MEDITS-based information on the deep-water red shrimps Aristaeomorpha foliacea and Aristeus antennatus ( Crustacea : Decapoda : Aristeidae ) *

The red shrimps Aristeus antennatus and Aristaeomorpha foliacea are the main target species of Mediterranean deep trawling. Both species were described in the Ligurian Sea, in the early nineteenth century by Risso, but were first exploited less than a century ago, while in some eastern areas of the Mediterranean their exploitation has just started. In updated systematic approaches (Pérez-Farfante and Kensley, 1997) A. antennatus and A. foliacea are the only Mediterranean representatives of the Aristeidae family. Important morphological characteristics of both species are i) light exoskeletons and long pleopods suggesting good swimming ability; ii) secondary sexual dimorphism concerning SCI. MAR., 66 (Suppl. 2): 103-124 SCIENTIA MARINA 2002


INTRODUCTION
The red shrimps Aristeus antennatus and Aristaeomorpha foliacea are the main target species of Mediterranean deep trawling.Both species were described in the Ligurian Sea, in the early nineteenth century by Risso, but were first exploited less than a century ago, while in some eastern areas of the Mediterranean their exploitation has just started.
In updated systematic approaches (Pérez-Farfante and Kensley, 1997) A. antennatus and A. foliacea are the only Mediterranean representatives of the Aristeidae family.Important morphological characteristics of both species are i) light exoskeletons and long pleopods suggesting good swimming ability; ii) secondary sexual dimorphism concerning body size and the rostrum, which is short in adult males and long in juveniles and adult females; and iii) an open telicum.The spermatophores are larger in A. antennatus, in relation to a greater fecundity: in fact the females of A. antennatus produce about four times more eggs than A. foliacea females of the same size (Orsi Relini and Semeria, 1983).The life history of the two species therefore begins with a very different energy budget and probably body development.
The distribution of the red shrimp recalls its origin in the Mesozoic basin of Tethys, which extended approximately from the present Caribbean to the Indian Ocean and included areas corresponding to the present Mediterranean.Red shrimps are therefore present in the Eastern Atlantic, the Mediterranean, and the Indian Ocean, but A. foliacea reaches as far as the Western Atlantic and New Zealand.
Information on their larval and early juvenile stages is scarce and particularly the distribution of larvae is still unknown.As a matter of fact, very young shrimps, less than 15 mm carapace length have been found generally in small quantities both in very deep waters (Sardà et al., 1994) and on fishing grounds (Martínez-Baños andMas, 1994, D'Onghia et al., 1997).Large numbers of juveniles have been followed regularly only in Sardinian waters (Mura et al., 1997).
What appears to be recruitment to the fishery (i.e. the first large group identified in length frequency distributions) are shrimps of both species of about 24-25 mm CL, but especially in A. antennatus, this pattern is not found in all fishing areas, and does not appear regularly in any given area each year.Sudden changes in the size structure of the fished stock of A. antennatus may occur also for sizes larger than those mentioned above (Orsi Relini and Relini, 1998a).
Consequently, though the population dynamics of A. foliacea seems a little clearer than that of A. antennatus (Ragonese et al., 1994;Spedicato et al., 1998;D'Onghia et al., 1998), given a more regular identification of cohorts in the fished stocks, the absolute age of recruits (i.e. if individuals of 24 mm CL represent the 0 group or the I or II group) remains an open issue.
Assessments of A. antennatus and A. foliacea have been attempted repeatedly.Table 1 (Fiorentino, 2000) gives the parameters used by different authors and their conclusions.It is well known that formulae for natural mortality (M) have been developed generally for fish stocks, in which the Von Bertalanffy parameters, and at times also longevity, have been introduced.Therefore the assessment of life span and growth can be the key to these attempts.Considering the evolution of growth studies, the emerging trend is to assign to red shrimps an increasingly longer life span and obviously a slower growth and a lower natural mortality rate.At present, the results of the first tagging attempts (Relini et al., 2000) are apparently supporting this new trend.
The main objective of this paper is to process MEDITS trawl survey data on depth distribution, abundance, length structure, and general biology (reproductive aspects excluded) of Aristeus anten-104 A. CAU et al.

