The experiment is the LTER nutrient by warming factorial treatments, which has 4 randomized blocks (Shaver et al.?). We randomly collected 3 individual ramets of Betula nana from each treatment within each block. The most recent 15 years of growth were analyzed for each ramet. Stems were cut and cross-sections made by hand with a razor blade. Ages were determined by staining the sections with 1% phloroglucinol in 20% HCl and counting the growth rings using a compound light microscope at 100x magnification. Stems were separated into different age classes trying to get at least one piece with same age on both ends for each age class. Lengths and weights of pooled segments for a given age class were determined after drying samples at 70 degrees C for 48 hours.
Because no external bud scars were usually visible, the boundary between age classes of stems usually could not be known with certainty. The lengths and weights of stem pieces that overlapped one year boundary were assigned half to one age and half to the next. Lengths and weights of pieces that overlapped more than one age class were assigned to the different age classes by dividing by one less than the number of years covered, and assigning half that amount to the two end pieces and the whole amount to intervening years, to avoid a bias toward the end pieces. When the end of the piece was the bottom of a branch, and therefore that boundary was known with certainty, the piece was divided by x.5, where x is one less than the number of years that the piece overlapped. Half that amount was assigned to the upper part of the segment, where the boundary was not known with certainty, while the whole amount was assigned to the other year classes covered by that piece, including the oldest piece. These procedures avoided a bias in assigning weight and length disproportionately to the years where a cut was made.
Betula nana makes two types of stem segments, short shoots and long shoots. Long shoots could be aged by making stem cross-sections, but short shoots could not because they have so little secondary growth. However, the maximum age of short shoots is one year less than the long shoot to which they are attached. Short shoot weight and length was assigned by dividing the total weight or length by the maximum age of those shoots and assigning that to each year class up to that age. Short shoots occasionally convert to long shoots, in which case they put on secondary growth and behave like a long shoot from then on (but obviously the stem segments covering the period of short shoot growth do not increase in length). These conversions from short to long shoots were classified separately in the data table.

Notes: Betula nana rep 1 from block 3 was excluded because it was diseased. No other data is missing. Data in columns with stippled headings were measured data, other columns are calculated from that data. Data from pieces that were overlapping age classes but were not conversions are entered in measured data columns as additions onto the original numbers (e.g. the formula shows =x+y+z).

See paper: Bret-Harte, M.S., G.R. Shaver, and F.S. Chapin III. 2002. Primary and secondary stem growth in arctic shrubs: implications for community response to environmental change. Ecology 90: 251-267.