Let us also quote F.-E.Verbanck (article from the magazine "La Vie Canine” in June 1964  named "Several varieties but only one breed") that states the following :
"Breeding between different varieties has to be done with carefully thought out trials,performed
with a specific aim, and foreseen as a future business, knowing that positive results will not be obtained immediately, but only after several generations. Our opinion has not changed, on the contrary, breeding between different varieties, performed by serious breeders, have proved to 
be right. Each time that a strengthening is necessary in one of the varieties 
of our Belgian Shepherds, it is only in the other varieties where we have to 
find the elements." 

The arguments to justify the exclusion of inter-variety breeding are not absent. At the hair texture level, mixing long hair, short hair, rough hair gives a quite big variety of patterns due to a phenomenon of incomplete dominance. At color level we can equally find arguments. As heterozygotes are often less nice than homozygotes, breeders, on behalf of quality, are interested in selecting the varieties for themselves and not in breeding them amongst themselves. That leads to genetic isolation, morphological and/or temperament differences. (‘working dog’ on one side, ‘companion dog’ on the other, for example).  

We still have to answer the question to know what is the place of the varieties in the genetic arrangement of a breed? The reasons in favour of a competent genetic variability in the Belgian shepherd are innumerable. Let us name some of them :

-          keep good breeding qualities for the Belgian Shepherd (fertility, strength, etc.). As many anomalies were recessive, the reduction of the genetic variability increased the outcome frequency. 

-          keep for the Belgian shepherd a certain homogeneity among the different varieties. It is necessary to avoid too evident morphological and temperament differences between the different varieties.

-          keep for the Belgian shepherds possibilities of evolution or modification which imply maintaining a certain variation. The model researched at a certain time will not necessarily be the model 30 years later.                                           

As a conclusion, I do not find any inconvenience to define well precise varieties, nor to prefer them, but I do find many problems that impoverish the breed by depriving it of valuable elements whose only defect is a question of color, developed according to changing opinions through the years. Let us use our energy to hunt hereditary illnesses like epilepsy or sterility, defects of structure like hypertype (light structure, head too long, no stop), teeth defects like prognatism and character defects like shyness - specimens with timid characters which cannot look after property or are incapable of defending her owner. Belgian shepherd, before all color consideration, is or should be a shepherd dog with all that that means such as construction type, about aspect and character. All this seems to me, more important than eliminating certain colors, which historically and genetically make up the color parameters of the Belgian Shepherd.