MATERIALS AND METHODS
A data base gathered in six surveys (1994 to 1999) during the international MEDITS Trawl survey (MEDITS programme, Bertrand et al., 2000Bertrand et al., , 2002) ) from late spring to early summer was considered.In total, an average 360 hauls per year were carried out according to a depth stratified random methodology.An experimental net (with 20 mm cod end stretched mesh size) was used in the survey.The hauls had an average duration of 1 h, and were carried out only during daylight at depths of between 201 and 800 m.Further details on the methodology can be found in Bertrand et al. (2000Bertrand et al. ( , 2002)).
The north and central Adriatic Sea was left out due to the scarce presence of the target species.
For each RAR, year, and species, the Biomass Indices (BI; kg/km 2 ) were estimated by single and pooled strata falling in the species/area distribution (i.e., 201-500 m, 501-800 m, and overall slope: 201-800 m) and presented in a synoptic table together with the corresponding coefficient of variation (CV%).These CV% were expressed as the percentage ratio between the standard deviation (S) of the sample and the mean (CV=S/X*100).Original data are presented in Appendices 1-4.
For each shrimp specimen, sex was determined and the carapace length (CL) was measured to the nearest mm, from behind the orbit of the eye to the posterior border of the carapace.The percentage contribution of both females and males to the sex ratio of the samples was estimated by species and survey.The Length Frequency Distributions (hereafter LFD) were derived and graphically represented by sex and year, both as percentage frequency per length class and as box plots of median, minimum and maximum CL.
The similarity and homogeneity among the LFD of the different years were tested by comparing the medians with the non-parametric test of Kruskal Wallis (herein KW).The KW test was applied using STATISTICA software (StatSoft, Inc., 1995).The KW test statistics is a function of the ranks of the observations in the combined sample for K>2 independent samples (Conover, 1980;Zar, 1996) H 0 = all of the K sample distribution functions are identical; H 1 = at least one of the samples tends to yield a higher observed value than at least one of the other samples.
In the event of the H 0 hypothesis being refused, a multiple comparison procedure (Tukey test) was used to check which years contributed most to the difference of the medians.The probability level of significant difference was set at p=0.05.
The growth curves for the females of the two species of each RAR were identified.To this purpose the Modal class Progression Analysis (MPA) routine available within the software "FiSAT"   (Gayanilo et al., 1996) was used.This routine was employed to discriminate the components of the LFD.Subsequently, using the means of each identified age group, we made a non-linear estimate of the growth parameters of the von Bertalanffy Growth Function (Analysis of length at age data in FiSAT).When it was not possible to calculate the growth parameters directly from MEDITS data, the corresponding estimates were derived from the specific literature in order to allow a general comparison.
A two-by-two comparison of the growth curves for each species was carried out considering two RARs each time and calculating the area between the two curves.The calculation was performed using the integral of the absolute distances between the integration extremes, t 1 and t 2 , where t 1 = age of fishing recruitment and t 2 = maximum age hypothesised per species.
For Aristeus antennatus, t 1 was assigned a value of 0.75 and t 2 a value of 6.0, while for Aristaeomorpha foliacea, t 1 was assigned 0.75 and t 2 3.5.The values obtained represent the distances between two curves that are proposed as comparison indices: a high index value indicates a broader distance and so a higher diversity between two curves.
DEEP WATER RED SHRIMPS OF THE MEDITERRANEAN 107

