Article 1

Regulation (EEC) No 2967/85 is amended as follows:

1. 1..

   In Article 3, paragraphs 1 and 2 are replaced by the following:

   ‘1.

   The standard statistical method for assessing the lean meat content of pig carcases authorised as a grading method within the meaning of Article 2(3) of Regulation (EEC) No 3220/84 shall be either ordinary least squares or rank reduced methods, but other statistically proven methods may be used.

   The method shall be based on a representative sample of the national or regional pigmeat production concerned, consisting of at least 120 carcases whose lean meat content has been ascertained in accordance with the dissection method laid down in Annex I to this Regulation. If multiple sampling methods are used, the reference shall be measured on at least 50 carcases and the precision shall be at least equal to that obtained using the standard statistical method on 120 carcases using the method in Annex I.

   2.

   Grading methods shall be authorised only if the root mean squared error of prediction (RMSEP), computed by a full cross-validation technique, is less than 2,5. In addition, any outliers shall be included in the calculation of RMSEP.’.

2. 2..

   in Annex I, point 2 is replaced by the following:

   ‘2.

   The reference lean meat percentage is calculated as follows:

   $$\mathrm{y\; =\; 0,89\; \times \; 100\; \times }\frac{\mathrm{Weight\; of\; tender\; loin\; +\; weight\; of\; lean\; (fascia\; included)\; in\; shoulder,\; loin,\; ham\; and\; belly}}{\mathrm{Weight\; of\; tender\; loin\; +\; weight\; of\; dissected\; cuts}}$$

   The weight of the lean in those four cuts will be calculated by subtracting the total of the non-lean elements of the four cuts from the total weight of the cuts before dissection.’.