Skip to content

Morisita horn similarity index

6 messages · moses selebatso, Martin Weiser, Roman Luštrik +2 more

Original
moses selebatso
# 
?Hello 
I am trying to analyse diet overlap (level of similarity) between two species. I have diet composition in %. I have tried to find the best tool, and thought Morisita horn will do, but I cant find the right package for. Is this the best tool?
Thank you,
Moses
Martin Weiser
# 
moses selebatso p??e v P? 27. 11. 2015 v 03:55 +0000:
Hi Moses,

this may be a matter of taste, but I like ordinations.
I assume you have data like this:

food.1 food.2... food.n species
2 2 ... 80 A
1 63 ... 5 A
30 20 ... 10 B
7 5 ... 70 B

Then I would run RDA - procedure rda() in the vegan package, with
something like: result<- rda(foodmatrix ~ SpeciesIdentity)
plot(result, disp=c("sp","cn"))
Interpretation of that plot is straightforward:
further the food.item is to the species.identity half, higher the
preferrence.

HTH,
Martin
Martin Weiser
# 
moses selebatso p??e v P? 27. 11. 2015 v 03:55 +0000:
Hi Moses,

this may be a matter of taste, but I like ordinations.
I assume you have data like this:

food.1 food.2... food.n	species
2 2 ...	80 A
1 63 ... 5 A
30 20 ... 10 B
7 5 ... 70 B

Then I would run RDA - procedure rda() in the vegan package, with
something like: 

result<- rda(foodmatrix ~ SpeciesIdentity)
plot(result, disp=c("sp","cn"))

Interpretation of that plot is then straightforward:
further the food.item is to the species.identity half, higher the
preference.

HTH,
Martin
Roman Luštrik
# 
Hi,

`vegdist` function of vegan package
<http://cc.oulu.fi/~jarioksa/softhelp/vegan/html/vegdist.html> implements
the morisita index. Is this what you're looking for?

Cheers,
Roman


On Fri, Nov 27, 2015 at 4:55 AM, moses selebatso <selebatsom at yahoo.co.uk>
wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Botta-Dukát Zoltán
    

    

      
      #
    

  


  

      
Dear Moses,

Vegdist supposes that input data are numbers of individuals not 
percentages. So, I'm afraid it cannot help to Moses.

Cheers,
Zoltan




2015.11.27. 13:17 keltez?ssel, Roman Lu?trik ?rta:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Jari Oksanen
    

    

      
      #
    

  


  

      
Vegan::vegdist has actually two related indices: Morisita index (?morisita?) that assumes that input data are integers (individuals, not percentages), and its Horn-Morisita (?horn?) variant that does not have this restriction. However, you should always check the formulae of indices, because the names may be used loosely. There is also Horn index that is not implemented in vegan::vegdist. For these indices, vegan follows Krebs?s ?Ecological Methodology?.

cheers, Jari Oksanen