Distribution and abundance
Based on the quantitative analysis of the space distribution of the catches (kg/km 2 ), the MEDITS data reflect a different distribution pattern for the two species.
The trend of the mean biomass indices recorded at depths of between 201 and 800 m in the six years on the whole (1994)(1995)(1996)(1997)(1998)(1999) (Tables 2-3: overall mean index) showed that A. antennatus was most abundant in the Sardinian Sea (RAR 3 with 5.42 kg/km 2 ), and least abundant in the easternmost part of the Mediterranean (Aegean Sea -RAR 6).On the contrary, A. foliacea showed fewer catches in the westernmost areas (Algerian Provencial -RAR 1) and greater yields in the centre (Strait of Sicily -RAR 4) (Tables 2-4).
In this macrostratum the most abundant catches of A. foliacea were recorded for RAR 2, 3, and 4. For these macroareas, the regression between biomass indices (except in 1994 for RAR 2 and 4) and the different MEDITS years showed a significant correlation (p>0.05) (Fig. 2).Concerning A. antennatus, high biomasses were estimated in RAR 1, 3, and 5, but the increase in catches in the years (positive trend) can only be seen in the Sardinian Sea (RAR 3) (p>0.05) (Fig. 3).
On analysing the single strata (201-500 and 501-800 m) we observed that the largest catches in both species were made at 501-800 m, while above this depth (201-500 m) the catches were smaller (Tables 2-3).

Sex ratio
A total of 60,460 specimens of A. foliacea were measured and sexed throughout the MEDITS area.A prevalence of females can be generally observed, representing 60% of the sampled total, except for the Aegean Sea (RAR 6), where only 46% out of a total of 1,961 individuals were females (Table 5).
A sharp prevalence of the number of females was always observed for A. antennatus, where they represented 77.4% out of a total of 40,827 specimens (Table 6).

Aristaeomorpha foliacea
The size frequency distributions analysed for the six years of the study in A. foliacea showed a trend in which for both sexes in all RAR's (except RAR 1), the fishing recruits are clearly distinct from the parental stock section, which presents multiple modes in the females and a strong asymmetry in the males (Fig. 4).
From an analysis of the basic statistics, it can be shown that, in the entire investigated area, the largest size found in the six years was 74 mm CL for the females (Straits of Sicily -RAR 4).From a more detailed analysis of the individual macroareas, the largest sizes were always larger than 60 mm CL, except in the Aegean Sea (RAR 6) and in 1994 in the Ionian Sea (RAR 5).The smallest sizes, on the other hand, measured between 10 and 25 mm CL (Table 5; Fig. 5).
Except for RAR 1, for which a statistical comparison was not possible because the catches were too few, the medians were significantly different for all RAR's, considering the general comparison among all the years of the survey (KW test, p<0.05) (Fig. 5).Statistically significant similarities were found in the Tukey test multiple comparison for pairs of years (Table 5).
In the Tyrrhenian Sea (RAR 2) the median was significantly similar in 1998 and 1999 both for the males and for the females (Tukey test, p>0.05) (Table 5, Fig. 5), while in the Sardinian Sea (RAR 3) the same similarity was observed between 1996 and 1999 for both sexes, and for the males also in 1997.In the Straits of Sicily (RAR 4), except for the 1994 and 1998 surveys, the medians in the females did not show and significant differences, while in the males there was a reduced variability.In the Ionian Sea (RAR 5), except for 1995, the females always showed similarities between medians (p>0.05), which also presented values indicating the presence of a high percentage of small individuals.Finally, in the Aegean Sea (RAR 6), the median showed a gradual, but not significant, increase in years.

Aristeus antennatus
Contrary to A. foliacea, in A. antennatus it has not been possible to define clear modes, corresponding to different cohorts, in the size frequency distributions of the individuals of both sexes in each RAR (Fig. 6).
The maximum size (66 mm CL) was recorded in the Strait of Sicily and in the Ionian Sea (RAR 4 and RAR 5, respectively).However, in all macroareas, these measurements were about 60 mm CL, except for RAR 3 (Sardinian Sea), where remarkably lower values were recorded (54 mm CL) (Table 6; Fig. 6).
In the Strait of Sicily and the Aegean Sea (RAR 4 and RAR 6, respectively), given the small number of samples, it was not possible to make statistical comparisons among medians.In the remaining macroareas no statistical affinity was found with the KW test, while the Tukey test showed a few similarities among medians.
For 1994 in particular, the values for the females of RAR 1 were quite high (41 mm CL), and were remarkably different from those found for subsequent years.On the other hand, the median found for the males in 1994 had the greatest number of significant similarities (Table 6; Fig. 7).
It should be pointed out, moreover, that the Sardinian Sea (RAR 3) showed the lowest median values for both sexes, and together with the Tyrrhenian Sea (RAR 2), it also showed few statistical similarities.In the Ionian Sea (RAR 5) the females showed very different median values that were statistically comparable only for 1995 and 1996.The males presented a greater number of similarities (1997-1994; 1997-1998).
DEEP WATER RED SHRIMPS OF THE MEDITERRANEAN 115

