Relative growth rate (RGR) is a prominent indicator of plant strategy with respect to productivity as related to environmental stress and disturbance regimes. RGR is the (exponential) increase in size relative to the size of the plant present at the start of a given time interval. Expressed in this way, growth rates can be compared among species and individuals that differ widely in size. By separate measurement of leaf, stem and root mass as well as LA, good insight into the components underlying growth variation can be obtained in a relatively simple way. These underlying parameters are related to allocation (leaf-mass fraction, the fraction of plant biomass allocated to leaf), leaf morphology (see Section 3.1), and physiology (unit leaf rate, the rate of increase in plant biomass per unit LA, a variable closely related to the daily rate of photosynthesis per unit LA; also known as net assimilation rate).

#### What and how to measure?

Ideally, RGR is measured on a dry-mass basis for the whole plant, including roots. Growth analysis requires the destructive harvest of two or more groups of plant individuals, grown either under controlled laboratory conditions or in the field. Individuals should be acclimated to the current growth conditions. At least one initial and one final harvest should be carried out. The actual number of plants to be harvested for a reliable estimate increases with the variability in the population. Size variability can be reduced by growing a larger number of plants and selecting *a priori* similarly-sized individuals for the experiment, discarding the small and large individuals. Alternatively, plants can be grouped by eye in even-sized categories, with the number of plants per category equal to the number of harvests. By harvesting one plant from each category at each harvest, each harvest should include a representative sample of the total population studied. The harvest intervals may vary from less than 1 week in the case of fast-growing herbaceous species, to more than 2 months or longer in the case of juvenile individuals of slow-growing woody species. As a rule of thumb, harvest intervals should be chosen such that plants have less than doubled mass during that interval.

At harvest, the whole root system is excavated and subsequently cleaned, gently washing away the soil (see details on procedure under Section 5). Plants are divided into three functional parts, including leaves (light interception and carbon (C) uptake), stem (support and transport) and roots (water and nutrient uptake, as well as storage). The petioles can either be included in the stem fraction (reflecting support; this is the preferred option), or combined with the leaf fraction (to which they belong morphologically), or they can be measured separately. LA is measured (for details, see Section 3.1) before the different plant parts are oven-dried for at least 48 h at 70°C and weighed.

Destructive harvests provide a wealth of information, but are extremely labour-intensive and, by their nature, destroy at least a subset of the materials being studied. Alternatively or additionally, growth can be followed non-destructively for several individuals (~10–15 per treatment), by non-destructively measuring an aspect of plant size at two or more moments in time. By repeatedly measuring the same individuals, a more accurate impression of RGR can be obtained. However, RGR cannot be factorised into its components then, and repeated handling may cause growth retardation. Ideally, the whole volume of stems (and branches) is determined in woody species (see Section 4.1), or the total area of leaves, in case of herbaceous plants. In the latter case, leaf length and width are measured, along with the number of leaves. To estimate LA, a separate sample of leaves (~20) has to be used to determine the linear-regression slope of leaf length × width.

#### How to calculate RGR?

From two consecutive harvests at times *t*_{1} and *t*_{2}, yielding plant masses *M*_{1} and *M*_{2}, the average RGR is calculated as

RGR = (ln*M*_{2} – ln*M*_{1}) / (*t*_{2} – *t*_{1}).

In the case of a well balanced design where plants are paired, it is probably simplest to calculate RGR and its growth parameters for each pair of plants, and then use the RGR values for each pair to average over the population. Otherwise, the average RGR over the whole group of plants is calculated from the same equation. In doing so, make sure to first ln-transform the total mass of each plant before the averaging. In the case of more than two harvests, average RGR can be derived from the linear-regression slope of ln(mass) over time.

The average unit leaf rate (ULR) over a given period is

ULR = [(*M*_{2} – *M*_{1}) / (*A*_{2} – *A*_{1})] × [(ln*A*_{2} – ln*A*_{1}) / (*t*_{2} – *t*_{1})],

where *A*_{1} and *A*_{2} represent the LA at *t*_{1} and *t*_{2}, respectively. Again, the simplest option would be to calculate ULR from each pair of plants.

Average leaf mass fraction (LMF) during that period is the average of the values from the 1st and 2nd harvest, as follows:

LMF=[(*M*_{L1} / *M*_{1}) + (*M*_{L2} / *M*_{2})] /2,

where *M*_{L1} and *M*_{L2} indicate the leaf mass at *t*_{1} and *t*_{2}, respectively.

Similarly, average specific leaf area (SLA) is calculated as follows:

SLA = [(*A*_{1} / *M*_{L1}) + (*A*_{2} / *M*_{L2})] /2.

*Special cases or extras*

(1) *Confounding effect of seed size*. Especially tree seedlings may draw on seed reserves for a long time after germination. Inclusion of the seed in the total plant mass underestimates the RGR of the new seedling, whereas exclusion of the seed mass will cause an overestimation of growth. For large, persistent seeds, the decrease in seed mass between *t*_{1} and *t*_{2} can therefore be added to *M*_{1} (excluding seed mass itself from *M*_{1} and *M*_{2}).

(2) *Ontogenetic drift*. As plants change over time, they readjust allocation, morphology and leaf physiology. Consequently, LMF, SLA and ULR may change with plant size, and RGR generally decreases over time, the more so in fast-growing species. This does not devalue the use of the RGR, as plant growth does not necessarily have to be strictly exponential. As long as plant growth is somehow proportional to the plant size already present, RGR is an appropriate parameter that encapsulates the average RGR over a given time period. However, ontogenetic drift is an important characteristic of plant growth, and a higher frequency of harvests may provide better insight into this phenomenon. In comparing species or treatments, it may be an option to compare plants at a given size or size interval, rather than over a given period of time.

(3) *Related to ontogenetic drift*, shrubs and trees accumulate increasing amounts of xylem, of which large proportions may die depending on the species. This inert mass would greatly reduce RGR. Previous studies express RGR (‘relative production rate’) on the living parts of large woody plants by treating the biomass increment over Year 1 as *M*_{1} and the increment over Year 2 as *M*_{2}. It could similarly be based on annual diameter (or volume) increments.

(4) *Smooth curves*. In the case of frequent (small) harvests, a special technique can be applied, in which polynomial curves are fitted through the data. This is an art in itself!

*References on theory*,* significance and large datasets*: Evans (1972); Grime and Hunt (1975); Kitajima (1994); Cornelissen *et al*. (1996); Walters and Reich (1999); Poorter and Nagel (2000); Poorter and Garnier (2007); Rees *et al*. (2010).

*More on methods*: Evans (1972); Causton and Venus (1981); Hunt (1982); Poorter and Lewis (1986); Poorter and Welschen (1993); Cornelissen *et al*. (1996); Rees *et al*. (2010).