Estimating time of death from rectal temperature: the Henssge nomogram in practice

Why temperature, and why not a linear rule

A body does not cool at a constant rate. Death is followed by a temperature plateau during which the core barely cools (roughly 0.5 to 3 hours), then by an exponential decline that slows as body temperature approaches ambient temperature. The old linear rule (Glaister and Rentoul: 0.83 °C per hour from 36.9 °C) ignores this course. It underestimates the interval in the plateau and in the late phase, overestimates it when cooling is fast, and its error is of the order of ± 3 to 4 hours even under favourable conditions.

The Marshall and Hoare model and the Henssge nomogram

Henssge (1988) built on the double-exponential cooling model described by Marshall and Hoare (1962). It starts from the ratio Q = (T_r − T_a) / (T₀ − T_a), where T_r is the rectal temperature, T_a the ambient temperature and T₀ the core temperature at death (37.2 °C by default). The cooling rate is set by the constant B, which depends on body mass M and the corrective factor f: B = −1.2815 · (f · M)^−0.625 + 0.0284.

For ambient temperatures up to 23.2 °C, Q = 1.25 · e^Bt − 0.25 · e^5Bt; from 23.3 °C, Q = 1.11 · e^Bt − 0.11 · e^10Bt. The equation has no closed-form solution for t, so the nomogram solves it graphically and apps solve it numerically.

The corrective factor: the largest source of error

The corrective factor adjusts the effective body mass for clothing, covers, moisture and air movement. A value of 1.0 corresponds to a nude body or a single thin layer in still air. As a guide: a body in water 0.35 to 0.5, a wet body or moving air 0.7 to 0.75, one or two thin layers 1.1, heavier clothing 1.3, heavy clothing with a blanket 1.8 and strong insulation with a duvet 2.4. Under strong insulation the factor also depends on body mass (Henssge 1992). A factor outside 0.35 to 2.8 signals that the conditions should be reconsidered.

How wide is the 95 % interval

The permissible variation is not a constant. According to the nomogram legend (Henssge 1988; Madea 2016) it is ± 2.8 hours for Q above 0.5. For 0.5 ≥ Q above 0.3 it is ± 3.2 hours without and ± 4.5 hours with a corrective factor, and for 0.3 ≥ Q above 0.2 ± 4.5 and ± 7 hours respectively. The calculation is meaningful only down to the so-called critical Q: 0.2 for ambient temperatures below 23.3 °C and 0.5 from 23.3 °C. Below this limit even a small temperature measurement error produces a large error in the estimate (Sauer et al. 2024).

The interval should be read as orientational. A validation of the Henssge and Madea compound method on 76 deceased with known times of death (Heinrich et al. 2025) found the true time of death inside the 95 % interval in only 37 % of cases: 27 % for bodies stored cold before measurement and 62 % for bodies kept warm. The temperature estimate is therefore best combined with other findings such as livor and rigor mortis, as in the compound method (Henssge, Madea and Gallenkemper 1988).

When the result does not hold

• The ambient temperature was not constant (the body was moved, a window opened, heating changed, the scene was manipulated). An open window can cool a small room by several degrees in winter, and returning to the original temperature takes up to three times as long as the window was open (Sauer et al. 2024).
• The body lies in water or was soaked and the conditions changed.
• Fever or hypothermia before death: the starting temperature was not 37.2 °C.
• Measurement in the plateau (Q above 0.9) or below the critical Q (0.2, or 0.5 in warm surroundings).

How 4n6 Tools handles it

The Henssge tool in 4n6 Tools helps choose the corrective factor from clothing and surroundings, solves the equation numerically, states the interval for the Q band and shows a warning under unvalidated conditions (warm surroundings with a low Q, the plateau, an extreme factor). From the time of measurement it calculates the date and time of death and adds the result with its citations to the case report. The Glaister rule is available only for comparison.