SPECS Interfaces 



The SPECS computer program system (Special Program for Ecological Science) 

 was developed for use with the experimental and long-term biological data of 

 Dr. Robert J, Livingston at Florida State University. While SPECS provides the 

 capability to retrieve and sort data-base information and to calculate values 

 of biological indices, it has only a limited ability to make these results 

 available in a form compatible with higher-level packages such as BMDP, SPSS, 

 EZGRAF, and MINITAB. MATRIX can act as a powerful interface between SPECS and 

 these programs. The SPECS data base comprises the following types of data: 

 fish, invertebrate, plant, trophic, and physical-chemical. Using the SPECS 

 LOAD and SORT commands, these data can be retrieved for any area(s), station 

 (or group of stations), and date (or range of dates). The resulting file is 

 called a load/sort file and may be input to MATRIX GENERATE using one of the 

 predefined formats described in Table A. Notice that, for each data type, 

 there are several date options. Prudent selection of one of these can greatly 

 reduce the user effort required for the collapse procedure soecif ication. For 

 example, suppose a load/sort invertebrate file is input to GENERATE and the 

 rows of the matrix file are to be individual months from January 1978 through 

 December 1982 (60 months). If the full date format is used, the date key 

 values will be listed as individual days (YYMMDD). It could be tedious here to 

 specify a monthly collapse procedure, because all the numerical assignments for 

 the days in 01/78 would have to be entered, then all the assignments for 02/78, 

 and so forth for possibly all of the 60 months. If the data are read with the 

 year/month format, the day field would be skipped and the listed values would 

 be YYMM (i.e., the monthly collapse is accomplished by the format instead of a 

 laborious user response). The user could then simply enter 999*1 and a new row 

 would be generated for each month. If each row were to represent one of the 12 

 months of the calendar year (i.e., row 1 would represent all January's, row 2 

 all February's, etc.), the "month only" format would be appropriate. This for- 

 mat causes the day and year parts of the date to be ignored, leaving only 12 

 possible values for the date key. 



The SPECS CALC command computes ten separate diversity, richness, and 

 evenness indices along with the total number of individuals and number of spe- 

 cies. These variables may be calculated for any area(s), station(s), date(s), 

 or time(s) of day or any combination thereof (see SPECS manual for details). 

 CALC outputs two files. One (keyword OUTPUT) is suitable for printing; the 

 other (keyword PLOTDAT) is suitable as input to MATRIX GENERATE. The use of 

 the MATRIX program on a SPECS CALC output file is the simplest way to make 

 these computed variables available for plottinq and/or statistical analysis. 



The SPECS and MATRIX systems can be run with maximum efficiency if the 

 user gives forethought to exactly what information is needed for his analysis. 

 A combination of LOAD, SORT, and SLECT orocedures in SPECS can be used to get 

 an input file for MATRIX with little or no extraneous data. If, for example, 

 the fish data for all dates and stations were retrieved to a load/sort file and 

 input to GENERATE when only the data for stations 3 and 5A from February 1978 

 through June 1980 were needed, two things would happen. First, MATRIX would 

 have to read a great deal of nonrelevant data, which would result in wasted 

 computer time and money. Second, there would be a very large number of key 

 values listed in the collapse procedure, so more user time and effort would be 

 required to specify the collapse correctly. The LOAD command causes an entire 

 data base to be read. The records that match the load parameters are written 

 to an output file. The SLECT command reads a load/sort file and writes the 

 records that match its parameters to a smaller load/sort file. If many subana- 

 lyses are to be run on a group of data, a LOAD command should be used to 

 retrieve all the data that will be required for all the analyses; therefore the 



141 