Growth
The Von Bertalanffy growth parameters for females of A. foliacea and A. antennatus are presented for each RAR area in Table 7, and graphically presented in Fig. 8.
For the females of A. foliacea in the Algerian Provencial region (RAR 1) and those of A. antennatus in the Strait of Sicily and the Aegean Sea (RAR 4 and RAR 6, respectively), we could not carry out a modal progression because the data were too scarce.The VBGF parameters for these areas were therefore not calculated.
The growth curves of A. foliacea are all quite similar.The results of the comparison using the integral of the absolute distance between the integration extremes show quite low values (Table 7) (especially for RAR4-RAR6 and RAR3-RAR5).
In A. antennatus, on the other hand, a similar slow growth in the specimens of the Tyrrhenian (RAR2) and Sardinian seas (RAR3) can be observed (Tables 7-8), while a faster growth is observed in those of the Ionian Sea (RAR5) and the Algerian-Provencial region (RAR1) (Tables 7-8).At any rate, A. antennatus appears to be the longer-lived species.

DISCUSSION
With the MEDITS international trawl surveys, which were carried out using a common sampling methodology, it was possible to carry out, for the first time at a large scale area, a global study of the distribution and abundance of the two species of red shrimps A. antennatus and A. foliacea, and therefore to have a comparative picture of the resources present in the different areas of the Mediterranean.
Based on the quantitative analysis of the space distribution of the catches (kg/km 2 ), a different longitudinal gradient was found: A. antennatus was relatively more abundant in the westernmost part of the Mediterranean, while A. foliacea prevailed in the central and easternmost part (D'Onghia et al., 1998).This distribution trend is also confirmed in several references, which show: 1) a clearly majority presence of A. antennatus in the Spanish Mediterranean (Abelló et al., 1988;Martín, 1991;Demestre, 1994 a,b;Martínez-Baños and Mas, 1994;Sardà and Cartes, 1994;Carbonell, 1994 a,b;Carbonell et al., 2000); 2) dominance of A. antennatus, as a result of the disappearance of A. foliacea, in the Gulf of Lions (Campillo, 1994); 3) contemporaneous presence of the two shrimps with a remarkable decline in A. foliacea in the Ligurian Sea (Orsi Relini andRelini, 1985, 1994;Relini and Orsi Relini, 1987); and 4) a gradual decrease in the percentages of A. antennatus in the Tyrrhenian and Sardinian Seas, until clear dominance of A. foliacea is reached in the Straits of Sicily and the eastern Mediterranean (Ardizzone et al., 1994;Campillo, 1994;Mura et al., 1992;Ragonese, 1993;Spedicato et al., 1994;Thessalou-Legaki, 1994;Petrakis, 1998).In this gradient the Ionian Sea is an exception, since here A. foliacea was less abundant than A. antennatus, as also found by several authors (D'Onghia et al., 1998;Matarrese et al., 1992;Tursi et al., 1993).Generally, it has been possible to observe, for the different RARs, an extreme interannual variability in abundance indices (kg/km 2 ).This could undoubtedly be related to several independent and at times uncontrollable ecological factors, such as overexploitation, adverse hydrological conditions, predatory pressure on young forms (Orsi Relini and Relini, 1985;Relini and Orsi Relini, 1987), and migratory phenomena (Matarrese et al., 1992), which by affecting the stocks would change the rela- tive and absolute abundance of the species over time.Only in a few cases (Tyrrhenian Sea and Straits of Sicily for A. foliacea, and Sardinian Sea also for A. antennatus) has it been possible to locate a positive trend in the catches.On the other hand, the data from the Aegean Sea are difficult to interpret.Here a fundamental role has probably been played by the fishermen's lack of experience in catching shrimps (Thessalou-Legaki, 1994;Politou et al., 1998).
