The controversy on how to measure fine root production of forests (P) most accurately continues. We applied four different approaches to determine annual rates of P in an old-growth temperate Fagus sylvatica–Quercus petraea stand: sequential soil coring with minimum–maximum calculation, sequential coring with compartmental flow calculation, the ingrowth core method, and a recently developed root chamber method for measuring the growth of individual fine roots in situ. The results of the four destructive approaches differed by an order of magnitude and, thus, are likely to introduce large errors in estimating P. The highest annual rates of P were obtained from the sequential coring approach with compartmental flow calculation, intermediate rates by sequential coring with minimum–maximum calculation, and low ones by both the root growth chamber and ingrowth core approaches. A carbon budget for the stand was set up based on a model of annual net carbon gain by the canopy and measurements on carbon sink strength (annual leaf, branch and stem growth). The budget implied that a maximum of 27% of the net carbon gain was available for allocation to fine root growth. When compared to the carbon budget data, the sequential coring/compartmental flow approach overestimated annual fine root production substantially; whereas the ingrowth core and root growth chamber approaches grossly underestimated P rates. With an overestimation of about 25% the sequential coring/minimum–maximum approach demonstrated the best agreement with the carbon budget data. It is concluded that the most reliable estimate of P in this temperate forest will be obtained by applying the sequential coring/minimum–maximum approach, conducted with a large number of replicate samples taken on a few dates per season, in conjunction with direct root growth observation by minirhizotrons.