I understand CRD and CRBD design in poultry feeding systems. However, if somebody already use Latin Square design in chicken feeding experiments for growth or egg production performance.
why do you want t use of Latin Square design ? Do you have some variations in your salon or farm? your place is windowless? and how is your ventilation?
Do you maintenance your animals in cages or ground? if you grow them in cage, how many floors do you have?
A latin square design is posible to use in feeding trail. It has to do with treatment assignment to the experimental units and also to have two sources of variation in addition to the teatment factor. I recommed you to read books such as those of Steel and Torrie and Cochran and Cox on experirmental designs.
Hi Fred I have read at least one paper where latin square design has been used in chicken feeding trials. It is possible to use this design. As indicated by Mohammed and Jose above you can use this design when you have two other factors (e.g. window vrs no window rooms or cage vrs ground floors etc) in addition to your treatment proper (e.g. different diets). Further details available if you know what you want to do. Good luck
Thank you Mohammad, Josh and Gariba for your wonderful suggestion. I would like to use open ventilation (windowless) and to maintain birds on the concrete floor filled with wood chips and 4 rooms as unit and 10 birds in each rooms, 4 treatments and these treatments to change/rotate for each rooms over a certain period (1 week). This rotation to be regarded as replicates. However, one thing not for sure is the age of birds over the experimental period. Hi Josh, thanks for the authors you mentioned is helpful.
It is a very unusual design more fitting to agronomic studies. however Siriwan et al. (1993) used it in their investigations on endogenous amino acid losses at the distal ileum in cockrels. If you work in this area you would know this paper as it is one of the seminal work. But I can mail it to you if you wish.
if you are having 4 rooms to be rotated then in this case you don't need latin square. Since you going to take the treatments round as replicates then you should just build an age effect into your design. This means you will have two main effects: Treatments and Age, which you would then evaluate using a factorial arrangement of your data. It might even turn out you dont find an age effect, making your data even easier to explain. Good luck.
Thanks David, I found Siriwan et al. (1993) from search engine, the procedure is helps. I appreciate your final suggestion. I will look through and consider, thanks!
Dear Fred
We can choose design according to the uniformity of the experimental units in our experiment , in case of Latin square design can be use when our experimental units have two direction of variation sources ( systematic variation ) , but this design has restriction rules like treatments = columns = rows , and the degree of freedom of the error very small ( d.f. = (t-1 ) (t-2 ) ) , so it is not suitable for small number of treatments .
I think that design is not common in use but the researcher may be forced to use this design .
Good Luck
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation