across Contexts | Sensitivity to Fragmentation | Primary Analytical Strength & Limitation |
|---|---|---|---|---|---|
| Number of Identified Specimens | NISP (Traditional tally) | Raw count of identified bone/tooth fragments | 100% Additive across strata | High (inflated by marrow smashing & taphonomy) | Simple & reproducible, but assumes fragments come from independent animals |
| Minimum Number of Elements | MNE (Binford 1984) | Overlap of diagnostic landmarks per skeletal bone | Non-additive | Low (collapses broken shaft splinters) | Basis for skeletal completeness & butchery cut-mark frequencies |
| Standard Minimum Number of Individuals | MNI (White 1953) | max(L, R) of most abundant element per taxon | Non-additive (sensitive to aggregation) | Insensitive to fragmentation once landmarks survive | Eliminates skeletal element count differences across taxa; exaggerates rare species |
| Paired / Lincoln–Petersen Estimate | MNIpair & LPI | L + R − P (MNIpair); (L × R) / P (Petersen *N**) | Non-additive | Requires intact osteometric & ontogenetic landmarks | N* estimates original dead population size rather than minimum recovered sample |
| Minimal Animal Units | MAU / %MAU (Binford 1978, 1984) | MNEelement / (count in 1 complete skeleton) | Non-additive | Low | Ideal for utility curves, carcass transport, and density-mediated attrition |
Bilateral Pairing, Ontogenetic Matching, and the Lincoln–Petersen Index
White's unadjusted MNI = max(L, R) assumes maximum anatomical pairing between left and right elements. In 1969–1971, Sándor Bökönyi, Raymond E. Chaplin, and Grover S. Krantz demonstrated that closer inspection of bilateral symmetry, epiphyseal fusion stage, dental eruption/wear, pathological lesions, and osteometric dimensions frequently proves that a left element and a right element cannot belong to the same individual (for example, an unfused juvenile left femur cannot pair with a fully fused adult right femur). If L left elements and R right elements yield P bilateral anatomical pairs from the same individuals, the pair-matched MNI increases to:
MNIpaired = L + R − P
Taking bilateral pairing a step further, Krantz (1968), Fieller & Turner, and Poplin showed that pairing left and right bones is mathematically isomorphic to capture–recapture ecology (the Lincoln–Petersen Index, LPI). Treating the L left elements as the first "marked sample," the R right elements as the second "recaptured sample," and the P bilateral pairs as the recaptured marked individuals yields an unbiased estimator of the original killed population size N* deposited at the site (or Chapman's small-sample bias-corrected estimator):
N* = (L × R) / P, or NChapman = [(L + 1)(R + 1) / (P + 1)] − 1
Using osteometric and visual pairing in commingled human ossuaries for paleodemography, Bradley J. Adams and Lyle W. Konigsberg (2004) formalised the Most Likely Number of Individuals (MLNI), providing hypergeometric confidence intervals for the true burial population size when taphonomic recovery is incomplete.
The Aggregation Effect and Power-Law Scaling with NISP
In Quantitative Zooarchaeology (1984), Donald K. Grayson demonstrated two structural mathematical properties of MNI that govern how zooarchaeologists interpret faunal tables:
- The Aggregation Problem (Non-Additivity): Because MNI is a non-linear maximum function, the MNI of a site depends heavily on how the analyst groups excavation units, pits, and stratigraphic layers recorded on the Harris matrix. Under the Minimum Distinction Method (pooling the entire site or major horizon as one aggregate), left and right bones from different pits can be paired together, producing the lowest possible MNI. Under the Maximum Distinction Method (calculating MNI separately within every individual hearth, trash pit, and 1-m grid square and summing them), a single butchered animal whose limbs were shared across four households (a process often tracked alongside tool use-wear analysis and the butchery chaîne opératoire) will be counted four separate times, inflating MNI toward NISP.
- Allometric Power-Law Dependence on NISP: Across hundreds of archaeological and paleontological assemblages, Grayson (1984) and R. Lee Lyman (2008) showed that MNI is tightly predicted by NISP according to a tight power-law regression:
MNI = a (NISP)b, with exponent b ≈ 0.50 to 0.75 Because b < 1, a rare taxon represented by a single bone fragment always has MNI = 1 (100% of NISP), whereas a dominant staple species represented by 1,000 fragments may have an MNI of only 30 to 50. Consequently, MNI systematically over-represents rare species when sample sizes are small.
Frequently Asked Questions
Why is MNI never simply the total bone count divided by the number of bones in a skeleton?
Animals rarely enter an archaeological site as complete skeletons, and different bones survive taphonomy at vastly different rates: dense enamel teeth, mandibles, astragali, and distal humeri resist carnivore chewing and soil leaching, whereas spongy porous vertebrae, ribs, and proximal humeri are preferentially destroyed. Therefore, MNI must be anchored to the single best-preserved, non-repeating anatomical landmark.
When should a zooarchaeologist rely on NISP versus MNI?
Most modern zooarchaeologists (following Grayson 1984 and Lyman 2008) report both NISP and MNI side by side. NISP provides an additive primary measure of ordinal taxonomic abundance when fragmentation is uniform across similar-sized mammals, while MNI (and MNE/MAU) provides a necessary conservative lower bound whenever whole articulated skeletons (such as microfauna, dog burials, or complete fish) would otherwise artificially inflate NISP for a single individual.
How does the Lincoln–Petersen Index (LPI) differ from MNI?
MNI measures only the minimum number of individuals physically present in the excavated sample on the laboratory table, which shrinks if only 10% of a site is excavated. The Lincoln–Petersen Index (LPI) and MLNI use the ratio of matched bilateral pairs (P) to unmatched left and right bones (L, R) to estimate the total original population of dead animals deposited at the site before taphonomic loss and partial excavation.
References
- White, T. E. (Apr 1953). "A method of calculating the dietary percentage of various food animals utilized by aboriginal peoples." American Antiquity, 18(4), 396–398. https://doi.org/10.2307/277116
- Grayson, D. K. (1984). Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas. Academic Press. https://openlibrary.org/works/OL4297365W
- Binford, L. R. (1984). Faunal Remains from Klasies River Mouth. Academic Press. https://openlibrary.org/works/OL2701262W
- Adams, B. J., & Konigsberg, L. W. (Jan 12, 2004). "Estimation of the most likely number of individuals from commingled human skeletal remains." American Journal of Physical Anthropology, 125(2), 138–151. https://doi.org/10.1002/ajpa.10381
- Lyman, R. L. (Mar 31, 2008). Quantitative Paleozoology. Cambridge University Press. https://doi.org/10.1017/CBO9780511813863