This appears to be one of the several slips in Kew Plant List. I have often been telling my students: "A species will either have no infraspecific taxon (subspecies, variety, forma), or it will have a minimum of two". If there is no other variety within (at least I could not find) C. benghalensis, then there is no sense in writing C. benghalensis var. benghalensis. There is no need to write even var. typica, if no other variant exists.
It is like this. Supposing there is a type a (herbarium specimen) on the basis of an author names and describes a particular species, say Commelina benghalensis in this case, all specimens identified with this species (say a1, a2, a3, a4, etc) will get this name. Now supposing an author (same or some one else) discovers a specimen (say type b) which he thinks belongs to the same species but differs in certain characters from type a, and gives it a name say Commenlina benghalensis var. gargii, it will automatically establish the type variety (type a) which will be named as var. bengalensis (with no author citation, because it is automatically created; earlier authors used to call it var. typica).
Perhaps we will have to find if there is any other variety ever described within C. benghalensis, to automatically establish var. benghalensis. If there is none var. benghalensis is redundant.
Dr. Gurcharan Singh
Retired Associate Professor
SGTB Khalsa College, University of Delhi, Delhi-110007
Res: 932 Anand Kunj, Vikas Puri, New Delhi-110018.
Phone: 011-25518297 Mob: 9810359089http://people.du.ac.in/~singhg45/