# Speed, Stride Frequency and Energy Cost Per Stride: How Do They Change With Body Size and Gait?

> Why does a mouse burn energy six times faster than a horse when they run? Timing the strides of 16 species on a treadmill, from mice to horses, two Harvard researchers found that each kilogram of animal uses about the same energy per stride at comparable speeds. Small animals simply take many more strides, and each gait’s speed range and stride rhythm change in a predictable way with body size.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/1988-heglund-taylor-speed-stride-frequency-energy/

## Citation and access

- **Authors:** Norman C. Heglund and C. Richard Taylor
- **Year published:** 1988
- **Journal or source:** Journal of Experimental Biology
- **Full citation:** Heglund, N. C. and Taylor, C. R. 1988. Speed, stride frequency and energy cost per stride: how do they change with body size and gait? Journal of Experimental Biology 138: 301–318.
- **DOI:** 10.1242/jeb.138.1.301
- **Access status:** Hosted (freely available elsewhere)
- **PDF:** https://www.originalwisdom.com/wp-content/uploads/2026/11/1988_Heglund_and_Taylor_speed-stride-frequency-and-energy-cost-per-stride.pdf

## Tags

- **Species:** Mammals > Blue wildebeest; Mammals > Brown rat; Mammals > Cattle; Mammals > Domestic dog; Mammals > Donkey; Mammals > Eastern chipmunk; Mammals > Eland; Mammals > Goat; Mammals > Grant’s gazelle; Mammals > Horse; Mammals > House mouse; Mammals > Kirk’s dik-dik; Mammals > Sheep; Mammals > Suni; Mammals > Thirteen-lined ground squirrel; Mammals > Waterbuck
- **Region:** Africa > Kenya; North America > United States
- **Tracking:** Gaits
- **Behaviour and ecology:** Behaviour; Energy budget; Movement
- **Methods and technology:** Body-size comparison; Linear regression
- **Data type:** Existing data; New field data
- **Observer qualification:** Not needed
- **Discipline:** Behaviour and ecology
- **Year published:** 1988

## Detailed keywords

- **Species:** Studied (16 species of quadrupeds, running on a treadmill): laboratory or white mouse (Mus musculus), eastern chipmunk (Tamias striatus), 13-lined ground squirrel (Spermophilus tridecemlineatus), white rat (Rattus norvegicus), domestic dog (Canis familiaris), suni (Nesotragus moschatus), dik-dik (Madoqua kirkii), Grant’s gazelle (Gazella granti), African domestic goat (Capra hircus), fat-tailed sheep (Ovis aries), wildebeest (Connochaetes taurinus), waterbuck (Kobus defassa), zebu cattle (Bos indicus), eland (Taurotragus oryx), domestic donkey (Equus asinus), and ponies and a horse (Equus caballus). Body masses ranged from 0.029 kg (mouse) to 680 kg (horse). Other animals mentioned: wildebeest, gazelles and zebras observed migrating across the African plains in an earlier study (Pennycuick 1975).
- **Location and habitat:** Laboratory treadmill studies. The African animals were captured or bought in Kenya and housed at the East African Veterinary Research Organization, Muguga, Kenya. The rodents, dogs, ponies and horse were captured or bought in the United States and housed at the Concord Field Station, Harvard University, Bedford, Massachusetts, United States.
- **Topics:** gaits (walk, trot, gallop); running speed; stride frequency; trot-gallop transition; “equivalent” speeds (minimum, preferred and maximum sustained speeds within a gait); preferred speeds in nature; scaling with body size (allometry); energy cost of locomotion; cost per kilogram per stride; muscle activation and force generation costs; elastic energy storage in muscles and tendons; mechanical advantage of the limbs; stride length; aerobic and anaerobic running.
- **Statistics or models:** Least-squares linear regression of stride frequency against speed for each gait; trot-gallop transition speed found where the trot and gallop regression lines cross; allometric equations (power functions of body mass) fitted by least-squares regression on log-transformed data, with 95% confidence limits and coefficients of determination (r²); means ± standard error. Inconsistencies in the paper: (1) the text on p. 316 says cost per stride at the trot-gallop transition scales with the -0.11 power of body mass, but Table 5 gives -0.011; (2) in Table 5 the preferred galloping exponent is printed as 0.04, with 95% confidence limits of -0.039 to 0.047, so 0.004 may be meant; (3) the size of “the horse” varies: the summary says “30 g mice to 200 kg horses”, the introduction compares a 30 g mouse with a 300 kg horse, p. 316 a 270 kg horse, and the minimum trotting speed range is given “from a 30 g mouse to a 170 kg horse”, while the heaviest animal in the tables is a 680 kg horse; (4) the Discussion says a mouse and a horse both increase their speed by about 2.6-fold “within each gait”, but the Results give 2.58-fold for the trot and 2.11-fold for the gallop.
- **Data type:** New data (treadmill measurements of speed and stride frequency), combined with existing data: some stride frequencies (mouse, rat, some dogs, horse) were published earlier by Heglund et al. (1974), and energy consumption data come from Taylor et al. (1982). Dates of data collection not given.