The extreme variability of the data can also be seen when comparing the median sizes.For the different analysed years, it has not been possible to find any statistical similarity among the years for either species (KW test).With multiple comparison, on the other hand, it has been possible to find similarities for the pairs of investigated years (Tukey test).In particular, the results obtained for the females of A. foliacea have almost always been confirmed also for the males, which could be justified by a similar recruiting frequency for both sexes, though a similar trend was not found for A. antennatus.
An analysis of the size frequency distribution and subsequent modal progression of female individuals in all studied regional areas has pointed out a similar pattern in the growth curve of A. foliacea, although, as seen from the comparison indices, slight differences in growth parameters were observed.As a matter of fact, the variance of L inf and K tend to compensate each other.For A. antennatus, a different situation is observed: in fact, in spite of similar values of L inf , the differences in K and t o determine two apparently different growth patterns.The latter could be due to the fact that it was not possible to properly segregate the size frequency distribution of A. antennatus into distinct cohorts (D'Onghia et al., 1994;Orsi-Relini and Relini, 1998b), and to the occurrence of a long, non-discrete recruiting season, which is not easy to determine on account of the difficulty of catching young shrimps at greater depths (Orsi-Relini and Relini, 1998a).In fact, the main differences between the growth curves appear in the younger individuals (age class 0+), which do not seem to have been fully recruited, except in RAR 2 and 3 (Tyrrhenian and Sardinian Seas).From all the VBGFs, it appears that this species has a long life cycle (more than 5 years) (Sardà and Demestre, 1987;Ragonese and Bianchini, 1996).
FIG. 1. -Map of the study area according to the MEDITS programme, with an indication of the Reference Areas (RAR) studied.
FIG. 4. -Overall length frequency distribution of Aristaeomorpha foliacea by sex for each RAR.
Aristeus antennatus: Mean abundance (in number of individuals per square km) estimated from the MEDITS trawl surveys by depth stratum, geographical sector and year Cont.).-Aristeus antennatus: Mean abundance (in number of individuals per square km) estimated from the MEDITS trawl surveys by depth stratum, geographical sector and year Aristaeomorpha foliacea: Mean biomass (in kg per square km) estimated from the MEDITS trawl surveys by depth stratum, geographical sector and year

TABLE 1 .
Fiorentino, 2000)gical parameters employed in assessment attempts for red shrimps in the Mediterranean (first part on A. antennatus fromFiorentino, 2000) *Reference point F 0.1 instead of F max ; ** mean of 7 values obtained by different formulae natus and Aristaeomorpha foliacea to support future assessment procedures.

TABLE 4 .
-Percentage contribution (in biomass) for Aristaeomorpha foliacea and Aristeus antennatus in each of the reference areas.
FIG.2.-Trends in biomass index (kg/km 2 ), from 1994 to 1999, of Aristaeomorpha foliacea for each reference area.The regression line is also shown when there is a significant correlation.

TABLE 5 .
-Annual and overall contribution (%) of males and females and basic descriptive statistics for the sizes (CL, mm) of Aristaeomorpha foliacea in each RAR.Letters in the Tukey test column indicate the pairwise similarity with the corresponding years.

TABLE 6 .
-Annual and overall contribution (%) of males and females and basic descriptive statistics for the sizes (CL, mm) of Aristeus antennatus in each RAR.Letters in the Tukey test column indicate the pairwise similarity with the corresponding years.

TABLE 8 .
-Growth curve comparison indices (The comparison indices indicate the distance between two curves).
Aristaeomorpha foliacea: Mean abundance (in number of individuals per square km) estimated from the MEDITS trawl surveys by depth stratum, geographical sector and year