-4- 
To qualify for retrieval, a record must match one of the data 
for every criterion cited. The retrieved records can then be set 
in an order given by further criteria. A useful ordering is 
by geographical co-ordinates and collectors. Classes of data 
in the record can be chosen to appear in the print-out of the 
retrieved information. In addition, individual kinds of data 
can be specified for listing in alphamerical sequence, with 
counts of the numbers of records in which each citation appears. 
Where geographical co-ordinates are specified for this, a dis¬ 
tribution map of the collecting sites is given on a computer- 
linked plotter. 
5. DATA COMPILATION AND RUN TIMES : Every effort has been made 
to strike a balance between the needs of both simplicity and 
shortening the time for compiling the raw data. These factors 
must also be seen in the need for avoiding over-long run times 
with modest computing facilities. Records with an average 
amount of data can be compiled at a rate of about 30 per hour, 
and proof-read at about 120 per hour. An IBM 1130 (16K core, 
512K disks) took, for 1000 records, 39 minutes for full editing, 
16 minutes for entry to the bank, 3 minutes for a simple retrieval 
operation and 4 1/2 minutes for plotting a small distribution 
map. These compilation and run times appear to be reasonably 
acceptable for a comparatively modest-sized machine. 
A FORTRAN-IV program listing will be published shortly, and 
a general account of the system has been accepted for publication 
in an early issue of TAXON. At present the system is being adapted 
for a UNIVAC 1106 with remote-access and time-sharing facilities. 
Further inquiries from readers of MUDPIE would be welcome.--A. V. 
Hall, Asst. Curator, Bolus Herbarium, University of Cape Town, 
Rondebosch, C.P., Africa. 
AUTOMATION OF DATA IN SYSTEMATIC COLLECTIONS 
The Conference of Directors of Systematic Collections, which 
includes 31 representatives of most of the larger institutions in 
the United States and Canada, established at its meeting in May, 
1970, a committee to study automation of data concerning specimens 
in natural history collections. This committee, composed of R. S. 
Cowan, Chairman (Smithsonian)? Roland Force (Honolulu); N. G. 
Hairston (Michigan)? and P. S. Humphrey (Kansas), reported to the 
1971 meeting of the Conference. The group is primarily concerned 
with data "about" specimens, not in data derived from the study of 
the specimens. They are concerned not only with improvement of 
the ways of collecting and distributing this type of data, but 
also with ways to integrate inter-museum objectives in a cohesive 
plan. The Committee has concluded that both purposes are desir¬ 
able, and has indicated certain stipulations and goals for immedi¬ 
ate attention. They will attempt to develop format in such a way 
that restrictions on computers, systems (and hopefully, languages! 
—ed. note) will be minimal. They are interested in inter-museum 