## Synopsis

### 1. What they did

Per gram of body, a small animal burns energy much faster than a large one when running. A 30 g mouse uses energy six times faster than a horse at the speed where each switches from a trot to a gallop, even though the mechanical work each gram of muscle does is about the same. The authors suspected that stride frequency (how many strides an animal takes each second) holds the answer.

They trained 16 species of wild and domestic four-legged mammals, from 29 g mice to a 680 kg horse, to run on a treadmill. Ten species came from Kenya (antelopes such as suni, dik-dik, Grant’s gazelle, wildebeest, waterbuck and eland, plus goats, sheep, zebu cattle and a donkey). Six came from the United States (mice, chipmunks, 13-lined ground squirrels, rats, dogs, ponies and a horse).

- Measuring strides. After at least several weeks of training, they timed 25 cycles of one foot at a steady treadmill speed, with a stopwatch for large animals and high-speed film (100 to 200 frames per second) for small ones.

- Finding “equivalent” speeds. For each animal they found five comparable speeds: the slowest trot, the preferred trot (the middle of the trotting range), the trot-gallop transition, the preferred gallop (the middle of the galloping range they measured) and the fastest sustained gallop.

- Scaling with body size. They worked out how these speeds and the stride frequencies used at them change with body mass.

- Energy cost per stride. They combined their results with published energy data for the same species to calculate the energy each kilogram of animal uses per stride.

Walking could not be analysed across body sizes: small animals would not walk steadily on the treadmill. Mice, chipmunks and rats ran to the front, stopped and rode to the back, over and over.

### 2. What they found

Trotting

- Small animals start trotting at much slower speeds. The slowest trot rose about nine-fold, from 0.19 m/s in a mouse to 1.8 m/s in a 170 kg horse.

- Every animal’s fastest trot was about 2.6 times its slowest trot, whatever its size.

- Stride frequency rose in a nearly straight line with speed, by about 1.55 times from the slowest trot to the trot-gallop transition, in all animals.

- Preferred trotting speed (in metres per second) is about 1.09 × body mass (kg) to the power 0.222. Stride frequency at that speed is about 3.35 × body mass to the power -0.130.

The trot-gallop transition

- At the transition speed, stride frequency was the same whether the animal trotted or galloped, so the transition can be found where the two lines meet.

- Transition speed rose with body mass to the power 0.216; stride frequency there fell with body mass to the power -0.150. A 30 g mouse began to gallop at 0.8 m/s, a 170 kg pony at 5 m/s.

Galloping

- The fastest sustained gallops were about 2.1 times the slowest gallops.

- Stride frequency barely changed with speed in a gallop: about 10% higher for double the speed. A mouse gallops at about 7 to 9 strides per second, a horse at about 2.

- Dogs and horses, built for endurance, gave much wider speed ranges than goats and cattle. Top speeds could not be measured, because only steady, aerobic running was studied.

Energy cost per stride

- All the preferred speeds rose with roughly the 0.21 power of body mass, and stride frequencies fell with roughly the -0.14 power.

- At four of the five equivalent speeds (all except the slowest trot), each kilogram of animal used about the same energy per stride, whatever its size. At the trot-gallop transition this was 5.34 ± 0.25 joules per kilogram per stride, from mice to horses.

- Cost per stride rose with speed: about 5.0 joules per kilogram at the preferred trot, 5.3 at the transition, 7.5 at the preferred gallop and 9.4 at the fastest sustained gallop (figures from the summary).

- At the slowest trot, cost per stride depended slightly on size: about 2.5 times higher in a 30 g mouse than in a 270 kg horse.

### 3. What they concluded

- Small animals are slower, have a narrower range of speeds and take more strides per second, but the relative changes in speed and stride frequency within each gait are similar from mice to horses.

- Because cost per kilogram per stride is about the same at equivalent speeds, the high energy cost of running in small animals is explained by their higher stride frequencies.

- The cost of running comes mainly from switching muscles on and off (activation, about 30% of muscle energy use) and from generating force over time (about 70%). Both should rise with stride frequency: muscles switch on and off once per stride, and faster strides need faster force generation.

- Cost per stride may rise with speed because longer strides change the leverage of the limb muscles, so they need more force for the same push on the ground. In a trot about three-quarters of the rise in cost can be explained by higher stride frequency; in a gallop, where speed rises almost entirely through longer strides, this leverage effect may matter more.

### 6. Other technologies, skills or innovations

- Treadmill running after weeks of training.

- High-speed cine cameras (Eclair GV-16 with a clock in view, and Photosonics 1PL with a clock mark on the film) to count strides in small animals.

- Defining “equivalent speeds” so animals of very different sizes can be compared fairly.

## Cite this summary

Lawrence, K. 2026. Summary of: Speed, Stride Frequency and Energy Cost Per Stride: How Do They Change With Body Size and Gait? (Heglund and Taylor, 1988). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/1988-heglund-taylor-speed-stride-frequency-energy/